April 13, 2012

Consistency in Test batsmen: a new look

A statistical analysis of consistency among Test batsmen
231

Allan Border has the lowest standard deviation among the top-20 batsmen
Allan Border has the lowest standard deviation among the top-20 batsmen © Getty Images

This is based on an idea given by Prashanth. After giving the idea and participating in a discussion or two, he disappeared off the radar. However I thank him for providing the spark.

This follows the article on "Consistency in Test bowlers: a new look" (click here). The relevant points are explained below.

1. I had used 5 Tests as the basis for bowling. However there are many Tests in which a batsman does not get a chance to bat, because of heavy top-order batting, innings wins, big wicket wins et al. Hence I have taken 10-Innings slices as the basis for batsmen analysis. This is a reasonable number and normally covers 2-3 months of Test cricket. This is normally 5-6 Tests.
2. 10 innings means that batsman can go through a Test or two of limited opportunities to bat or non-batting because of emphatic wins etc. There will be enough opportunities within the 10-innings slice to catch up.
3. There is enough time to get over short duration loss of form.
4. To measure consistency, only runs scored will be used. The fundamental cricket dictum that batsmen should score runs and bowlers should take wickets is followed. Averages are important mainly over a career and for comparisons across players.
5. Why not average? Let us take couple of examples to understand why not. Sehwag and Younis Khan have career averages just over 50 and RpT values of around 85. In a 10-innings period, match context being comparable, Younis scores 330 at 55 and Sehwag scores 450 at 45. Who has performed closer to his career figures and for that matter, better. Certainly Sehwag, despite the lower slice average.
6. Let us not forget that we remember numbers like 974 (Bradman), 774 (Gavaskar) and 688 (Lara) rather than the averages.
7. The career slices should be non-overlapping and equal, other than the last one. Gooch's 333 should be part of one career slice only. Hence the concept of rolling number of innings is not valid.
8. 10 innings might seem arbitrary but represents a long enough career slice. It represents a long 5/6 Test series.
9. The keyword is consistency with reference to the player's own career performance levels.
10. We are not looking about high and low values but only relative to the concerned player's career figures. Over a 10-innings stretch Graeme Smith is expected to score 408 runs and Habibul Bashar is expected to score 300 runs. This will be the basis. If Smith scored 350 runs, it is a below-average performance and if Bashar scored 350 runs, it is an above-average performance.
11. Adjustment is made for the last career slice if the same is fewer than 10 innings.
12. The criteria for selection is 3000 or more Test runs. 162 batsmen qualify. It is unfortunate that a few top batsmen like Graeme Pollock and George Headley do not make the cut.
14.The Standard Deviation (SD) of the slice ratios is used to determine consistency.
15.There were suggestions that I should use more Tests/innings as the basis. I have resisted that idea mainly because I want to be hard on the players. If English batsmen had a great five- Test stint in summer and a poor five-Test sojourn in winter, I want these to be treated as two out-of-the-normal occurrences and do not want to get the 10 Tests together, get a nice, middle-level performance which papers over cracks. Same with all teams. Let us also agree. If a batsman scores 180 runs in 10 innings, it is a major cause for concern and should not be covered up by 600 runs in 10 innings before or after this barren period..

The following 5 groups are formed for purposes of determining consistency. For each career-slice of 10-innings, a ratio is formed between that concerned slice's runs and the career-average runs for 10 tests. This ratio is called SPF (Slice Performance Factor). Suppose the batsman has scored 284 runs and his 10-innings and his career-RpI value is 40, the SPF value is 0.71. If he scored 501 runs, the SPF is 1.25.

A. SPF  below 0.67:  Well below average - Falls into the inconsistent bracket.
B. SPF 0.67 - 0.90:  Below average
C. SPF 0.90 - 1.10:  Around average
D. SPF 1.10 - 1.33:  Above average
E. SPF  above 1.33:  Well above average - Falls into the inconsistent bracket.

Groups B, C and D are considered to be well within the average levels. Standard Deviation is also used to determine the consistency.

First some data tables. The complete table is available for download. The tables and graphs are presented with least comments. Let me allow the erudite readers to come out with their own comments.

BatsmanTeamInningsRunsAvgeRpIMeanStdDevMid3%GrpsGrp AGrp BGrp CGrp DGrp E
         %<6767-9090-110110-133>133
Tendulkar S.RInd3111547055.4549.70.990.32568.832410756
Dravid RInd2861328852.3146.50.990.33272.42949574
Ponting R.TAus2761319653.4347.81.010.41260.72855846
Kallis J.HSaf2571237956.7848.21.000.34869.22645674
Lara B.CWin2321195352.8951.50.990.27875.02438373
Border A.RAus2651117450.5642.20.990.24181.527261063
Waugh S.RAus2601092751.0642.01.000.33361.52647546
Jayawardene D.PSlk2171044351.1948.11.000.35259.12254454
Gavaskar S.MInd2141012251.1247.31.000.35068.22246453
Chanderpaul SWin234970949.2841.51.010.30466.72445564
Sangakkara K.CSlk183938254.8751.30.990.42557.91944524
Gooch G.AEng215890042.5841.40.990.36372.72245652
Javed MiandadPak189883252.5746.71.000.38457.91936415
Inzamam-ul-HaqPak200883049.6144.11.000.37660.02053363
Laxman V.V.SInd225878145.9739.00.990.30769.62337634
Hayden M.LAus184862650.7446.90.990.38547.41961624
Richards I.V.AWin182854050.2446.90.990.40673.71927613
Stewart A.JEng235846539.5636.01.000.36866.72446644
Gower D.IEng204823144.2540.30.990.29785.72127651
Sehwag VInd167817850.8049.00.990.42852.91745314
Boycott GEng193811447.7342.00.990.33270.02034643
Smith G.CSaf174804349.6546.20.990.35066.71828224
Sobers G.St.AWin160803257.7850.21.000.30768.81633442
Waugh M.EAus209802941.8238.41.000.28376.22126643
Fleming S.PNzl189717240.0737.91.000.24784.21918442
Chappell G.SAus151711053.8647.11.000.25581.21624271
Bradman D.GAus80699699.9487.51.000.27275.0812131
Flower AZim112479451.5542.80.980.43666.71224402

To clarify the table contents. RpI mean Runs per innings. Mean is the mean of the SPF values and is close to 1.0 for all batsmen. StdDev is the Standard Deviation for all the SPF values. Mid3% is the % of the Groups B, C and D over the total number of Career Slices, which is the next column: Grps. Grp A to Grp E are self-explanatory. The complete file is available for downloading. The link is provided at the end. The first one is the core table of batsmen who have scored over 8000 runs in their Test career. In addition, Don Bradman (no need to explain), Greg Chappell (a modern great), Stephen Fleming (New Zealand) and Andy Flower (Zimbabwe) are included.

Contrary to what all of us may have perceived, Lara is remarkably consistent on this 10-innings basis. His SD of 0.278 is second only to Border amongst the top-20 batsmen. Just to confirm that this is not a fluke, look at his Mid3% which is quite high at 75.2. Again, bettered only by Border and Gower.

Consistency is determined in two ways. The first is statistical. The Standard Deviation (SD) is determined for all the ratios. Low SD values indicate consistent players and high SD values indicate inconsistent players. The usual method of using the Coefficient of Variation is not required since the means for almost all players is around 1.00. Shown below are the SD tables with the low-20 SDs indicating very consistent batsmen.

BatsmanTeamInningsRunsAvgeRpIMeanStdDevMid3%GrpsGrp AGrp BGrp CGrp DGrp E
         %<6767-9090-110110-133>133
Greig A.WEng93359940.4438.70.990.171100.01002530
Redpath I.RAus120473743.4639.51.000.19591.71204431
Ranatunga ASlk155510535.7032.91.010.20293.81607441
Hassett A.LAus69307346.5644.50.990.20485.7711230
Fredericks R.CWin109433442.4939.81.000.205100.01103440
Pietersen K.PEng143665449.2946.51.010.21086.71513731
Knott A.P.EEng149438932.7529.51.000.22886.71514541
Saeed AnwarPak91405245.5344.51.020.230100.01004150
Smith R.AEng112423643.6737.81.000.23683.31214421
Hutton LEng138697156.6750.50.990.23785.71415251
Wright J.GNzl148533437.8336.01.000.23880.01516242
Border A.RAus2651117450.5642.20.990.24181.527261063
Ijaz AhmedPak92331537.6736.00.980.24690.01004231
Fleming S.PNzl189717240.0737.91.000.24784.21918442
Mushtaq MohammadPak100364339.1736.41.000.24870.01013222
Hunte C.CWin78324545.0741.61.000.24887.5803311
Collingwood P.DEng115426040.5737.00.980.24991.71213440
Strauss A.JEng167660441.0239.51.000.25082.41709233
Sutcliffe HEng84455560.7354.20.980.25277.8912411
Chappell G.SAus151711053.8647.11.000.25581.21624271

Tony Greig is the surprise leader in this table, with a low SD value of 0.171. The most notable modern batsman in this table is Pietersen with an excellent SD value 0.210. Other than Pietersen there is no current batsman in this list. Like Lara. he has certainly surprised us. Maybe there is a lot of substance behind that exaggerated swagger. He talked about the many hours of practice put in while talking of his Colombo classic. Maybe that is paying off. It is also possible that unlike what one associates with him, he does not have extensive bad patches nor purple patches. I also wish he stops making silly statements.

The alternate method is common-sense-based rather than on a statistical measure. The two extreme group numbers, A and E, are considered significant departures from the career levels. The middle three group numbers are added and divided by the total number of slices to get the Mid3%. This reflects the consistency of the players. Shown below are the SD tables with the high-10 Mid3% values.

BatsmanTeamInningsRunsAvgeRpIMeanStdDevMid3%GrpsGrp AGrp BGrp CGrp DGrp E
         %<6767-9090-110110-133>133
Fredericks R.CWin109433442.4939.81.000.205100.01103440
Saeed AnwarPak91405245.5344.51.020.230100.01004150
Greig A.WEng93359940.4438.70.990.171100.01002530
Ranatunga ASlk155510535.7032.91.010.20293.81607441
Redpath I.RAus120473743.4639.51.000.19591.71204431
Collingwood P.DEng115426040.5737.00.980.24991.71213440
Ijaz AhmedPak92331537.6736.00.980.24690.01004231
Hunte C.CWin78324545.0741.61.000.24887.5803311
Pietersen K.PEng143665449.2946.51.010.21086.71513731
Knott A.P.EEng149438932.7529.51.000.22886.71514541
Gower D.IEng204823144.2540.30.990.29785.72127651
Cook A.NEng135618448.6945.81.000.29185.71415521
Hutton LEng138697156.6750.50.990.23785.71415251
Slater M.JAus131531242.8440.50.980.26385.71414441
Hassett A.LAus69307346.5644.50.990.20485.7711230

These are the batsmen with high middle three group % values indicating a high degree of consistency. In the bowler tables, there were six bowlers with 100% of their groups in the middle-3 groups. It seems like batting is slightly more difficult since there are only three batsmen. These all belong to the 70s/80s/90s. Roy Fredericks, the attacking West Indian batsman leads the three-some, followed by Saeed Anwar and Tony Greig. Collingwood is there as also Pietersen and Cook. Possible reason for England's pre-eminence.

Now for some special graphs.

Top run-scoring batsmen

Top run-getters in Tests career
© Anantha Narayanan

The top-9 batsmen, who have crossed 10000 Test runs, are featured. It can be clearly seen that most of these batsmen do not exhibit a high level of consistency. The only exceptions seem to be Allan Border and for the first two-thirds of his career, Jayawardene.

Most consistent: Based on low SD values

batsmen with low standard deviation values
© Anantha Narayanan

As already discussed this table is led by Tony Greig. A fairly low SD of 0.171 indicates a very consistent career. This is borne out by his placement in the next graph also. However it should be noted that the lowest SD value for bowlers is a much lower 0.124. Pietersen finds a place in both the consistency graphs.

Most consistent: Based on high Middle-3-group % values

Batsmen with high middle-3 group % values
© Anantha Narayanan

Unlike bowlers where there were six with 100% in the middle categories, amongst batsmen, there are only three: namely Fredericks, Saeed Anwar and Greig.

Least consistent: Based on high SD values

Batsmen with high standard deviation values
Batsmen with high standard deviation values © Anantha Narayanan

These graphs look like the dying person's cardiograph. These batsmen have had moves up and down throughout their career. Exemplified by Gambhir who had a poor start, great move up and then fell off equally badly. Vettori has had such a Jekyll and Hide career that it is not surprising to see him here. In the first 70 innings Vettori averaged 18. In the next 100 innings he averaged well over 35.

Least consistent: Based on low Middle-3-group % values

Batsmen with low middle-3 % values
© Anantha Narayanan

It is clear that these two methods of determining consistency are quite different. There are different sets of batsmen in the two graphs.

Batsmen with top RpI figures

Batsmen with highest RPI figures
© Anantha Narayanan

Just to complete the analysis I have given here the charts for the top batsmen - by Runs per innings, since most of them would have missed the first chart: by career runs scored. Again inconsistency seems to be the trend here.

I think mention must be made of two batsmen, Tony Greig and Kevin Pietersen. Tony Greig never went off the middle three groups. That is some level of consistency. Pietersen, amongst the modern batsmen, has surprised us with his high degree of consistency.

To download/view the Excel sheet containing the complete data for 162 batsmen please click/right-click here. I have strengthened the Excel sheet by colour coding the individual SPF values through dynamic formatting.

Ed Smith's thought-provoking piece on randomness and form "When is poor form just randomness?" (click here) made me realize that this particular measure I have created can be applied to Ed Smith's axiom. Suppose I summed the SPF values of the top six batsmen or top four bowlers for every Test/innings, we would know what are the lowest SPF averages (very poor form, as a group of six/four players) and the highest SPF averages (very rich form, as a group of six/four players). That, for a later article.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems

Comments have now been closed for this article

  • Vinay on April 25, 2012, 8:40 GMT

    Consistency against individual career average itself is inconsistent. There should be a bench mark just like your 3000 run marks, either by particular number for career average or particular number of runs per innings. What is the purpose if we are analysing against whether a batsman consistently scoring 30 runs or 25 runs if his career average is 30 or 25? First of all, what is consistency ? Whether he is consistently 'an average batsman' or consistently 'a good batsman' or consistently 'a weak batmsan'? There should be some purpose. Secondly for sure slice of 10 is too big. Better to have slice of 4 and 6 and 8 etc, and then average it. By doing all thes,e We can excel in mathematics,charting,analysing or deriving something, but what is the purpose? We should have some target to derive something.

  • Alex on April 23, 2012, 16:25 GMT

    @Anant, @Prashanth: Ananth, Prashanth is only trying to take you back to his unanswered original qn, which was quite relevant. The qn: how often did a batsman let down his team vs how often he delivered? I suggest the following answer. Let's make 3 cases.

    Case 1 = only 1 innings; Case 2 = he played in both innings & upto 6 wkts fell in 2nd; Case 3 = both innings completed.

    Let's say fail = 1 if

    1. less than A1= min of (30 runs & 25% team runs) if Case 1. 2. less than A2= min of (50 runs & 25% team runs) if Case 2. 3. less than A3= min of (70 runs & 20% team runs) if Case 3.

    I use "minimum of" function to pass smaller scores in low scoring matches. Then, success = 1- fail. In this binary rating, fail = 1 or 0, & success = 0 or 1. This can be fine tuned to

    fail =

    1. B1 = runs scored/ A1 (if Case 1); 2. B2 = runs scored/A2 (if Case 2); 3. B3 = runs scored/A3 (if Case 3).

    This lets fail be between 0 & 1. We may still rate success as 1 or 0, as per Prashanth. [[ Might be, but there is a way of doing things especially when I have said that I will look at it later. I really do not want to look at it now. Thanks, Alex. Ananth: ]]

  • Prashanth on April 23, 2012, 13:14 GMT

    Ananth, ???! OK- my last post on the matter. Complete logical inconsistency and breakdown on your part I fear. Your argument seems to be that since this has turned out to be a statistician’s delight we must ignore reality. A Test “series” is a “collection” or a “series” of INDIVIDUAL matches. It is not ONE single match, it is a collection of individual matches. The “end” result of a series depends on the results again of INDIVIDUAL matches. Again ,to use an extension of my previous example if we have a 3 Test series: If a batsman “lets his team down” in say the first 2 and then scores heavily in the 3rd…then overall the batsman may actually be labelled “matchlosing” ! Say a string of 30,30,25,35, 130,60 The 2 matches which may have been “saved” or won have been lost…and higher scores in the 3rd do NOT impact them. The summing up of the entire “string” does NOT have any real cricketing meaning. Re. all the various statistical delights you point out – that is all very well and good fun. But they unfortunately do not have any connection to cricketing reality.

    Refusing to accept this fact is actually colouring reality in favour of statistical abstractions. [[ Thank you for your contributions so far. Your negative views, the disdain you are showing for me and the other readers and the words you have used make me think that you should find better blogs to visit. Ananth: ]]

  • Prashanth on April 23, 2012, 10:28 GMT

    Milpand Thank you for taking time out for your clear and lucid explanation. All well appreciated. I only wish I was capable of such clarity. But I will try one last time. My point is that the statistical methods and theory used (however valid and fundamentally acceptable in themselves) must actually find parallels in “real life” if they are to be of any value. There exists no such thing as a “slice” (of more than a single match) in cricket. It is a pure fabrication. So, straight off, we have committed a cardinal sin. We are immediately indulging in fantasy. When we use multiple “strings” or “slices” of innings we immediately lose contact with “real life”. A real life Test match comprises of maximum 2 innings. The longest string we can use is 2 innings and that too ONLY in the match in question. Anything else is artificial and un-real. As mentioned if a batsman is credited with a “matchwinning” performance in a particular match that performance is perforce restricted to that particular match alone. The scores in that particular match cannot be spread out, dispersed, or clumped together with other innings in other matches- which is precisely what is being done when clumping together various strings/slices. This is non-sensical in cricketing terms. Perhaps the easiest way of looking at it would be from the team’s point of view. If a batsman underperforms significantly in any given match this may lead to a team losing that particular match. Subsequent or prior innings on either side have absolutely NO impact to the match in question. An eg. would be scores of 20, 30 (in one match), and then 100 in the first innings of the next match. If the 20, 30 from a main batsman “leads” to a team losing, the result cannot be reversed by the subsequent 100 in the NEXT (or prior) match. This is simply not how things transpire in real life. A slice/string comprising of 8/10 (or such) innings is a purely artificial construct with no parallels whatsoever in real life cricket. A poor match by a batsman may well contribute to his team losing. Scores on either side simply do not matter. The only slice we ACTUALLY ever encounter in cricket is that of a single, solitary match. I full well know I am being repetitive – but hope that I am getting at least some of my argument across. [[ So, according to you, a Test Series is a fabrication and analysis based on a Test Series a cardinal sin. For that matter, an ODI tournament. Okay a Test series has the benefit of same opponent and same location. The slice, however may be against different opponents and in different locations but has the advantage of uniformity in size for comparison purposes. I agree that I did not address your specific point. I have also told that I would look at it later. But why are you pulling down this analysis. It has a major weakness in the impact of starting point. But how many people have derived good insights and revealing facets of player careers. And how many valuable aspects of individual innings distribution it has led us into. And what are all the types of analysis of basic statistical measures it has taken us into. Coloured glasses are great in the sun, not inside where they should be taken off to get a clear view. Ananth: ]]

  • Anshu N Jain on April 23, 2012, 10:09 GMT

    Thanks Milind for your lucid example.

    Ananth: As Milind pointed out, Individual scores suffer from the asymmetric distribution syndrome. But samples of averages/RpIs from a population (career innings) are normally distributed, and the method that I wrote to you about in detail could be evaluated further.

    Also, an analysis based on averages/RpIs in various innings, against various countries, across various results etc. with S.D. as a consistency measure can be used in conjunction with the consistency measure derived using the boundary values of Median, HQ and LQ, either for combination or for comparison? [[ Yes, all this makes sense. At a later date, let us re-visit this. You guys don't forgot what you wrote so that if I miss something you could remind me. Ananth: ]]

  • Samir Gupta on April 23, 2012, 8:36 GMT

    Oh! and BTW, may be a similar analysis for bowlers. I suspect that Hadlee, Murali and Kapil may do rather well [[ Samir, I think you must have had a tough day today !!!. The article before the Test batsmen consistency analysis was the Test bowlers consistency one. Ananth: ]]

  • Samir Gupta on April 23, 2012, 8:34 GMT

    Ananth,

    I would like to know who are the Atlases of cricket. How well does a batsman do when others have failed? I know that you have done analysis of individual innings that have stood out in a match. Have you done an analysis over a career? An interesting stat is how many times in his career (as a ratio of the total number of innings) has a batsman faced more than 33% of the balls bowled or scored more than 40% of the runs. It may be a good idea to give a higher weightage to a fourth innings score as compared to a first innings score. A match saving innings is again more valuable than a match winning innings (because we are looking at Atlases). My suspicion is that batsmen like Dravid, Border, Gavaskar, Headley and Lara will do well as compared to Tendulkar, Ponting and Richards. Not sure about Bradman. He was so far ahead of everyone else, it may be a good idea to exclude him from this analysis. [[ Samir The article starts with a reference to the Batsmen peer comparison analysis which was done a few months back. Ananth: ]]

  • Alex on April 23, 2012, 4:23 GMT

    @Ananth & @milpand: Nice observation on Martin. Pommie Mbangwa was even more consistent, I think. He too has a median of zero and trumps Martin on both ave (2 vs 2.16) and SR (17 vs 20). [[ I am not going to take this nonsense lying down !!!. I admit, when I did the worst Test batsmen piece a few years back, Pommie upstaged Martin and many Kiwis were miffed, justifiably so. No, sir, not now. One day I will revisit that theme and this time it WILL be Chris Martin since I will set the cut-off sufficiently high for poor Pommy to stay out of the analysis and stay in the commentator's box and lament. I may compromise on Lara but never on Martin. Ananth: ]]

  • milpand on April 22, 2012, 23:21 GMT

    Ananth, a simple measure that ranks these selected players on the basis of their likelihood to post a score within a range closer to their own middle values (Mean, Median, RpI, Average) as below:

    Greig: 1.38 Border: 1.41 Dravid: 1.51 Lara: 1.56 SRT: 1.66 Ponting: 1.8 Bradman: 1.9 Gambhir: 2.11

    makes sense to me.

    [[ THis is the (HQ-Median)/(Median-LQ) value. Looks to be a very sensible methodology. Anyhow we have more time for the follow-up work. Ananth: ]]

    I am a great fan of Border. If this measure rates him highly then it must be good.

    I like Chris Martin for his batting. This measure is rubbish if he does not fare well. [[ I am not going to take this from you. I am a greater fan of Martin than you are. He is the only one I will pay to watch bat. Between 1 and 10 balls of fun. So much so, I created Martin's career file and that is summarized below. 12, 7,7,7, 5,5,5,5, 4,4,4,4,4,4,4,4, 3,3, 2,2,2,2,2,2,2, 1 (14 times) and 0 (59 times). HQ=2; Median=0; LQ=0. The above ratio is (2-0)/(0-0) ??? "What do I do when the edian is 0 ???". You solve the problem. I would give him an honorary 1.00. Ananth: ]]

  • milpand on April 22, 2012, 23:18 GMT

    I have deliberately included a number of innings to get past the familiar figures of 334, 309, 299, 254 and 0. Let us revisit LQ, Median, Mean, HQ at the end of 40th and 80th innings. After 40: 17, 46, 88.2, 125 After 80: 17.5, 56.5, 87.5, 133.5

    Stdev can be used in a normal distibution. Hence statements like: "About 68% of values drawn from a normal distribution are within one standard deviation away from the mean; about 95% of the values lie within two standard deviations; and about 99.7% are within three standard deviations." are not unfamiliar.

    But skewed distributions are asymmetic, not normal. Skewness is a measure of asymmetry. If distribution is symmetric then mean equals median. Therefore there is some value in looking at the ratio of median to mean to understand asymmetry. Also HQ & LQ are the best available alternatives where stdev can't be used.

    Box plot is the pictorial view for a skewed distribution that combines 6 important values other than famous cricket average. [[ Excellent explanation, Milind. I am sure readers would benefit a lot. I think there is a lot of value in looking at the three key figures, HQ, Median, LQ and the Mean. For cricket distributions the High score is useless. What does it matter if Lara's HS is 300 or 400. And almost all will have a string of low scores, at 0. Ananth: ]]

  • Vinay on April 25, 2012, 8:40 GMT

    Consistency against individual career average itself is inconsistent. There should be a bench mark just like your 3000 run marks, either by particular number for career average or particular number of runs per innings. What is the purpose if we are analysing against whether a batsman consistently scoring 30 runs or 25 runs if his career average is 30 or 25? First of all, what is consistency ? Whether he is consistently 'an average batsman' or consistently 'a good batsman' or consistently 'a weak batmsan'? There should be some purpose. Secondly for sure slice of 10 is too big. Better to have slice of 4 and 6 and 8 etc, and then average it. By doing all thes,e We can excel in mathematics,charting,analysing or deriving something, but what is the purpose? We should have some target to derive something.

  • Alex on April 23, 2012, 16:25 GMT

    @Anant, @Prashanth: Ananth, Prashanth is only trying to take you back to his unanswered original qn, which was quite relevant. The qn: how often did a batsman let down his team vs how often he delivered? I suggest the following answer. Let's make 3 cases.

    Case 1 = only 1 innings; Case 2 = he played in both innings & upto 6 wkts fell in 2nd; Case 3 = both innings completed.

    Let's say fail = 1 if

    1. less than A1= min of (30 runs & 25% team runs) if Case 1. 2. less than A2= min of (50 runs & 25% team runs) if Case 2. 3. less than A3= min of (70 runs & 20% team runs) if Case 3.

    I use "minimum of" function to pass smaller scores in low scoring matches. Then, success = 1- fail. In this binary rating, fail = 1 or 0, & success = 0 or 1. This can be fine tuned to

    fail =

    1. B1 = runs scored/ A1 (if Case 1); 2. B2 = runs scored/A2 (if Case 2); 3. B3 = runs scored/A3 (if Case 3).

    This lets fail be between 0 & 1. We may still rate success as 1 or 0, as per Prashanth. [[ Might be, but there is a way of doing things especially when I have said that I will look at it later. I really do not want to look at it now. Thanks, Alex. Ananth: ]]

  • Prashanth on April 23, 2012, 13:14 GMT

    Ananth, ???! OK- my last post on the matter. Complete logical inconsistency and breakdown on your part I fear. Your argument seems to be that since this has turned out to be a statistician’s delight we must ignore reality. A Test “series” is a “collection” or a “series” of INDIVIDUAL matches. It is not ONE single match, it is a collection of individual matches. The “end” result of a series depends on the results again of INDIVIDUAL matches. Again ,to use an extension of my previous example if we have a 3 Test series: If a batsman “lets his team down” in say the first 2 and then scores heavily in the 3rd…then overall the batsman may actually be labelled “matchlosing” ! Say a string of 30,30,25,35, 130,60 The 2 matches which may have been “saved” or won have been lost…and higher scores in the 3rd do NOT impact them. The summing up of the entire “string” does NOT have any real cricketing meaning. Re. all the various statistical delights you point out – that is all very well and good fun. But they unfortunately do not have any connection to cricketing reality.

    Refusing to accept this fact is actually colouring reality in favour of statistical abstractions. [[ Thank you for your contributions so far. Your negative views, the disdain you are showing for me and the other readers and the words you have used make me think that you should find better blogs to visit. Ananth: ]]

  • Prashanth on April 23, 2012, 10:28 GMT

    Milpand Thank you for taking time out for your clear and lucid explanation. All well appreciated. I only wish I was capable of such clarity. But I will try one last time. My point is that the statistical methods and theory used (however valid and fundamentally acceptable in themselves) must actually find parallels in “real life” if they are to be of any value. There exists no such thing as a “slice” (of more than a single match) in cricket. It is a pure fabrication. So, straight off, we have committed a cardinal sin. We are immediately indulging in fantasy. When we use multiple “strings” or “slices” of innings we immediately lose contact with “real life”. A real life Test match comprises of maximum 2 innings. The longest string we can use is 2 innings and that too ONLY in the match in question. Anything else is artificial and un-real. As mentioned if a batsman is credited with a “matchwinning” performance in a particular match that performance is perforce restricted to that particular match alone. The scores in that particular match cannot be spread out, dispersed, or clumped together with other innings in other matches- which is precisely what is being done when clumping together various strings/slices. This is non-sensical in cricketing terms. Perhaps the easiest way of looking at it would be from the team’s point of view. If a batsman underperforms significantly in any given match this may lead to a team losing that particular match. Subsequent or prior innings on either side have absolutely NO impact to the match in question. An eg. would be scores of 20, 30 (in one match), and then 100 in the first innings of the next match. If the 20, 30 from a main batsman “leads” to a team losing, the result cannot be reversed by the subsequent 100 in the NEXT (or prior) match. This is simply not how things transpire in real life. A slice/string comprising of 8/10 (or such) innings is a purely artificial construct with no parallels whatsoever in real life cricket. A poor match by a batsman may well contribute to his team losing. Scores on either side simply do not matter. The only slice we ACTUALLY ever encounter in cricket is that of a single, solitary match. I full well know I am being repetitive – but hope that I am getting at least some of my argument across. [[ So, according to you, a Test Series is a fabrication and analysis based on a Test Series a cardinal sin. For that matter, an ODI tournament. Okay a Test series has the benefit of same opponent and same location. The slice, however may be against different opponents and in different locations but has the advantage of uniformity in size for comparison purposes. I agree that I did not address your specific point. I have also told that I would look at it later. But why are you pulling down this analysis. It has a major weakness in the impact of starting point. But how many people have derived good insights and revealing facets of player careers. And how many valuable aspects of individual innings distribution it has led us into. And what are all the types of analysis of basic statistical measures it has taken us into. Coloured glasses are great in the sun, not inside where they should be taken off to get a clear view. Ananth: ]]

  • Anshu N Jain on April 23, 2012, 10:09 GMT

    Thanks Milind for your lucid example.

    Ananth: As Milind pointed out, Individual scores suffer from the asymmetric distribution syndrome. But samples of averages/RpIs from a population (career innings) are normally distributed, and the method that I wrote to you about in detail could be evaluated further.

    Also, an analysis based on averages/RpIs in various innings, against various countries, across various results etc. with S.D. as a consistency measure can be used in conjunction with the consistency measure derived using the boundary values of Median, HQ and LQ, either for combination or for comparison? [[ Yes, all this makes sense. At a later date, let us re-visit this. You guys don't forgot what you wrote so that if I miss something you could remind me. Ananth: ]]

  • Samir Gupta on April 23, 2012, 8:36 GMT

    Oh! and BTW, may be a similar analysis for bowlers. I suspect that Hadlee, Murali and Kapil may do rather well [[ Samir, I think you must have had a tough day today !!!. The article before the Test batsmen consistency analysis was the Test bowlers consistency one. Ananth: ]]

  • Samir Gupta on April 23, 2012, 8:34 GMT

    Ananth,

    I would like to know who are the Atlases of cricket. How well does a batsman do when others have failed? I know that you have done analysis of individual innings that have stood out in a match. Have you done an analysis over a career? An interesting stat is how many times in his career (as a ratio of the total number of innings) has a batsman faced more than 33% of the balls bowled or scored more than 40% of the runs. It may be a good idea to give a higher weightage to a fourth innings score as compared to a first innings score. A match saving innings is again more valuable than a match winning innings (because we are looking at Atlases). My suspicion is that batsmen like Dravid, Border, Gavaskar, Headley and Lara will do well as compared to Tendulkar, Ponting and Richards. Not sure about Bradman. He was so far ahead of everyone else, it may be a good idea to exclude him from this analysis. [[ Samir The article starts with a reference to the Batsmen peer comparison analysis which was done a few months back. Ananth: ]]

  • Alex on April 23, 2012, 4:23 GMT

    @Ananth & @milpand: Nice observation on Martin. Pommie Mbangwa was even more consistent, I think. He too has a median of zero and trumps Martin on both ave (2 vs 2.16) and SR (17 vs 20). [[ I am not going to take this nonsense lying down !!!. I admit, when I did the worst Test batsmen piece a few years back, Pommie upstaged Martin and many Kiwis were miffed, justifiably so. No, sir, not now. One day I will revisit that theme and this time it WILL be Chris Martin since I will set the cut-off sufficiently high for poor Pommy to stay out of the analysis and stay in the commentator's box and lament. I may compromise on Lara but never on Martin. Ananth: ]]

  • milpand on April 22, 2012, 23:21 GMT

    Ananth, a simple measure that ranks these selected players on the basis of their likelihood to post a score within a range closer to their own middle values (Mean, Median, RpI, Average) as below:

    Greig: 1.38 Border: 1.41 Dravid: 1.51 Lara: 1.56 SRT: 1.66 Ponting: 1.8 Bradman: 1.9 Gambhir: 2.11

    makes sense to me.

    [[ THis is the (HQ-Median)/(Median-LQ) value. Looks to be a very sensible methodology. Anyhow we have more time for the follow-up work. Ananth: ]]

    I am a great fan of Border. If this measure rates him highly then it must be good.

    I like Chris Martin for his batting. This measure is rubbish if he does not fare well. [[ I am not going to take this from you. I am a greater fan of Martin than you are. He is the only one I will pay to watch bat. Between 1 and 10 balls of fun. So much so, I created Martin's career file and that is summarized below. 12, 7,7,7, 5,5,5,5, 4,4,4,4,4,4,4,4, 3,3, 2,2,2,2,2,2,2, 1 (14 times) and 0 (59 times). HQ=2; Median=0; LQ=0. The above ratio is (2-0)/(0-0) ??? "What do I do when the edian is 0 ???". You solve the problem. I would give him an honorary 1.00. Ananth: ]]

  • milpand on April 22, 2012, 23:18 GMT

    I have deliberately included a number of innings to get past the familiar figures of 334, 309, 299, 254 and 0. Let us revisit LQ, Median, Mean, HQ at the end of 40th and 80th innings. After 40: 17, 46, 88.2, 125 After 80: 17.5, 56.5, 87.5, 133.5

    Stdev can be used in a normal distibution. Hence statements like: "About 68% of values drawn from a normal distribution are within one standard deviation away from the mean; about 95% of the values lie within two standard deviations; and about 99.7% are within three standard deviations." are not unfamiliar.

    But skewed distributions are asymmetic, not normal. Skewness is a measure of asymmetry. If distribution is symmetric then mean equals median. Therefore there is some value in looking at the ratio of median to mean to understand asymmetry. Also HQ & LQ are the best available alternatives where stdev can't be used.

    Box plot is the pictorial view for a skewed distribution that combines 6 important values other than famous cricket average. [[ Excellent explanation, Milind. I am sure readers would benefit a lot. I think there is a lot of value in looking at the three key figures, HQ, Median, LQ and the Mean. For cricket distributions the High score is useless. What does it matter if Lara's HS is 300 or 400. And almost all will have a string of low scores, at 0. Ananth: ]]

  • milpand on April 22, 2012, 23:17 GMT

    His next 8 scores are 1,4,8,14,131,232,254,334. Median is still 49. He still fails to reach 13 (not 32) once every 4 innings (LQ). But HQ rises to 125 from 87 whereas mean climbs to 90.4 from 59.

    Further 8 scores of 0,2,25,43,112,152,223,226 confirm that Bradman still can't reach 13 once every 4 innings. Median has marginally moved up to 50.5 and RpI to 92.9. HQ now stands at 136.

    After 8 more scores of 0,8,24,66,76,103,167,299 mean does not change. LQ too remains unchanged with 2 more scores of less than 13 Two scores above 136 indicate that HQ is also unchnaged. Only median rises to 62.

    In the next 8 innings, Bradman addresses his single digit starts with 13,25,29,30,36,48,71 & 304. This increased LQ to 17. 7 two-digit scores bring down the median to 46. Runs per Innings marginally down to 88.2. At halfway stage of his career HQ is 125.

    He ended his career after 40 more innings but LQ remained 17.5. Median increased to 56.5 and HQ to 133.5. His final RpI of 87.5 is well known.

  • milpand on April 22, 2012, 23:17 GMT

    I understand why a few terms I used were labelled 'jargon'. Hence I am going to use an example:

    Weigh randomly selected 100 people. Most will weigh around average but very few will be far too obese or severely malnourished (but none will weigh exactly zero). This is an example of "normal distribution considered most prominent" - see http://en.wikipedia.org/wiki/Normal_distribution A list of batting scores say 1,18,37,40,58,79,112,123 where a batsman always starts at 0 indicates it is a SKEWED distribution. Batsman is more likely to be dismissed early.

    The bell shaped normal curve has its peak located around the mean. The avg is 59 for the 8 scores listed here but median is 49, halfway between 40(4th obs) & 58(5th obs) becuase this is not a normal distribution. Here batsman failed to reach 32 once every 4th innings (first quartile or LQ) but scores more than 87 in 2 out of 8 times (third quartile or HQ).

  • Samir Gupta on April 22, 2012, 20:49 GMT

    Ananth,

    The BCL vs SRT vs Ponting vs Kallis debate has been raging for very long. Is there some statistical method to resolve it?

    Maybe a combination of RPI, significant contribution in saving matches, winning matches, drawing them, the analysis you did about the quality of opposition and runs scored against them, The sheer longevity of the batsman, the peer analysis you have done, giving more weightage to runs scored in a low scoring match.....

    I suspect that difference between the four will be statistically insignifant.

    BTW, I do support your pleas to not compare greats. [[ Yes, the bottomline is that one can only appreciate A's greatness if he learns to appreciate B's claim. If I do not understand and appreciate what Rod Laver achieved, I cannot say I can appreciate Federer. Personal preferences are personal. There they should stop. Ananth: ]]

  • Anshu N Jain on April 22, 2012, 18:21 GMT

    Limited overs games have their own share of challenges - no second innings for batsmen and bowlers to improve their showing, and restrictions on over resources in the only innings available(both for batting and bowling teams). It is an examination of character not any less than that required of a Test match. [[ Anshu I am sure you have seen the stat I had put in response to Shrikanth. In 5% of matches, teams scoring below 200 have successfully defended the low totals. In another 5% of the matches, these sub-200 totals have been chased with difficulty, the team losing 6 wickets or more. On the other hand, 40 matches (just over 1%) of matches where 300+ totals have been chased successfully. So the game is a finely nuanced one once people put their mind to understand and appreciate. Ananth: ]] That said, I agree that T20 is not a sport and pure entertainment with a lot of paraphernalia around the core, so much so that the essence of the core has been diluted. I personally dont hold either T20I or the IPL circus in any worthwhile regard.

  • Anshu N Jain on April 22, 2012, 18:11 GMT

    My 2-pence in this interesting debate on cricket's "existential crisis" as shrikanthk put it:

    A Test is won by the team that dismisses the other team twice for fewer runs. Resources available to each team - wickets only. Resources available to both teams - time and overs. An ODI or a T20 is won by the team that scores more runs given the same amount of resources to each team - overs, wickets and time. Except for a couple of artificial constructs (field restrictions and over restrictions for bowlers), the fundamental skills being tested - batting and bowling - and the end objective remain the same in both formats - making the most of resources available. The limited overs game does not allow for a "Draw" as a likely outcome because both teams get the same exact resources to make the most of. It would be unfair to still end up with a "cant decide who won/lost". contd..

  • shrikanthk on April 22, 2012, 16:56 GMT

    Some final clarifications -

    In non-international matches, time limitations have been a norm, for example, in school and local/club cricket

    Even proper cricket matches have never had a strict time limit. A game of cricket can be played over 1 day, 5 days or even to a finish. No matter what the time limit is, you cannot win a cricket match without bowling the opposition out. The basic premise of limited overs cricket contradicts that. Nothing wrong with that as long as people understand why Limited overs cricket came into existence in the first place. If they do understand that, they won't make this mistake of taking ODI stats too seriously.

    I haven't heard rapid fire chess being called in a different name

    Firstly rapid fire chess is more or less the same game. It is not structurally different from traditional chess the way Limited overs cricket is different from proper cricket. Also rapid chess has not cannibalised the parent game!

  • Vimalan on April 22, 2012, 16:24 GMT

    Otherwise they would not insist on calling the limited overs game as Cricket. They will openly rebel and form a new game of their own which will compete with Official Cricket for eye-balls. But they don't have the confidence to do it. ----- @shrikanthk

    Most of the people who love ODIs at least don't hate Test cricket like some of the hard core Test cricket enthusiasts like you hate ODIs. Many of them follow Test cricket through scores and highlights if not sitting through 5 days. So why would there be a need to rebel and call ODIs as a different sport ? Why can't all these formats co-exist under the same umbrella ? I haven't heard rapid fire chess being called in a different name. [[ There are three formats in Chess. Blitz, Rapid and Standard time control. The first two are independent formats as also tie-breakers. Ananth: ]]

  • Alex on April 22, 2012, 15:18 GMT

    @shrikanthk & @Ananth: I don't think anyone on this forum puts test cricket down. IMO, test cricket is #1 followed by the ODIs. T20 is an experiment to take cricket beyond the confines of commonwealth. IPL is just money making & entertainment disguised as sport.

    Until Packer popularized ODIs in '79, I don't think ODI's caught on in earnest after the concept was first introduced @1970. Ideally, ODI's should have restricted only the duration of the game to 1 day and allowed for draw, more than 10 overs/bowler, fewer fielding restrictions, etc. That would have made the game more absorbing (indeed, early ODI's were less restricted 60-over affairs). However, that still won't ensure more art (in your sense): even in tests, supporters prefer ugly 50 runs to stylish 22 ... Chanderpaul's entire career is built on that premise!!

    Forget Ambanis & Tatas, Sharad Pawar is the shrewdest businessman of modern India. Money has always dictated how the sport evolves and he is doing the same.

  • Aditya Nath Jha on April 22, 2012, 15:13 GMT

    @shrikant (with apologies to anantha/readers, for arguing on a point that's orthogonal to the article) - "Otherwise they would not insist on calling the limited overs game as Cricket" is like saying beatles shouldn't call what they produced as music. Music has also always been run by establishment - earlier it was the record labels and now it's apple. They have a portfolio approach to music (look at the categories in grammy). The cricket establishment is also taking a portfolio approach. As for your other point about "The truth is that the so-called classical art and popular art have always co-existed in history" - all forms of cricket have co-existed. In non-international matches, time limitations have been a norm, for example, in school and local/club cricket. The analogy may not be exact, but they are pretty similar. I love test matches and i think pollard can't bat for more than 30 balls. But I am arguing against your stand of "anything other than test cricket is not cricket".

  • shrikanthk on April 22, 2012, 12:10 GMT

    There will be "conservatives" like you - and you have every right to your views

    I have no problem being labeled a conservative. Which I proudly am. The only issue I have with your remark is that the White ball enthusiasts are in some respects even more conservative and hidebound than me!!

    Otherwise they would not insist on calling the limited overs game as Cricket. They will openly rebel and form a new game of their own which will compete with Official Cricket for eye-balls. But they don't have the confidence to do it. Packer tried it once and his breakaway leagues weren't a huge success. Hence the truce with ACB following which Ch.9 got the telecast rights for the official cricket season.

    The One-dayer/T20 is where it is today because of the benevolence of the "Establishment".

    Anyway, thanks Ananth for letting me rant a bit. Shame that it was a one-sided discussion. But it's worth it if it makes atleast a few people think harder about Cricket's current identity crisis.

  • shrikanthk on April 22, 2012, 11:29 GMT

    whenever a classical/orthodox format moves away towards a more popular one, opinion gets divided

    Aditya: I have heard this line many many times. The assumption is that there is this linear progression from classical art to popular art. The truth is that the so-called classical art and popular art have always co-existed in history.

    When the likes of James Joyce and Virginia Woolf were writing their novels, there were other Edwardian writers who wrote pulp thrillers, racy romance novels among other things, at the same time.

    Folk music is for instance is much older and richer in its heritage than classical music.

    The parallel with Limited overs cricket is flawed because the ODIs/T20s are not popular forms of art that evolved in parallel. Unlike Jazz or Rock'n'roll, these cricket formats are direct descendants of cricket created by the bureaucratic pen at Lord's. They have cannibalised the parent game without necessarily expanding its base beyond the English speaking world!

  • Prashanth on April 22, 2012, 11:28 GMT

    Ananth, Thanks. I have again browsed through the comments and to me the basic premise with which we started this trend has not been addressed. The topic started with regard to “matchwinning” batsmen or innings. The definition of which was taken to mean a single performance well above norm in a single particular match. I countered with a scenario wherein the same batsman due to his lack of runs in other matches should perhaps not be considered “matchwinning” in the “long run” – since his non-performance or lack of performance in other INDIVIDUAL matches (not slices, or “averages of slices”) may result in more losses to the team than the wins which the odd blinding innings may help contribute towards. Surely if a batsman or innings is considered “matchwinning” in a particular match due to performance well above the norm- then other matches wherein performance is well below the norm may be labeled “matchlosing”?

    In effect only match scores well below norm are what should be accounted for when considering a “matchlosing” performance. A “matchwinning” performance in a match is just that. The amount of runs scored in a SINGLE particular match. These runs simply cannot be “dispersed”, so to speak, over various “slices”. A “slice” – of any amount, or calculated in anyway (rolling etc) simply has no real-life-cricketing meaning whatsoever when discussing particular individual matches. A “matchwinning” performance can only possibly apply to the particular match in question. Similarly a “matchlosing” performance has necessarily to be taken on a match-by-match basis. Slices defy all actual, real-life cricketing practicalities. A batsman’s performance over a “slice” (except on an individual match basis) has no bearing whatever on individual match results- in real-cricket-life results individual match scores are ALL that matter. “Slices” (except matchwise) are an imaginary and un-real non- cricketing construct. I appreciate that “slices” may be a theoretical convenience. However, at some point theory must relate to reality. In no way can a “slice” of several innings be equated with consistency since the average of a slice has almost zero relation to individual matches – and performance in individual matches is ALL that matters. Not an average of a random “slice” of matches. So, the thought process by which I started the original line of inquiry has been completely “square-cut”. [[ Twice I have thanked for the spark you provided. I am not going to withdraw that. I will still thank you one more time. Maybe interpreted wrongly, but a most discussed article. One day I will look at your idea more carefully and do proper justice to it. Ananth: ]]

  • Aditya Nath Jha on April 22, 2012, 7:57 GMT

    @shrikanthk - whenever a classical/orthodox format moves away towards a more popular one, opinion gets divided. There will be "conservatives" like you - and you have every right to your views, of course - and there will be those who actually enjoy the new, more "popular" formats. Music is a great example. Classical music remains the orthodox format, but rock n' roll, pop, hindi filmi music - and even rap - have their own followers. Many people have tastes that cut across genres and many more stick to one. Any debate about which form is better or who is a better singer is actually moot. There are different kinds of skills involved and your choice of who/which is better depends on which set of skills you resonate with more.

  • shrikanthk on April 22, 2012, 6:36 GMT

    you have made the assumption that I have not considering the Test finishing skills as worhwhile. My point was only on your comment related to ODIs

    I made no such assumption. My replies are directed at a larger audience which thinks that ODIs have somehow introduced this new art of "finishing games" unknown to cricket until the 80s. People have been "finishing" games ever since they started playing this great game in the 17th century.

    Maybe I am deliberately being argumentative. But the objective is to make people think hard and objectively about this "phenomenon" of Limited over cricket, instead of getting carried away by the propaganda that goes on in the media.

    One such evil propaganda is that One-dayers have hastened the "pace" of the traditional game. Bull. Ananth will tell you that run-rates in tests haven't changed all that drastically over the years. And in fact, they've declined a bit recently. A Hussey who motors at SR of 150 in T20 grafts at an SR of 40 in Tests!

  • shrikanthk on April 22, 2012, 6:09 GMT

    demeaning to the finishing skills of players like Bevan and Hussey, who made finishing the game a great treat to watch

    Not demeaning anybody. I am discussing broader questions here and not specific players or specific innings like Hussey's 60 which keeps squeezing its way into these conversations.

    "Finishing the game" is not a specific limited over game skill. It is a cricketing skill. Alan Davidson and Richie Benaud displayed remarkable "finishing" skills when they put on that memorable partnership on the 5th day of the 1960 Tied test (against a very good attack mind you, with no formulaic field restrictions to help them out). Same is true for the batting of Watson and Hussey in the recently concluded Barbados test. How about the finishing skills of Arthur Morris and Don Bradman in the Leeds test match of 1948 when Aus chased 400 on a dusty turning track!

    And these skills are harder to execute in a test given the draw option which puts dilemmas in the batsman's mind! [[ See, again, you have made the assumption that I have not considering the Test finishing skills as worhwhile. My point was only on your comment related to ODIs. Why bring in extraneous Test skills, which have never been questioned. Ananth: ]]

  • shrikanthk on April 22, 2012, 5:38 GMT

    I think the conceptualization & execution level of western paintings (including former USSR) since 1900 is far higher than it used to be over 1600-1900. Some may still prefer Da Vinci, Monet, & Rembrandt but, IMO, guys like Picasso, Dali, and Kandinsky far surpass the old greats

    There's no point rating artists of different hues on a scale. At any rate, Modern art as pioneered by the likes of Picasso was a genuine artistic expression. It was a product of a distictive sensibility. Not the product of commercial equations like T20 or the Gillette Cup (which was the first ever limited over tournament). I simply don't see a parallel between the emergence of limited overs cricket and evolution of new art movements over the last few centuries.

  • shrikanthk on April 22, 2012, 5:14 GMT

    That wondering has to do with the format of the game Maybe. In which case invent a new game. Don't call it cricket. Nobody has the stomach for that. Because they very much need the parent game's brand. Its history, its traditions, its skills. Without which their new game is a non-starter. [[ "" Nobody has the stomach for that."" Again, "nobody" when you mean a small section of the followers. You and I may have similar views on Test cricket but differing views on the other formats. Why should you think I want to see a new game called "Crycket". Ananth: ]] Also let's not forget that when the ODIs were introduced in the seventies, Test cricket's popularity was very high. The ODI was not a panacea that filled in the void left by this anachronistic traditional game. The new "format" essentially cannibalised the traditional game. It did not exactly bring in as many new followers as often imagined (except maybe in the subcontinent). BTW, nobody expects anybody to "sit through" 5 days of a test match. That's not how Test match economics works. A Test is like a book fair with different sets of people attending each day (each session in fact). The cumulative attendance in a MCG Boxing day test exceeds the crowds in any WC final!

  • shrikanthk on April 22, 2012, 5:01 GMT

    Alas, shrikanthk will rush to argue that item numbers are not works of art ... despite artists world over spinning money with those decades in decades out

    I would never say such a categorical thing. An Item number is not just a spectacle but an expression of personal and collective euphoria and irresistable impulses. That's true of all arts. Be it the impulse to write down a story, the impulse to put bat on ball or the impulse to kick a ball around.

    The structure of Limited overs cricket suggests that the only impulse is to bring in audiences. The impulse to play a cover drive or a square cut is subordinate to the requirement of "playing an aerial stroke to take advantage of powerplay". The impulse of a bowler to bowl a flighted leg-break is subordinate to the requirement of preventing the batsman from playing an aerial stroke. [[ Shri, this is the sweeping generalization which prevents other readers from accepting your views. How can you say the only objective is to play an aerial stroke or prevent the batsmen from playing aerial strokes. You think ODIs are aerial strokes and nothing else. I have done a special study. Out of the 3000+ matches played, 154 have been won by teams defending sub-200 totals. You think these bowlers would have been content to just deny runs and wait. This is 5%. 145 teams have chased targets below 200 runs and only won by 1/2/3/4 wickets. In other words, the matches have been close and they have not gone and smashed everything. That is another 5%. In these 10% of matches the bowlers have held sway: first by limiting the opposing team to below 200 and then either defending the low total successfully or fighting to the end. I have not even looked at other matches such as 220 and 210 or 220 and 221 for 9. What you are saying is also demeaning to the finishing skills of players like Bevan and Hussey, who made finishing the game a great treat to watch, by using the gaps, taking singles and pacing the chase. Ananth: ]]

  • shrikanthk on April 22, 2012, 4:40 GMT

    Shrikanth looks at the world through two glasses. Test and non-Test. He would not bother that the non-Test has multiple hues and he should appreciate that and not dump everything under one basket

    I only use my "cricketing glasses". And I'm pretty clear in my mind as to what sort of a sport cricket is. A lot of people seem confused about what this game is all about. Hence this support for alternative formats and constant lack of confidence in the parent game.

    Cricket is the only sport in the world which keeps questioning itself. Questioning its ethos, questioning its aesthetic, questioning its relevance. It is a game whose "followers" suffer from an inferiority complex. Hence this urge to pander to the non-followers by tinkering with rules.

    A Football fan doesn't wonder - will my game be "better" if it's played over half-an hour, with restrictions imposed on player movements. Only cricket fans think this way. Despite Cricket being one of the most popular sports in the world! [[ That wondering has to do with the format of the game. Can you afford to take a week off to watch a Test match in New Delhi. You probably cannot. You are happy at following the game through various media. Some one else, equally interested in Test cricket but unable to take that time off, says "let me watch the ODI game on Sunday". He realizes that that is also a game of skill. Someone else wants to go to watch T20. And the fourth lot, and the most boisterous and noise-making lot, watches IPL. There cannot be any comparisons between Football and Cricket when the football game, with two extra-times and a sudden-death period would finish in less than a T20 game. An England/Chelsea follower can come back from work, have a pint or two, let go a pint or two, pack a pint or two and a sandwich and THEN go to Stamford Briidge or Wembley to watch a game. Can a Test enthusiast do that. Ananth: ]]

  • shrikanthk on April 22, 2012, 4:21 GMT

    Beauty is in the eye of the beholder........ No sport can sustain over a long period of time if it does not have the substance

    Vimalan: Pornography has survived since the days of cave paintings. If sustenance alone is the true test of an art form, then pornography is the greatest of arts and prostitution the greatest of all professions!

    Gamblers & aristocrats owned the first teams in England and dictated how the game should go on --- and this was 5 day & 3 day cricket!

    But the game was not designed in a manner to "optimize" gate receipts. If that were the motivation, the game would've evolved very differently. Even in the days of Queen Anne, audiences would've liked a 20-over game more than a cricket match as we understand it.

    Also, cricket matches last for 3/4/5 days not to satisfy an idiot's whim. But because cricket is a FUNDAMENTALLY SLOW game unlike say Football or Hockey. You need that much time in cricket to figure out which is the "better side".

  • shrikanthk on April 22, 2012, 4:20 GMT

    Beauty is in the eye of the beholder........ No sport can sustain over a long period of time if it does not have the substance

    Vimalan: Pornography has survived since the days of cave paintings. If sustenance alone is the true test of an art form, then pornography is the greatest of arts and prostitution the greatest of all professions!

    Gamblers & aristocrats owned the first teams in England and dictated how the game should go on --- and this was 5 day & 3 day cricket!

    But the game was not designed in a manner to "optimize" gate receipts. If that were the motivation, the game would've evolved very differently. Even in the days of Queen Anne, audiences would've liked a 20-over game more than a cricket match as we understand it.

    Also, cricket matches last for 3/4/5 days not to satisfy an idiot's whim. But because cricket is a FUNDAMENTALLY SLOW game unlike say Football or Hockey. You need that much time in cricket to figure out which is the "better side".

  • Vimalan on April 22, 2012, 3:10 GMT

    ----- My position will soften if the greed factor lessens and the following things happen. - Maximum of 30 days and about 40 matches. - Make sure that the Gayles and Bravos of the world prioritise their duty. IPL's initial grandoise ideas on country-first have gone down the drain. IPL should think how India would feel if Sehwag missed Indian international cricket for XBL at a million dollars a season. - Tone down the hyperbole. Use the word "great" less often. Send Morrison back or ask him to quieten down. Ananth: ----------------------- @Ananth, I fully agree with your views about IPL and I share the same sentiments.

  • Alex on April 22, 2012, 3:10 GMT

    @Ananth: As someone who has studied paintings in detail, I think the conceptualization & execution level of western paintings (including former USSR) since 1900 is far higher than it used to be over 1600-1900. Some may still prefer Da Vinci, Monet, & Rembrandt but, IMO, guys like Picasso, Dali, and Kandinsky far surpass the old greats. [[ I saw a Dali or two also at Cologne. I spent 5 minutes trying to understand the painting but could not. A painting has to be a work of beauty and understood and appreciated by public and not just by the art students. When you see the positioning of the two ambassadors, the care with which the objects have been drawn, the exquisite way the floor has been painted, the symbolism with the skull, the detail to which the globe has been drawn, the size of the painting and the clarity at some distance away, one can only admire the genius at work. Similarly Turner has brought out the interaction between sun and sea. The colours used and the shades are marvellous. In fact it is my wallpaper. On the historic side, the harshly drawn tug pulling a great battleship to the breaking yard adds poignancy to the painting. But I would not question anyone who admires Dali, Kandinsky, Picasso, Matisse, Cezanne, Seurat et al. My lack of knowledge should not make me do so. Ananth: ]] Considering the bludgeoning nature of Hussey's classic 60*, if it is not to be a hard rock then it best likened to a Bollywood item number ("Chikni Chameli" or "Munni" or whatever) than a prim & proper Beatles song!! Alas, shrikanthk will rush to argue that item numbers are not works of art ... despite artists world over spinning money with those decades in decades out.

    I daresay the most popular form cricket 150 yrs down the line will be quite unlike anything played today, the only common factors being bat & ball (which might also undergo some structural modifications).

  • Alex on April 21, 2012, 23:48 GMT

    @shrikanthk: Cricket is not an exhibition art like paintings but a *sport* where two teams compete to win. It is not an *entertainment* like drama where everyone on stage collaborates. In both cases, the objective is to have fun & put on a show for which the audience will pay money. Even penniless galli cricket matches of kids are about having fun & winning a match.

    As for T20 being a market research outcome, pl look up cricket's history in 17th & 18th century. Organized cricket emerged because gamblers felt that the best way to money by betting was to own teams. Gamblers & aristocrats owned the first teams in England and dictated how the game should go on --- and this was 5 day & 3 day cricket! So, the precedent for Packer's ODI's and IPL was laid well over 250 yrs back.

    Finally, what is meant by art is always subjective. IMO, Hussey's 60 off 24 was art but more like Guns N Roses' "Sweet Child" rather than a Bach concerto. [[ When we had one day last year at London, I chose British Museum and National Gallery. I could stand hours in front of classics like Holbein's "Ambassadors", Turner's "Fighting Temeraire", Claude's "Seaport", Botticelli's "Venus and Mars" and so on. I did not select Tate Modern because I do not understand and appreciate paintings of the past 100 years. In Cologne we rushed through the Museum and had a look at couple of Picasso's paintings and it justified my decision. But can I say that the modern paintings are no good and that only the pre-1900 painters should be exhibited. No. There is an information gap from my side and that should not come in the way. Since I am not a "Guns N Roses" guy and my favourite group is Beatles, I would liken Hussey's 60 to Sgt Pepper. Ananth: ]]

  • Anshu N Jain on April 21, 2012, 18:46 GMT

    Yes Ananth, what I should have said is "the closer to 1.0, the more the consistency". It must have been at the back of my mind that (HQ-Median) is greater than (Median-LQ) for all batsmen that you posted values for. [[ Milind What do you think of this simple factor. Works to Lara: 1.56 SRT: 1.66 Ponting: 1.8 Dravid: 1.51 Border: 1.41 Bradman: 1.9 Greig: 1.38 Gambhir: 2.11 !!! Ananth: ]]

  • Vimalan on April 21, 2012, 18:43 GMT

    @shrikanthk

    Beauty is in the eye of the beholder...If the spectator who is one of the main stakeholders of the game, like competitive T20 matches then it'll become an art form of its own after some years of refinement. It may be 10 years or 30 years. No sport can sustain over a long period of time if it does not have the substance. Same will be the case of ODIs or T20s. [[ Vimalan, I am one of the crazy guys who watches the Test matches chugging along at 2 RpO while the guys are belting sixes on the other channel as if there is no morrow. However I take care to distinguish between the T20-Intls and the greed-based, glitzy, IPL matches which pass off as Cricket. Harsha Bhogle wrote an entire artickle recently extolling the virtues of IPL without referring to the word IPL once. He used the term T20 to suggest respectability for IPL. I am not like that I clearly differentiate between the two. Shrikanth looks at the world through two glasses. Test and non-Test. He would not bother that the non-Test has multiple hues and he should appreciate that and not dump everything under one basket. However I will never doubt that IPL is here to stay and has a place. It caters to the family, runs for a long-movie-lenth, is marketed in a colourful, if obnoxious, manner and even at TRPs of 3% is a potent force: as highly commercilized formula films are. My position will soften if the greed factor lessens and the following things happen. - Maximum of 30 days and about 40 matches. - Make sure that the Gayles and Bravos of the world prioritise their duty. IPL's initial grandoise ideas on country-first have gone down the drain. IPL should think how India would feel if Sehwag missed Indian international cricket for XBL at a million dollars a season. - Tone down the hyperbole. Use the word "great" less often. Send Morrison back or ask him to quieten down. Ananth: ]]

  • shrikanthk on April 21, 2012, 17:45 GMT

    Now, after 3000 matches, the game has certainly evolved and has a place of its own

    So does a long history alone legitimize an art form? 30 years from now, will you regard the T20 as a separate art form, just because it has accumulated enough sentimental value over the years?

    Cricket (and by that I mean the traditional game) is a legitimate art form. Not because it has been played over the past 200 years. But because like all great art forms, it evolved on its own shaped by the genius of great artists (ranging from the early Hambledon Masters to Muttiah Muralitharan). Same is true for other genuine art forms. For eg : Rock n roll, Jazz or for that matter Carnatic music. These arts were not the outcome of market research or business plans like the T20s and One-dayers.

    T20 is the only sport that I am aware of which is an outcome of market research, as opposed to gradual piece-meal evolution. It may be a legitimate sport. Yes. But not a legitimate art form.

  • shrikanthk on April 21, 2012, 15:56 GMT

    Putting down one format over the other is like belittling the achievement..in GRE or CAT..over IIT-JEE..

    Cannot resist replying to this. Ananth - I think this debate won't hurt because it is fundamentally important.

    Nobody is putting down achievements in one "format". Limited overs cricket isn't exactly a distinct art form like say Baseball that has evolved independently from cricket. The One-dayer/T20 owe their very existence to "Cricket" by which I mean the game played with a red ball where a team wins only when it bowls a side out twice!

    These "new formats" were devised simply to attract people who don't like cricket! So to say that Tests are akin to JEE while ODIs to CAT is to miss the point. CAT and JEE are two totally different exams testing different skills. The one-dayer was not devised to "test different set of skills". It was devised to bolster county club finances. [[ My one concern, Shri, is that you keep on repeating thiis particular statement. It might have been true at one point in time. Now, after 3000 matches, the game has certainly evolved and has a place of its own. Let us accept that it is over-killed normally with too many inconsequential bi-lateral matches. However many a match has been played in which sub-par totals have been defended with skill or middling totals chaed, again with skill. And I would never agree that the 189*, 175*, 149, Gower's 158, Styris' 141, Jayasuriya's 141, Tendulkar's 134, O'Brien's 113 et al are not innings comparable to the best Test innings and will stand the test of time. We have not even talked of bowling yet. Gilmour's 6 for 14, Bichel's 7 for 20, Vaas's 8 for 19, Bond's 6 for 23 and so on. Anyhow today Country club finances are all supported by various 20-20 tournaments, not ODIs. Ananth: ]] BTW I'm someone who cracked CAT but flunked JEE. Trust me, cracking CAT is not an "achievement".

  • SWARZI on April 21, 2012, 15:40 GMT

    TO ANANTH, I am not disputing the enjoyment that 'all of us' get out of comparison exercises regarding our heroes. All I am trying to say to those fans who are seeking to have you use every statistical methodology until you come up with with one that gives the player the advantage, would not necessarily mean their player is better than any other player. Comparisons are fairest, or best,or more truthful, or most accurate when two players are tested via the same standards/conditions - a scientific truth. ANAND, There is no putting down of any format under the other here! I don't think you would dispute the fact that the traditional format of cricket which is designed to be played in the most grueling of sporting conditions cultivates many more men of deep thought, unyielding tenacity and hard work than any of the shortened formats. The shortened formats inculcates a certain level of laziness in certain players - especially by some of our best fast bowlers who tke advantage and sit down and wait until the BIG BASH tournaments come around to go out and play and collect big money for their little work! I think that this attitude is a disgrace to the founding fathers of cricket! I imagine it makes them really uncomfortable in their graves!

  • Anand on April 21, 2012, 13:45 GMT

    Putting down one format over the other (particularly ODIs over tests) is like belittling the achievement of a student in GRE or CAT or GMAT over a performance in the IIT-JEE format of the 90s just because GRE and CAT were in the objective format. Saying that ODIs dont require technique is like saying it is easy to perform in objective type exams because one can just keep guessing the answers. I agree that lack of slips, field restrictions etc make it conducive for some run scoring in ODIs but one cannot write off a format for that reason. Can anyone deny Bevan's skills just because he doest have similar numbers in tests? Truly great players adapt to the different formats. I only have a complain against T20 because my opinion is 20 overs is too small a sample size to determine anything. The 50 over format allows for initial onslaughts, planned middle overs and a final slog. T20 misses those planned middle overs which reduce the beauty. But surely I wont dismiss any T20 achievements [[ Since most of the other points have been covered I will cover the last point. I will look only at T20-Intl performances. I do IPL work because it is part of a contract I have. Otherwise, no. In T20-Intls, I would place Mike Hussey's WC Semi Final innings of 60 in 24 as a truly wonderful innings. Likewise Mendis' spell of 6 for 19, all quality batsmen. But these are few and far between. Ananth: ]]

  • SWARZI on April 21, 2012, 13:43 GMT

    Ananth, the intellectual exercises that your blogs stimulate among your readers are outstanding! I don't envy your resolve to please all of us! However, I think that it is has become necessary to remind your contributors that, based on the way that you carried out the statistical analysis for this current topic, the findings that you’ve posted would only help to rank the various batsmen in terms of their consistency over the periods of their entire careers. But these indicators would not ‘NECESSARILY’ determine whether or not a player is better than the other! The reason being that when genuine comparisons are ever to be made between two individuals or two scientific scenarios, etc., the most reliable way to make any factual avowal regarding the superiority or inferiority of one against the other is to always try to assess their performances under the same/similar conditions. Hence, if one wants to know ‘GENUINELY’ who is the most consistent of two batsmen, the authentic answer can only be arrived at by comparing their performances within a particular period (long enough) when they played against similar bowlers; on similar pitches, via the same rules; in similar batting conditions; etc. It is noticeable that as a result of this premise, how difficult it would be to determine GENUINELY, who is more consistent in a head to head matchup between Hutton and Pontin! Hence, the alternative use of all these other statistical methodologies that are being suggested by all these specific player fans, who are trying to find a method which when employed would give ‘their man’ the advantage would still not give an accurate result, unless the same/similar conditions are put into the equation of assessment for every two players being compared. [[ But comparisons cannot be avoided. Pele and Messi are going to be compared, come what may. Similarly Federer, Sampras and Laver. Nicklaus and Woods (if he revives his winning habit), Owens and Lewis. Kasporov and Anand. One could go on. The bottomline is that all of us have fun, without calling each other names. Ananth: ]]

  • Anand on April 21, 2012, 13:35 GMT

    What I am trying to drive is, if a stat about the number of times a bowler has dismissed a batsman comes up we can use that to see insights about which bowler a batsman was more vulnerable against but cant use that to compare two batsmen. Sachin's or Lara's performance against various teams only shows how they fared at different points in time (slices of innings) and space (venues and oppositions). To make it a basis for comparing them is just plain ridiculous. Of course people have their choice of favorite player. To me, my favorite batsman in the Indian team was and is Kris Srikkanth. But that is my personal choice. But in a blogspace like this which brings about the beauty of the game and its players through numbers, it is counter productive to use them for projecting personal choices. Another point is about tests vs odis. I always look at tests as descriptive exams and odis as objective ones. A good student will adapt to both. One cannot belittle one format over the other. (contd)

  • Anand on April 21, 2012, 13:23 GMT

    Ananth:

    You come up with gems do often. Your analysis and more importantly as a big data analyst myself, am fascinated by your techniques and tools. I love the game and to me any analysis that brings out finer details of the game (like yours do) are always great. I also am disappointed by the Lara vs SAchin debate. Its an extreme fortune to be born in an era where we can watch these two legends play. Seriously it is an insult to both the players when we compare them. I would only read these numbers as indications of few finer facts about a player but never in the light of who is better than whom. I mean statistically Glenn McGrath has probably dismissed Lara more often than he did say Venkatesh Prasad. But honestly who do you think McGrath will prefer to bowl to? Lara or Prasad? Wont it be foolosh to say Prasad handles McGrath better than Lara based on these numbers? I am sure no one made such a comment here. (continued)

  • Anshu N Jain on April 21, 2012, 13:08 GMT

    My last word on your comparison of our 2 examples. Your example comes up with a different Median value, mine doesnt. So strictly speaking, the outlandishness of the 2 examples is not comparable :-) (all in good fun!) [[ No, I deliberately decided on the distribution to show that the median value can be so varying. Ananth: ]] I'll repeat that mixed use of boundary values (Percentiles) and aggregates (RpI) does not seem intuitively correct to me to account for varied distributions. Which is what I intended to convey through my example, in REJECTING my original suggestion of using RpI/Median. I did not intend to run down any methodology.

    I'll also suggest that if you are really keen on using Boundary values to derive consistency, you can use them without the RpI. Simply use the ratio (HQ-Median)/(Median-LQ); the lower the value, the higher the consistency. This ratio conveys the range within which 50% of the scores lie, and their spread on either side of the Median. The tighter the better. [[ I will take couple of days off this fascinating topic and then do some work on all these ideas suggested, including your alternate suggetsion made in the earlier comment. "the lower the value, the higher the consistency" ??? In your last example, should it not be that the closest the value is to 1.o, the more consistent the player is. After all if HQ-median is equal to Median-LQ, it indicates a very balanced and hence consistent situation. HQ=60, Median=50, LQ=10 leads to an index of 0.25 and this distribution does not indicate consitency. Ananth: ]]

  • Prashanth on April 21, 2012, 9:55 GMT

    3)But what if the "expected" runs a batsman is supposed to put up in a match is just a bit fewer? Say 80 instead of 90? So, a 1/0 method is slightly impractical . 4)Can we then "grade" the batsman's effort per match? If a batsman scored within 80-90% on either side of what is expected to he should be deemed to be pretty consistent. 5)As I type I realise the complexity of this seemingly simple topic! 6)Again, the way I see it the concept of a "slice" negates the original premise; Which was to see how many times a batsman has let his team down,so to speak vs. the number of times he has performed at or above expectations. The above are just a few thoughts. I will have to brush up on my statistical knowledge to catch up with the likes of Milpand and Anshujain who are currently on a different wawelength! [[ Thanks, Prashanth. Unfortunately you have come in nearly 300 comments and two articles ater the first ome was published. It is impossible for us to go back to your original ideas. Pl go through the comments and get the drift. Once again many thanks for the spark. Ananth: ]]

  • Prashanth on April 21, 2012, 9:49 GMT

    Hello Ananth, Checked in again after hearing about this from some colleagues. Surprising how soon you latched onto this. I thought it would take a while. Also it has become exceedingly complex !

    The origin was relatively simple.I was simply wondering whether the odd so-called matchwinning innings can make up for other failures. Also,that from a purely cricketing standpoint a 70 in each inn. of a match may be equivalent to a nail-biting 140 in the last inn.

    As mentioned some of the hi-fi statistical jargon being thrown around is a bit beyond me. The way I see it is - Did the batsman perform as per what was expected from him in a match ? or no?

    So,the ways to go about it are: 1)A simple 1/0. Performed above or below expectations (per match). The exact expectation (median,avg.,RPI,etc) to use is unclear. 2)If a batsman performs below expectation s in say 3 matches - he has let his team down and a great hundred in the 4th match may not nullify the damage already done. TBC-

  • Anshu N Jain on April 21, 2012, 8:45 GMT

    I'm surprised that my comment came across as negating a methodology. Nowhere have I rejected Milind's or your suggestions. In my view, any formula needs to be tested for consistency (no pun intended) and fidelity. Unreserved apologies nonetheless. I started with suggesting the RpI/Median as a possible measure, but then realised that combining an aggregate value with a boundary value may not be ideal, given the nature of distribution. If you recall, my final suggestion was using (RpI of scores above Median)/(RpI of scores below Median). Your example is way too extreme. Maybe a few more ranges of scores will help. Even in my simulated distribution, I have arrived at the SAME exact boundary values for Lara, including the Median. If you apply my suggested formula to your example, it gives the same result in both situations. [[ Anshu, I agree that my example is extreme. It is ridiculous to think of scores being only 0 and 100. Unfortunately your example is equally so. Assuming no occurences of 10-25, 41-65, 100-299 is totally outside the possiblity of occurence. Ananth: ]] Also, I am kind of confused with your switching from pure statistics (reply to Alex's comment)to only cricket practicalities. Apologies again for not getting it. [[ Yes, I agree that I had talked about a statistical solution since I was tired of getting 100s/inns, not outs % etc. However we should also keep in mind the normal expected distribution. That is where I brought in the cricketing consideration because statistically anything can occur but not necessarily in the real game. The moral is none of us should use a way-out unlikely example. Coming back to the point in question. Let us take the simple but effective fraction suggested by you: RpI(Mean)/Median. Values close to 1.00 clearly indicates a degree of consistency. The ideal situation is that the RpI is 40 and Median 40. 1.00 is the index. As many entries above mean as below mean. The congruence of the two values represents a very balanced distribution. Now take two cases. In the first case HQ is 70, Mean=50, Median=40 and LQ is 10. This indicates that the 75 percentile is as far away from the median as the 25 percentile. If we ignore above 75% and below 25% values, this indicates perfect consistency. Your index of 1.25 will remain the same. In the second case HQ is 70, Mean=50, Median=40 and LQ is 20. This indicates that the 75 percentile is 75% away from the median while the 25 percentile is 50% away. There is a skew and your factor of 1.25 (or whatever) should be adjusted probably by (70/40)/(40/20). Or by 70*20/40*40. Or something else similar. This is the essence of Milind suggestion. Let me try out the other suggestion also. Anyhow I think we are certainly on track to get to something looking at the individual scores. That is one excellent option. The other can always be the rolling slice method of x Tests (either Runs or 50s or both) and something based on that. My apologies simce I know the effort both you and Milind have put on this. Ananth: ]]

  • Anshu N Jain on April 21, 2012, 7:32 GMT

    @Milind, @Ananth I am not convinced about using the Boundary values (which these 5 are) because they mask the magnitude of scores within any two boundaries. I did a random scenario testing on Lara's scores (232 inn). It is possible to have the same boundary values and career RpI with starkly different distributions. e.g. here are two scenarios (Actual and Simulated) with the same Boundary values, same RpI, but different distributions (sadly, cant list individual inning scores: Score Actual Simulation Range Innings Innings 0-9 62 115 10-25 40 0 26-40 28 58 41-65 40 0 66-99 28 41 100-199 25 0 200-299 7 0 300-399 1 11 400+ 1 7 One has scores evenly spread out across Min and Max. The other has pockets of score densities. Which of these is the more consistent? What is the statistical interpretation of this? [[ Anshu, a few things. First I myself have mentioned that the two extreme values, High and Low are useless and will not even come into the picture. So we are considering only three values + Median. Milind has suggested a way of tweaking your base value of RpI/Median with a number based on the two Quartile values. I think there is considerable value in that suggestion. I have suggested an alternative to keep this tweaking factor to near 1. There is no point in bringing a completely impractical and theoretical distribution to negate some methodology or other. We should expect normal distributions and there is nothing to suggest even one such distribution, now or in 100 years. Look at what I can do. Player A scores 49 zeros and 50 hundreds in his career. His Median is 100. Player B has 50 zeroes and 49 hundreds. His Median is 0. The RpI(Mean) value is almost the same, around 50. Then RpI/Median will go for a hit. However the comforting thought is that such a distribution WILL NEVER HAPPEN. Similarly your simulated alternate distribution will also NEVER HAPPEN. Let us now resort to Cricketing practicalities. The actual Lara distribution is the one which will happen 99.9999% of the times. There may be minor variations. That is all. In which case, can the HQ and LQ be used to strengthen your base value of RpI/Median and get a better handle on consistency. Ananth: ]]

  • milpand on April 20, 2012, 22:00 GMT

    9th and 91st percentile as whiskers makes a lot of sense.

    Batting scores are skewed distributions therefore Standard Deviation does not get used in analysing the dispersion in batting scores. First and third quartile give some sense of symmetry in these skewed distribution.

    Box plots are used to indicate degree of dispersion and skewness (& outliers).

    Combining the 4 useful terms of Median, RpI, HQ and LQ in (HQ/MEDIAN) * (RpI/MEDIAN) * (MEDIAN/LQ) we get: Border: 9.38 Dravid: 9.58 Ponting: 10.43 Tendulkar: 10.82 Lara: 14.05

    This hardly constitutes to be a decent measure but I hope that it helps in developing a better alternative. [[ As you yourself have mentioned, this can be reworked as (RpI/Median)*(HQ/LQ). This is Anshu's base vaue multiplied by a factor which reflects how the HQ and LQ have been dispersed. Anshu: I would appreciate your specific comments on these. Milind: Just a googly. What about the addl multiplying factor as (HQ+LQ)/2*Median. Or something which reflects the dispersion of HQ and LQ around the median. Then this will become a multiplying factor around 1. Ananth: ]]

  • Alex on April 20, 2012, 17:48 GMT

    @Ananth: Fine, thanks. The key is to look at the moving average taken over 5-6 innings. One more illustration:

    IMO, someone like Lara should be considered less consistent than, say, SRT and Chappell since, statistically, Lara of 1990-2000 was vastly different from Lara of 2001-06. Till 2001, he played 125 innings and crossed 50 in only 27% innings (incl. 14 hundreds). After that, he exceeded 50 in 37% innings (and his conversion of 50's into 100's jumped from 48% to 104%!!). This came at the cost of a big dip in ODI's where his % of 50+ scores decreased from an excellent 33 to a poor 22 (albeit his avg had a much small decrease: from 42 to 38). In contrast, Chappell performed almost like a clockwork.

    So, I think the time-evolution of how frequently a batsman scored 50+ on RpI (or 70+ on RpT), or exceeded his own median, is a good indicator of consistency ... just take std deviation of this series to quantify it.

  • Alex on April 20, 2012, 16:16 GMT

    @Ananth & @Anshu: Fine, thanks. In that case, pl consider the following modification of Anshu's scheme.

    1. Compute average RPI for 6-innings slices (or 3-test slices) ... do this on a _moving average_ basis. For example, one number for innings 5-10, another for 6-11, and another for 7-12, etc. The # elements in this series equals the # innings of the batsman. Let's call it series A. Plot its violin plot for the batsmen of interest.

    2. For series A, compute the ratio of Anshu's two metrics.

    3. IMO, standard deviation of series A will give a good idea of the consistency ... this is a commonly used approach in engineering (e.g., for noise cancellation, disturbance rejection, etc.). [[ Thanks, Alex. I will look into it. I understand that the High and Low values are almost always useless. The first is some high figure conveying very little and the second one will almost always be 0. Hence what may be more meaningful are the 9th and 91st percentiles. I will do that and post. Let us see. Ananth: ]]

  • Alex on April 20, 2012, 14:01 GMT

    @Ananth and @milpand: Box plots is probably not the best solution here. I suggest you use "violin plots" (see http://www.r-bloggers.com/boxplots-and-beyond-iii-violin-plots/ and http://www.jstor.org/stable/2685478). Anyway, Ananth's # for the 5 batsmen show:

    1. SRT & Lara scored most in their top 25 percentile innings; Border scored least (400*, 375, & 277 bloated this up for Lara ... else, he probably averages the same as Ponting there).

    2. Lara scored the least in his bottom 25 percentile innings.

    LQ and Low don't give any added information ... Low=0 for almost everyone and most top batsmen will have LQ in the range of 8 to 12. Median is a pretty useless statistical property.

    I suggest Ananth replace these 3 factors with (1) innings (or completed innings) per 100, (2) innings (or completed innings) per 35+ score, (3) innings (or completed innings) per 60+ score since a 35+ score is a moderate success, 60+ score is significant, and 100+ is big score. [[ Alex, for once you are off the mark. I do not want any cricketing measures such as innings per 100, not outs and the like. I have millions of such numbers floating everywhere in my space. Anshu had made an excellent suggestion using two of these numbers. I want a purely statistical suggestion of how to use these numbers to define consistency or otherwise. You have to think without bringing in cricket. Ananth: ]]

  • Boll on April 20, 2012, 13:47 GMT

    @Ananth - just to check, is HQ the RPI of top 25% innings regardless of not outs, LQ similarly? or average of top/bottom 25%/

    On first perusal, I haven`t got much further than the thought that `Everyone gets a bloody duck sometime or other.` [[ Boll, these are numbers. Nothing cricket-related. An innings is a batting effort and runs are runs scored. A "not out" is a cricketing concoction to derive another cricketing concoction, the average. I want non-cricketing numbers based derivations. Ananth: ]]

  • Ananth on April 20, 2012, 13:21 GMT

    In continuation
    What do these figures tell us.
    Lara:      RpI=51.5 High=400 HQ=72 Median=33 LQ= 8 Low=0
    Tendulkar: RpI=49.7 High=248 HQ=74 Median=34 LQ=10 Low=0
    Ponting:   RpI=47.4 High=257 HQ=66 Median=30 LQ=10 Low=0
    Dravid:    RpI=46.5 High=270 HQ=68 Median=33 LQ=10 Low=0
    Border:    RpI=42.2 High=205 HQ=62 Median=31 LQ= 9 Low=0
    

  • Sudarshan on April 20, 2012, 12:55 GMT

    How does this analysis affect players like Imran and Sangakkara who had distinctly different phases and roles. Imran played many games purely as a batsman and his powers increased towards the end of his career. Similarly how does one classify catches held by Sangakkara. Or the variation in bowling strike rate and average of Sobers by the type of bowling he was doing. [[ This looks at players in their batting roles. Why bring in bowling, catching, wicket-keeping, bowling strike rate, type of bowling etc. Ananth: ]]

  • Ananth on April 20, 2012, 12:07 GMT

    Readers might have seen Milind's comment on the Box-whisker style graphs (also called Box plots). The graphs look great. What attracted me was the use of 5 numbers, maximum, minimum, median and the upper and lower quartile values. Can we think of using these 5 (or less) numbers to come to a single number to indicate presence or absence of consistency. I am looking at extending Anshu's idea of Mean(RpI)/Median. How do we integrate the two quartile values into this formula. The graphs can be used to support the conclusions. Ananth

  • Boll on April 20, 2012, 12:02 GMT

    No mud being slung from sunny Kumamoto Ananth, just the last remaining cherry blossoms!

    Apologies to (the new and more strictly traditionalist) shrikanthk, for taking things off track here, but I do feel as if there may be a strong positive correlation between the ratio of comments to the batsman/bowling consistency articles, and the number of MoM awards given to batsmen and bowlers - 174/60, circa 3/1...sounds about right.

    Apparently the toil is its own reward however, leather slingers. [[ I am with you all the way. This malaise has spread to IPL also. Amongst other negative points. Levi's 50 gets instead of Ojha's 2 for 17 (both top wkts). Pathan's 42 gets instead of Morkel's 3 for 18 (all top wkts). Smith's 39 gets instead of Dinda's 4 for 18 (the worst case so far). du Plessis' 71 gets instead of Morkel's 28 in 7 and 2 for 37. Ananth: ]]

  • MSingh on April 20, 2012, 11:05 GMT

    Jusst now saw that the message got truncated. Here is teh continuation...

    Since we are looking at the contribution of a player, we need to be careful in leaving out the innings that he never had an opportunity to bat. So, these will be a) innings when he was never needed to bat and b) innings when he remained not out when his individual score was less than, say 20(Meaning that he may have had an opportunity to have contribute more. Guess when we change this number different results may come). It is possible there are some other innings that should not be included, which am not able to think of.

    Would it be possible to get results for this query? Thanks in advance [[ Long time back I had done an analysis called Extended Batting average where I allowed a not out innings to continue. I added the RpI value to the undefeated innings since it seemed natural. You could say, why not just add the Not outs x RpI to the run tally and divide by the number of innings. However care had to be taken that the extension can only be done when there are other batsmen around to support and the match was not already ended. One day I will revisit that analysis. One thing is certain. At least in that analysis the reader set who are on a permanent mission to sling mud at me cannot do that since Lara would benefit very little. He has only 2 qualifying not outs. Ananth: ]]

  • Sarosh on April 20, 2012, 10:42 GMT

    If a batsman performs above par this is a positive - and should be discounted from calculating consistency.

    4) As an aside –for batsmen with more than 3000 runs in ODIs. Openers - 12 batsmen avg. above 40. Batsmen posn 3-4 – 12 batsmen avg. above 40.

    Depending on one’s point of view one would think that the white ball is most dangerous when new. Perhaps the easiest place to bat in ODIs would be the middle order- especially for players like Lara more at ease against spin and medium pace.

  • Sarosh on April 20, 2012, 10:41 GMT

    FYI -Lara – 1.77 inns. per Test. (The least among batsmen with more than 10,000 runs being Steve Waugh at 1.55/Test – over 131 matches the difference works out to a whopping 29 inn.- about 1,500 runs at Waughs avg ). Lara also had just 8 DNBs. Out of Lara’s 6 N.O – just 3 are low scores for this analysis- 13,14,48 (below career avg) .Lara truly had the luxury of completing almost all his innings and fulfilling his potential to the max. 3) We must only look at scores below par. Inconsistency has a negative connotation. The very essence of skill is consistency .Many sportsmen can produce the goods once in every ten attempts or so. If a Federer did that he would hardly ever reach any Finals. A batsman has the relative luxury of failing as long as the rest of the team pulls along. If they don’t the main batsman may often have to take the rap -A fate that has often befallen Tendulkar- but curiously not Lara.

    Contd....

  • Sarosh on April 20, 2012, 10:40 GMT

    This is clearly ridiculous and defeats the very purpose of the analysis itself. Clearly this spikiness may manifest itself differently depending on the number of innings used as a slice. Using an arbitrary slice of 10 inns. may result in a single big inn. in a sequence making the entire 10 inn. slice appear consistent. Instead of using multiple methods but a fixed and same inn. slice- the correct method to determine spikiness or inconsistency is to see if the results hold for different inn. slices. If you obtain similar results ( or average out )for say 2,3,4,5…..10 inn.- only then may a particular level of consistency or inconsistency be attached to a batsman. This simple (or simplistic method as you may want to call it) would cover most areas.

    2) Re. the RPT issue- the simple average is still the best measure available. For low N.O scores (less than the career batting avg. for this particular analysis) we can simply ignore them in order to calculate consistency. Contd....

  • Sarosh on April 20, 2012, 10:39 GMT

    One sometimes wonders whether your obsession with Lara reduces your objectivity and blinds you to flaws in your reasoning and analyses which should normally be obvious. You must be the only cricket follower on the planet, past or present, who without some peculiar bias can make statements such as Tendulkar has had a “very good” Test career. Apparently Bradman and Lara are the only batsmen with great Test careers. To paraphrase you – one doesn’t know whether to laugh or to cry. You may as well be talking about Yousuf, Langer, Kirsten...etc.etc…rather than Tendulkar.Truly amazing statements which beggar belief.

    1) Re. this particular analysis the idea of consistency should be to deduce “spikiness”. A basic, hypothetical and overly simplistic e.g. for a batsman with an avg of 50- 50, 50, 50, 50….is consistent. 0, 0, 0,200 …is inconsistent Even with a 4 inn slice the above sequences would show up as equally consistent. A similar trend may be carried on to 8 inns. etc. Contd....

  • MSingh on April 20, 2012, 9:18 GMT

    Hi Ananth,

    This is an excellent article. Thanks for the meticulous work.

    I am thinking of a different way to rate batsmen. We all know that the average gives us only half the picture. So, below is an attempt to find out the most contributive/Valuable Player in a different way. I am not sure if something similar is already done before. Please share the details if it was done. Below I have tried to explain the way

    We all know a 60 in a team score of say 126 has more value than a 100 when the team scores in excess of 600. So, this is an attempt to find out what is the contribution/value of a player in his career.

    In this method, if a player scores 60 out of the total score of 120, in his very first innings, his contribution score is 50%. And in the second innings(same or the next test), if he scores 30 out of team score of 300, his score for the innings is 10%, and his career contribution becomes (50+10/2) = 30%.

    Since we are looking at the contribution of a player, we n

  • m.waqas latif on April 20, 2012, 7:29 GMT

    lara is great among all irrespect of all blames of inconsistency on him he was certainly more consistent than any other his era fellow stars he is far better even than sachin

  • Ranga on April 20, 2012, 4:42 GMT

    I like Alex's comment of including both SRT & Lara in an all time XI!! Of course, I am never a fan of an all time XI concept but still, it is some way of recollecting and ranking players with peers (batting positions). An Indian flab in cricinfo might have resulted in SRT over Lara in the all-time XI (though I see the balance in the gentlemen from the jury, but the audience is more Indian than any other). However, it is only when you see SRT (OR) Lara we get this dichotomy. Why not have SRT (AND) Lara?

    And an all time XI cant span across generations, though it is a theoretical concept. I would only have an all time XI in an era. So it would not be an alltime XI but an XI of the "Era's best".

    1 Rupee(or $) in 1945 does not have the same value of 1 rupee in 2005, how can we equate 1 run in 1945 to 1 run in 2005 (though I am talking of the same "Currency" unit and same quantity, the value changes). I cant & shouldnt compare my salary to my grandpa's even if they are numerically equal

  • shrikanthk on April 20, 2012, 2:54 GMT

    I am ready to accept that T20s are ouside the ambit of traditional cricket in that 5 for 40 or 50 in 40 would be a failure

    Swarzi, Ananth: Let's not use words like "purest", "greatest" while comparing Tests with ODIs. The words may be warranted, but it weakens our case. It makes us sound like self-righteous old fogeys. This kind of language doesn't help us win the white-ball enthusiasts over to our point of view.

    All we are asking for is intellectual integrity from all fans. Nobody is saying 5-40 by Murali in a T20 is a failure. But you wouldn't use this to bolster Murali's claims as a cricketer (as here's someone who has taken countless 5-wicket hauls in tests).

    BTW, who decides what is cricket and what is not? If the MCC tomorrow introduces a T10 competition to sustain their creaky county FC circuit, would we regard it as cricket? If everything can be cricket, then nothing is cricket!!!!

    Cricket is the only game in the world which faces this peculiar identity crisis!

  • A. Khan on April 20, 2012, 2:28 GMT

    @Ananth I don't know if this has been brought up or not. The starting point issue can be tackled for each player individually. We have 10 potential starting points, check the one which gives best s/d, use that starting point for that player, the first slice can be handled like the last slice. The issue of not outs is definitely a tricky one and any solution will have pros and cons. Same is the case with slice interval, to some 10 may seem a bit more, to some it would appear otherwise. I would say, if possible, we can do this for 6, 8, 10 and may be 12 innings and lets see what changes come out! [[ While I agree with the idea, there is a lot of work. I am not saying it cannot be done but 10 runs across for each player, determining the starting point, handling the either-end slices: But possibly the REAL solution. Let me see. I at least i now have all innings available within the Player database. Ananth: ]]

  • A. Khan on April 20, 2012, 1:51 GMT

    @Swarzi Such a long and eloquent commentary. I agree with most part except for that holding records doesn't make one better than other, its consistent significant contribution to the team. Also I didn't like your misleading facts about Sachin not scoring any century when not opening. He did score 4 hundreds when not opening. On top of it he has scored 21 half centuries. In total he has scored 3116 runs with an average of 33.14 (30.87 if you exclude minnows) and strike rate of 78.29 (75.59 on excl. minnows). About 66 of those innings were before he actually started opening the innings. In the 46 innings after that when he batted in the middle order he averaged 36 (31 when you exclude minnows). One interesting stat is that in the matches in which Sachin opened, India has win/loss ratio of 0.92 (132/142) against 1 (49/49) when he did not open the innings (both excluding minnows). No difference at all!

  • Alex on April 19, 2012, 23:28 GMT

    @Swarzi: I often feel I should have done PhD in statistics instead. In addition, SRT & Lara are my weak points: I know one and am a very big fan of the other. Anyway, besides the records for most # runs, #centuries, #50's, which are partly because of playing so much, SRT holds the record for most # runs scored without getting dismissed: 496 (241*, 60*, 194*, 1). These were scored away and vs Oz & Pak ... fairly impressive, I think.

    The 1998 injury was a major setback for him. Also, barring the injury-ridden 2004-06 phase and the last year, SRT failing in 3 innings on trot was a very rare phenomenon. He might very well be the most economical batsman ever in terms of the body movement used to play a shot: I personally think that is the true test of an art. Several others certainly surpass him on top 5 innings or top 5 series etc. but there is a very strong case to view him as the best, apart from the Don, in tests.

  • Boll on April 19, 2012, 23:17 GMT

    @Ananth - forgive me if I`m being obtuse, but even for a top-order player such as Bradman, if we look at his first 5 slices of 5 tests - played 10 innings, 6 innings, 6 innings, 6 innings and then 10 innings. Isn`t there an in-built inconsistency here which offsets any advantages of using RPT as the standard? [[ I can bet there would have been quite a few innings wins strewn amongst these AND Bradman would have scored over 600 runs in most of these slices. Ananth: ]]

  • milpand on April 19, 2012, 21:17 GMT

    Box whisker plots are used to graphically depict six number summaries. Lowest and highest score are the two whiskers with median appearing as central data point. Then we depict lower quartile & upper quartile within a box and RpI/Average superimposed with an additional character. [[ Let me check it out and try. I love BCG charts and have used in the next article. Ananth: ]] http://en.wikipedia.org/wiki/Box_plot : "The spacings between the different parts of the box help indicate the degree of dispersion (spread) and skewness in the data, and identify outliers. Boxplots can be drawn either horizontally or vertically." http://support.microsoft.com/kb/155130 : Microsoft support page for plotting such charts in Excel.

    Data tables: Colour code a data point and use it as a hyperlink for additional information. If players appear in a row and calendar year are columns, then RpI for each year can be blue/red to indicate above/below par performance and underlying data for that year is provided through hyperlink to another part on the webpage or a downloadable file. [[ Only problem is that I don't have complete control over the web pages. I send my Html file to Cricinfo and they publish. Ananth: ]]

  • Alex on April 19, 2012, 20:18 GMT

    @SWARZI and @Ananth:

    1. SRT is better than Lara on freq. of 100's (0.164 vs 0.146) and on freq. of 50+ scores (.373 vs .353). Lara was quite deficient on these vital metrics in his first 10 years.

    2. SRT did score 100s in middle order: 4 in 61 ODI's at #4 with ave=39 & SR=77 (Lara: 14 in 190 at #3 or #4 with ave=40 & SR=81). In more than 80% of the 97 ODIs he batted in at #4 or #5, SRT was either less than 21 yrs old or in the injury plagued '05-'07 phase. Don't compare a baby SRT with Lara in prime!!

    3. Who prevented Lara from opening? He was a great opener with five 100's in 52 ODI's at ave=46 & SR=75. And after the '92 WC, it was clear that the opening slot was best suited for max damage. Still, like Viv, he chose to bat at #3 or #4.

    4. Since Viv was the best ever vs fast bowling, my XI takes him as an opener so that Lara & SRT can both walk in: Hobbs/Hutton, Viv, Lara, Don, SRT, Sobers, Gilly, Marshall, Warne, McGrath, Imran/Ambrose --- 12th: Ponting/Chappell/Kallis.

  • SR on April 19, 2012, 18:34 GMT

    This is regarding the comment "In fact I will take a challenge that I can show ANYONE, barring Bradman to have had his weak spot." Since you said anyone- not just batsman :) -I submit Glenn McGrath

    Averages:

    v Bangladesh 24.8 v England 20.92 v ICC World XI 14 v India 18.64 v New Zealand 25.33 v Pakistan 21.7 v South Africa 27.33 v Sri Lanka 22.24 v West Indies 19.38 v Zimbabwe 15

    Averages in:

    in Australia 22.43 in England 19.34 in India 21.3 in New Zealand 18.4 in Pakistan 31 in South Africa 23.62 in Sri Lanka 29.2 in U.A.E. 7.4 in West Indies 20.7 in Zimbabwe 15

    by continent:

    in Africa 22.14 in Americas 20.7 in Asia 23.02 in Europe 19.34 in Oceania 22.05

    by innings:

    1st match innings 22.92 2nd match innings 21.2 3rd match innings 22.57 4th match innings 19.49

    So he has been really good home, away (all conditions), all innings vs all opponents. I know I'm biased..but I cant see a weak spot there unless you include "ability to avoid a cricket ball when playing soccer"[[ There are some minor weak spots: in Pakistan and Sri Lanka. But I get the point. In fact Barnes is also like that. Ananth: ]]

  • SWARZI on April 19, 2012, 17:32 GMT

    Cont'd: To put the foregoing views into the perspective of the Tendulkar vs Lara debate, I always thought that the ESPN/Cricinfo jurors were wrong when they were asked how come they picked Tendulkar into their ‘All time Test Match 11’ instead of Lara, and they indicated that Tendulkar’s ODI records would have had some influenced on their decision. On a fair basis though, Tendulkar would always head the pick for an ODI All Time 11. But people don’t realise that despite his very very long career playing test cricket, he only holds two of the many illustrious records in the purest format of the game. That is, the most runs and most 100s which is not in any way surprising. All of the remaining records (with one or two exceptions) are still either held by Bradman or Lara. And as I said before, it is easy to rate Lara and Tendulkar in test cricket, as they both batted for the overwhelming majority of their careers in that format at the same position No. 4. It is not fair to compare them in ODIs because, Tendulkar is an opening batsman who played in 80 innings in the middle order and did not score a single 100 at any time batting the middle. On the other hand, Lara batted both as an opening batsman and in the middle order and scored HUNDREDS In both positions, so there would be no fairness in trying to find parity between them in this format [[ Yes, that is what a few have suggested, that batting position in ODIs is very important. We are always ready to put down Bevan's and Hussey's averages as the benefits of middle-order batting and increased number of not outs. But we are not ready to accept that accumulating runs in the opening position is easier than accumulating at no.4. Ananth: ]].

  • SWARZI on April 19, 2012, 17:30 GMT

    Ananth, Sir Don Bradman’s records are the indisputable ‘almost’ perfect measures that set the international standards to determine great batsman-ship in cricket. Sir Don built his peerless reputation by plying his trade via the principles and precepts of the purest or traditional format of the game only. And so did all of the other greats before and after his time (the Hobbs’, Huttons, Sobers, Pollocks, Headleys, Mohammeds, etc.). I therefore support the purist society who unapologetically refuse to combine the records achieved by any contemporary player via their exploits in the lesser/easier or ‘watered down’ versions of the game with those that such player achieved in the purest format, just to bring his status as a batsman on par with those traditional greats, who achieved their reputations via the arduous gruel and grind and hard work conditions, which exclusively describes the conditions under which the original matches were designed to be played. The purist society take this stand as they are adamant that it was by no fluke that the founding fathers of this incomparable sporting discipline designed it to be an institution for the overall development of young men; whereby they are put in match situations where the playing conditions are characterized by very very long days (more than a week in the earlier days) under the wretched heat of the sun (during the hours when it was at its hottest) in order to make them: well disciplined; very fit; mentally indomitable and tenacious; intellectually inquisitive and alert; highly ambitious and proudly patriotic; fiercely competitive but socially humane and humble; and achievement oriented but via the virtues of hard work only. The purist also deem it grossly unfair to denigrate the invaluable accomplishments achieved in the traditional set up by trying to bring parity between them and achievements gained via the ‘Wall Street’ oriented cricket culture that dominates cricket arenas today; where ‘the dollar’ sets the goals and objectives of the game; and the need for such virtues as mental indomitability and tenacity; patriotism; intellectual alertness; self worth; discipline and especially hard work is significantly reduced, and, to nothing in certain cases! None of this ‘stock market’ type cricket can really be compared with the traditional format of the game! Many Long days of hard and taxing work in the burning heat of the sun for a peanut pittance cannot be equal in any way to a few minutes of wishy-washy swash buckling Roman gladiatorial-like furor in a cricket arena, for which an undeserved fortune is paid. In fact, I think that both ODIs and T/20 formats lack too many of the unique ingredients of the traditional format to be classified as ‘pure cricket’. American Baseball also looks like traditional cricket in some ways; but it is not traditional cricket – hence, if Adam Gilchrist had taken up a contract to play for any of the American teams, he could not combine the runs he scored in that format to enhance his pure cricket records! [[ All your points are valid. I am ready to accept that T20s are ouside the ambit of traditional cricket in that 5 for 40 or 50 in 40 would be a failure. However ODI cricket cannot be dumped with T20 cricket just as I agree it cannot be dumped with Test cricket. 10-1-60-5 and 100 in 120 are most likely to be match-winning performances. But I will certainly not do an analytical piece combining Tests and ODIs. They will remain separate. Ananth: ]]

  • Boll on April 19, 2012, 17:13 GMT

    @ Ananth, re. `Giilchrist will be measured only against HIS own RpT figure, only 58.0, not some arbitrary number.` - yes, I understand that Ananth, but in numerous cases for a player such as Gilchrist (or players who batted at 5 or 6) you will be comparing a group of 5 test matches in which they batted 10 times, against another group in which they only batted 5/6 times. [[ 5 Tests and 5/6 times batted in. You are only thinking of the Australian teams of 2000 and 2012 which might have won a few by an innings. Anyhow the problem does not really exist for upto batting position 6. It may only be for no.7. And specifically for Gilchrist. Which other country had a 45-50 average guy batting at no.7. Look at all the benefits the Test as a unit will bring and we should accept this minor situation which might affect at best, 5-10 players. Vettori, when he moved to averaging at 30, moved to no.6. Boucher at 7.18 averages 30.30. Kapil Dev with 7.23 averages 31.05. Imran Khan at 7.06 averages 37.69. Pollock at 7.70 averages 32.32. Cairns at 7.06 averages 33.54. Vettori at 8.09 averages 30.31. These are the batsmen at 7+ and 30+. But don't forget that these are completely within a player's career. Ananth: ]]

  • Boll on April 19, 2012, 16:50 GMT

    @Ananth, re.`the Runs per Test figure which has virtually no bearing on the batting position.` I must respectfully disagree. Even for the top 5 players you mention, there are 2 openers, 1 x No.3, 1 x No 3/4 and 1x No.4 - hardly middle-order players are they?

    Players batting at 5,6,7 are going to be severely disadvantaged. [[ For once, Boll, you are not getting it. Where is the disadvantage. Gilchrist, batting with a BPA of 6.72, indicating he batted at no.7 most of the time, averaged 58+ runs per Test is being evaluated against this figure. He has gathered 5570 in 96 Tests at an average RpT of 58. I may look silly repeating it. But the 58 is Gilchrist's own figure. If he scored 60 he has exceeded par, if he scored 56 he has fallen below par. So where is the disadvantage to him. Ananth: ]]

  • Raghav Bihani on April 19, 2012, 14:57 GMT

    @Anantha: I agree we should look at the figures per test. Runs per test can be taken as the expected performance. The good part is that this takes care of not outs to some extent. A group of 4 tests where a player does not achieve the expected performance can be classified inconsistent. [[ I am more inclined towards 3. Let us hear from readers. However, 4 tests is a reas onable Test series, probably the correct limit. 6-8 innings. Ananth: ]] One can be inconsistent on the upside as well, but the downside is what is a concern. By the way I loved your analysis of Bradman's high average being mainly due to big hundreds and frequent ones. In fact you can breakdown a player's average into various parts/factors. Player having a average of 60 is attributable 5% to not outs, 65% due to 100s, 20% due to 50s, 10% failures. Could throw up some surprises. [[ Yes, Bradman's numbers were a revelation so much so soon I am coming out with an article on the same theme. We might very well have solved Fermat's last theorem. Ananth: ]]

  • Boll on April 19, 2012, 14:45 GMT

    Again, re RPT vs RPI - in the case of someone like Adam Gilchrist, would you not be treating his 5-test series (in which he played between 5-9 innings) as equivalent? [[ Boll, there is single correct solution. There will always be things to take care of. At the end of the day Gilchrist sure does not want to figure in any consistency list. He wants to be known as a game-changer extraordinary. Anyhow Giilchrist will be measured only against HIS own RpT figure, only 58.0, not some arbitrary number. Ananth: ]]

  • Anshu N Jain on April 19, 2012, 14:41 GMT

    @Vikram: I took your comment most positively. Apologies if it came out any different. I was sharing how discovery kind of preceded evidence in my case :-)

    @Anantha: For measuring consistency, you neednt fix any datum at all. The player's career RpT sets the standard by which to measure HIS own career-wide performances for consistency. [[ But if I don't have a minimum, some odd selections would be there. Let us say, at least 2000 runs so that the two wonderful 2000+ runs 60+ average batsmen would get in. Ananth: ]] But yes, the problem posed by Innings victories persists for batsmen in strong bowling units. Looking forward to the tweaks that you have in mind to handle this. I'm guessing there wont be many such instances for the top 15 odd batsmen by career runs scored.

  • Boll on April 19, 2012, 14:30 GMT

    @Anshu N Jain, the problem with using RPT instead of RPI as the benchmark, is that opening batsmen (particularly in weaker teams) have a significant advantage over middle-order batsmen (particularly in stronger teams).

    Michael Atherton for example averaged 1.84 innings per test (67.2 runs per test at an average of 37.69)

    Steve Waugh averaged 1.55 innings per test (65.0 runs per test at 51.1)

    Hardly fair is it? [[ Not really, Boll. It does not matter that one person gets slightly more innings per Test. He opens and has a better chance of getting out. Anyhow we will only look at the Runs per Test figure which has virtually no bearing on the batting position. The top-5 values are Bradman 134.5 Weekes 92.8 Lara 91.2 Hobbs 88.7 Hutton 88.2 The top three positions are three middle-order players. Even in the example given what is the problem. In fact Atherton would be expected to score 2 more runs per Test. Ananth: ]]

  • Vikram on April 19, 2012, 13:12 GMT

    Ananth, I was wondering if taking a career to date RPI of the player will help solve the issue of the starting point for the slices (very similar to how you compute the average of bowlers). This will also take into account the fluctuations that people like RT, SRT, BCL have had over their long careers. As for the topic of not-outs, having done some quick calculations, there is no ideal solution because some will favour one player and some the other (more not-outs vs. less not-outs). One way would be to look at balls faced because a 12n.o. by RD could be from 60 balls as he tried to save the match. To me, that is a complete innings. Or you could use median of balls faced in career, and then within 25% of it is a completed innings. [[ Vikram, available only for 65% of Tests and let us not bring balls played. We should not bring in context at all. That is for the Ratings work. Ananth: ]] Anshu: When I said theoretical, I meant it in a positive way, meaning that this is the best definition of consistency.

  • Anshu N Jain on April 19, 2012, 12:40 GMT

    Prima facie, using RpT instead of RpI seems to make imminent sense. It also takes away the problem posed by Not outs. Not many instances where front-line batsmen suffer from 2 small not out scores in a test. Also, it underlines the very essence of Test cricket - you get another chance to get back. If you fail in both innings, well, you failed! [[ And if we fix the datum as 90 or so runs, it allows a player to have a 100 & 0. I think this is worth looking at in depth. And special situations such as innings wins can be identified and tweaked. Ananth: ]]

  • Hitesh on April 19, 2012, 11:24 GMT

    Hi Ananth,

    Just checked on the starting point issue which was raised by Raghav and it seems to be a very valid one... This one can be tackled by only having a rolling or moving average... I think it's the best to be done in such analysis(did for a couple of players and thr is a huge diff from ur results)... We can't draw conclusion based on any start point as there might be players who may benefit from them... I knw as u have mentioned this was already done previously and it is very time consuming but if we want the best results then this wud be the ideal way 2 do it... U can add the 50+ to this as well and then we can see how things are...

    @ Alex: We r talking abt ODI Batsman and I dont think we need to get fielding into this. then y leave out bowling? just coz Ponting hasnt bowled much? ;)Also, the best of SA or Oz bowlers wud have bowled at SRT in Ind as well... So do I call it their failure 4 not containing him in away matches?? The quality of the attack was still gud rgt??

  • Raghav Bihani on April 19, 2012, 10:04 GMT

    @Anantha: It is a nascent thought. Consistency is the ability of a player to deliver close to his average / median RPI every time he plays. If he makes more than this career AVG/RPI, it should not be an issue. Obviously scoring much more than median is going to take this Median higher and more difficult to achieve in a longish career. e.g. Bradman by scoring big hundreds takes his Career RPI higher and more difficult to achieve in each innings.

    Now the computation. Take every stretch of 4/5 innings below expected RPI as inconsistency. then add up the innings in the stretch and show as a % of total innings. e.g. SRT has 8 streches of more than 5 innings below his RPI of 5. These 8 slices have 57 innings in them, making him 57/311 or 18% inconsistent. The same figure taking 4 innings criteria gives SRT another 8 streches totalling 32 innings below par. Thats says he is (57+32)/311 or 28.6% inconsistent. the same figure for Lara is 70/232 or 30% in consistent. [[ If we talk of performing below-par, then we have to change our way of thinking. In that case a Test seems to be a better unit. Your either-end innings might be offset by brilliant innings in the same Test. Just an example. Lara's Aus/Nzl stretch of 17, 5, 0. 1, 1 was preceded, in the same Test of 17, with 226, an all-time classic, only dwarfed by the 153 and 277. SRT's 0, 0, 8, 0, 41 was preceded by 117, a match-winning innings.. It would be unfair. The more I think of it, I think Tests should be the basis for such analysis. Ananth: ]]

  • M. Siddique on April 19, 2012, 9:51 GMT

    Dear Mr. Ananthanarayanan

    Thanks for the wonderful analysis but I think a batsman's greatness cannot just be determined by SD or CoV. For example GR Vishvanath's 97 and so many other match saving and match winning innings from him, Dravid, Laxman and Ganguly will be obliterated if we go just by Standard Deviation as the Yardstick. I have consciously not included Tendulkar, as I always beleive that he does not always score when it matters. I am thinking of working on an alternate measure which will take into account various factors such as Home/Abroad, Quality of the Opposition Bowlers, Contribution from others, State of Innings etc. Ofcourse since my resources will be limited, I might be doing it only for Indian Players like Tendulkar, Dravid, Laxman, Sehwag, Ganguly, Gavaskar, Vishvanath, Chauhan, Mohinder, Azhar, Vengsarkar and Sandip Patil.

    But ...keep up the Good work.

    Regards

    Siddique [[ I think you are probably new to this blog. Millions of words have been exchanged on this topic. What you are saying is Wisden-100 done by me 10 years back and on the anvil for some fine tuning now. Will come out with a fresh list in about 2/3 months' time. Ananth: ]]

  • KT on April 19, 2012, 9:06 GMT

    @Ananth - good job dude...:) must have been pain staking analysis.

    i am often baffled when people compare likes of gavaskar, border, richards with Tendulkar, Ponting et al. Do we have the same green pitches?? do we have bowlers like lillie, marshall, roberts? but we do have so much protective gear..can a bowler bowl 3 bouncers??

    we will never get a level playing field to compare players from different genres. But once you complete test bowling analysis, i would like you to apply some correction factors to batsmen on basis of bowling ranks. e.g - a batsman scoring runs against roberts, marshall etc in WI in 70s shud get more weigthage/marks as compared to one scoring runs against merv dillon, roach etc in India...

    But nevertheless, great job.. waiting for ur bowling analysis

  • Pujari Rajkumar on April 19, 2012, 8:15 GMT

    But this analysis completely ignores the conditions in which the runs are scored, a quintessential dimension of an innings. Please give weighted mean type of weights to each innings a batsman played depending upon the match score which is between 0 and 2, multiply it to the score of the batsman and calculate all these again. That would be humungous amount of work, but would be a treasured possession once it is complete. May be a near complete analysis as well. And relate the performances to the performance of the bowling attack as well. Give another factor varying between 0 and 2 for the bowling attack and divide the score by that factor. Because sometimes, Rahul Dravid and Kevin Pietersen in recent whitewash of India in England, batsmen had to face different bowling attacks with varying amount of steel in them. This analysis is absolutely wonderful and thorough. Tremendous work,looking forward for more.... [[ I think you are probably new to this blog. Millions of words have been exchanged on this topic. What you are saying is Wisden-100 done by me 10 years back and on the anvil for some fine tuning now. Will come out with a fresh list in about 2/3 months' time. Ananth: ]]

  • Anshu N Jain on April 19, 2012, 8:09 GMT

    contd.. It appears that the ratio of metrics 1. and 2. above (previous comment), i.e. (RpI of scores above Median/RpI of scores below Median) would best encapsulate how tightly the scores lie in a range, and therefore the consistency. The smaller this ratio, the greater the consistency. Though this needs to be checked and verified by doing the same for the other batsmen. [[ Anshu, I will do this as part of my follow-up. At the outselt the idea is simple and good. Ananth: ]] @Vikram: Yes, it is theoretical, though I instinctively came up with the metric and only later researched the theory: The ratio of Median to Mean can be used as a measure of Skewness, a word that appears to fit this analysis. Again, my limited understanding of statistics to blame if this isnt correct! Apologies beforehand.

    Scores above 50 would, as Anantha said, best convey contribution and not consistency, something I pointed out in one of my earlier comments as well.

  • Raghav Bihani on April 19, 2012, 7:56 GMT

    @Anantha

    The other small point I wanted to make is with regard to not outs. I understand the logic of using RPI and the igonirng the notout situation. I have no issues with 153* or 80*. These are full innings and can be taken as such. The issue comes when we have small one like 8* or 12*. It is accentuated if you more than 2 not outs in a single slice.

    I do not know how many slices are affected but even 1 slice in 2-3 players would be substantial. Best to ignore such efforts or add them to the next innings to give extra credit. Penalising is a bit harsh on the player. [[ I am very clear on this as far my other rating work is concerned. I consider an innings as an innings only if there is a dismissal or the runs scored is in two digits. In other words ignore single-digit not outs. Can I throw a Solomonic (or Quixotic, some might say) suggestion. Add the number of runs scored in not out innings below 50 (or the player's RpI). Depending on this sum, not likely to be big, add 1 or more innings to the innings tally. Lara (6): 14*, 48*, 13* will lead to 75 runs and 2 innings. (400*, 153*, 80* are all full-fledged innings). Watson (2): 41* will be one innings. (Other inns is 120*). Cook (8): 34*, 1*, 14*, 49* will lead to 2 innings. (104*, 139*, 109*, 235*). Bradman (10): 37*, 30* will lead to 1 innings. (299*, 103*, 144*, 102*, 56*, 57r, 127*, 173*). Pollock RG (4): 33* will lead to 1 innings. (65*, 77*, 67*). I think is an excellent idea and is probably the fairest one. Tough to work out but the program will take care of that. Let me hear from all. This will put this vexing problem to rest, once and for all. Ananth: ]]

  • Raghav Bihani on April 19, 2012, 7:42 GMT

    @ Anantha,

    I have no issues with slice of 10 innings. Anything between 8-12 works for me. The real issue was with the "starting point". Do you start your slices from innnings no.1 or 3 or 6 of a career. You have potentially 10 starting points. Each will reveal a player in a different light.

    I have given you example of Lara (not prove anything about the player). I have done analysis by starting at innings number 3 and 6. Now we have 2-3 slices below 300 runs and SD has also gone up. Mid-3 value is different. He is no longer that consistent. Similarly starting at some other point could make Lara even more consistent. The same is true for all players. Basic conclusion is that "Starting Point" affects the players consistency figures a lot.

    Changes the number of innings can also affect the analysis but here I get the feeling that the factor is not that important. But I would like you to test check a little. [[ Raghav, many thanks. I understand clearly now the point which was made on the importance of the starting point. There are about 5 really poor stretches in Lara's career. It has so happened that these poor stretches have all been split and gone across slices. Let me assure you that this has happened just like that. Even if I wanted to, I could not have manufactured that. I have not done the SD of these values. But even an inspection makes me think that the SD will exceed 0.35. Of course the mean would be around 1.0. That does not mean anything. Possibly the 8-innings slice should take care of that. However the problem might be there also with the starting point. The impact might be less. Same with 6 slices. Even the number of 50s would have similar problems, with less impact. However the final point, Raghav, is that what I have done is natural. Start from innings 1, apply 10-innings slices until we reach the last slice. Do a bit of adjustment for only the last slice, if the number of innings falls below half. And this is done for all in an identical manner for all. And then I have interpreted the numbers to have come out with my own definition of consistency. And starting with innings no 6 or 8 was not ever envisaged. It is clear that if that is done it changes the whole distribution of runs. And defeats my definition of consistency. I think there is a need to look at this in depth. If I am going to follow the fixed number of slices I should do the distribution pattern for every number from 6 to 12 and draw some conclusions. To me now, Anshu's suggestion seems to make sense. And the 50-run method would bring out whether there were significant contributions in a consistent manner. Once again, many thanks. Ananth: ]]

  • Anshu N Jain on April 19, 2012, 7:35 GMT

    [What does this difference of 3+ runs indicate.]

    Simplistically speaking, it probably suggests that Ponting (RP) had more failures (Score l.t. 25 - 47%) and more big scores (41 100s) compared to Dravid (RD) (Score l.t. 25 - 43%, 36 100s) the essence of inconsistency, relatively speaking, in my view. RP does score 1 run more than RD per Inning though. I went back to the numbers for these two, and did further crunching. Observations: 1. RpI for scores below median - RP (11.3) l.t. RD (12.8) by 1.5 2. RpI for scores above median - RP (83.8) g.t. RD (81.6) by 2.2 3. % innings with scores between 26 and 100 (~50% of RpI to ~100% of RpI range) - RP (39%) l.t. RD (44%) 4. RpI for scores between 50 and 100 - RP (68.4) l.t. RD (71.1) by 2.7 5. RpI for scores >50 - RP (99.6) g.t. RD (98.6) by 1 6. RpI for scores >100 - RP (146.1) l.t. RD (146.8) by 0.7 contd..

  • kantipur on April 19, 2012, 7:29 GMT

    I wanted to check the accuracy of Anantha calculation. For the consistency comparison of Tendulkar VS Lara, I looked at the ranking graph and time specific ranking in the ICC website. Lara remained within top 5 most of the time he played and never gone down to number 20. Tendulkar on the other hand spent a long time below top 10. He even gone down to below top 20 twice. Just showing Lara was indeed more consistent than Tendulkar.

  • Vikram on April 19, 2012, 6:36 GMT

    Very interesting stats, Anshu. I guess all 3 methods that we have been discussing, the one done by Ananth , the idea about 50+ scores as well as your method, all approach consistency from a very different perspective. Your method is the most theoretical , Anshu, analysizing consistency on an innings by innings basis. Ananth's method is more about checking consistency from a series performance perspective. The difference is clear when you take the example of SRT vs BCL. SRT is more likely to be consistent on an innings basis (if we look at RPI based calc.), but Lara through his big scores is more likely to be consistent from a series perspective. A comparison of the results through these two methods would prove or disprove our hyptotheses abt players like KP or BCL. Do what anyone may though, it is difficult to undervalue AB's importance to Australian revival through his doggedness and consistency. Hats off to a legend!!! The 50+ score method evaluates who was consistently valuable. [[ Vikram, my method has the problem that it is very much dependent on the starting point. Raghav has done some alternate starting point work and the changes are significant. He will post that comment soon. I think this requires a serious re-look. Anshu's is most accurate and the 50s-based one is very practical from a contribution point of view. Ananth: ]]

  • Alex on April 19, 2012, 5:54 GMT

    @Ananth: Let us know what you want to plot and we will make some suggestions.

    @shrikanthk: The "comments" space on Ananth's blog was never restricted to discuss strictly the main topic of the article. That decision has helped Ananth get new ideas. My own spectrum of competency in cricket history is fairly narrow but deep. I make suggestions within its realm and never talk about anything else.

  • Ananth on April 19, 2012, 5:03 GMT

    I have sent my next article for editing. This relates to the Test bowler peer analysis within the same innings. Much more nuanced than the Test Innings peer analysis which I had done during Dec 2010. I find that many readers are extracting Statsguru segments on player performances in different formats and presenting to push up or put down players. How do we draw inferences when we are not sure about the authenticity, accuracy and the assumptions made. And Statsguru does what you ask it to do. An omnipotent but dumb tool. That is the way it should be. Hence I have decided to take up a major task. After my next article, I will do the following four articles, not necessarily contiguously. I may re-visit the Test batsmen consistency with the tweaks in between. There is also the ODI consistency analysis to be looked at. ODIs: Top batsmen performances, vs country, career/home/away. ODIs: Top bowler performances, vs country, career/home/away. Tests: Top batsmen performances, vs country, career/home/away. Tests: Top bowler performances, vs country, career/home/away. This will be a definitive compendium of data tables which readers can download and use to analyze themselves. My current idea is to have 10 buckets for ODIs: Australia, England, India, New Zealand, Pakistan, Sri Lanka, South Africa, West Indies, Bangladesh&Zimbabwe and all other countries. The problem is not in compilation of data. Last night I mailed my article for editing and I have got the first set of tables in an hour today. the problem is presentation within the article. It is a huge amount of data and I have to look at ways of presenting the tables in an efficient and readable form. Use of graphs is also an option. Depending on the response, I may do the same thing, by year. Ananth

  • Alex on April 19, 2012, 3:43 GMT

    @Boll & @Ananth (& @shrikanthk also!): ODI record is one factor where the big 3 definitely separate themselves from the likes of Dravid and even Kallis --- despite a major upgrade since 2006, Kallis' career SR is still only 73 and his record vs Eng in Eng and vs Oz in Oz is quite poor.

    I am eagerly awaiting Ananth's article on Lara. Due to a variety of reasons, media & audience in 80's had come to to praise "dominating" 50 or 100 more than big hundreds. More than anybody else, Lara made big scores fashionable again and also pushed the envelope further than anybody else.

    Finally, I looked up http://www.youtube.com/watch?v=RJLZBjIjDHI to observe SRT in Oz (his weakness in ODI's) during his disastrous 2000 Oz tour. The quality of stroke play in this innings is amazing and Pak attack was outstanding. And yet, SRT's ODI record in Oz is pretty poor!! [[ Alex: I have outlined some near-future plans of mine in a separate comment. Ananth: ]]

  • shrikanthk on April 19, 2012, 3:09 GMT

    By the way, talking of Bradman's consistency - One thing to be borne in mind is that his most inconsistent phase was from 1932 to 1934.

    There was a reason for this inconsistency of his - Bodyline in '32-33 and near-fatal ill health in '34. [[ The sequence of scores during this period was 0, 103, 8, 66, 76, 24, 48, 71, 29, 25, 36, 13 and 30. A total of 529 at 40.7. Only 4 50+ scores out of 13. The average dropped from 112.29 to a "pitiable" 89.56. We should do well to remember the two scores before and after this period of darkness: 167 & 299 and then 304 & 244. Who are we talking about and what are we talking about. And there are suggestions that Bradman was not that good. Ananth: ]] So if one makes allowances for these two extraneous factors, his figures should appear more consistent than what they are now.

    Also, the other name that stands out for me is Jack Hobbs. His low SD is astonishing given that his career lasted nearly 30 years! Hobbs was nowhere as consistent as this in his First-class career. He averaged below 40 in several first-class seasons prior to WWI. However he turned into this run-machine after turning 40 post War. Goes to show that 20s was indeed a great time to be a batsman as compared to the noughties/teens of the 20th century.

  • shrikanthk on April 19, 2012, 2:40 GMT

    I agree with you and as I said it should be a weightage of 2:1 in favour of tests

    Vikram: It's not about the weightage being 2:1 or 3:1. My question is - why would you hire an auto-rickshaw when you have a Mercedes lying unused in your garage!

    Anyway, I give up as always having failed to change people's minds. This was meant to be an article on Tests! Yet we had comments on SRT in ODIs vs Lara in ODIs and more recently Ponting in Home ODIs vs Ponting in Home Tests. This is what irritates me. When Ananth posts an article on ODIs, I don't see people dragging Tests into the picture in the comments!

    It wasn't my objective to turn this thread into a Test vs ODI debate. I was forced to do it by other comments.

  • Boll on April 18, 2012, 22:02 GMT

    Ah, apologies for completely misreading the revised consistency calculations - did all seem a bit simple...@Alex, Ponting`s overall record against India is quite puzzling. In tests he averages 86 at home and 26 away, in ODIs he averages 28 at home and 47 away. Go figure.

  • Alex on April 18, 2012, 15:01 GMT

    @Hitesh: Thanks. Ponting's poor record vs Ind in Oz is unexpected & puzzling indeed. Do note that his SR is better than SRT's in Oz & SA. Outside Oz & SA, SRT was certainly most destructive (vs anybody) of these 3 and overall, as said, I put him within plus or minus 2% of Ponting and maybe 5% ahead of Lara. It is just that SRT's two bleeps vs Oz & SA should be noted since Oz & SA were easily the best bowling+fielding sides of his era.

    I am a bit partial towards Ponting since his fielding gave Oz such an edge in ODI's --- IMO, he is the greatest ever all-round fielder (of course, Viv also was a great fielder and, like SRT, a fairly effective bowler for 5-6 overs in ODI's).

  • Anshu N Jain on April 18, 2012, 14:46 GMT

    contd...

    The same can be done as the ratio of Career Average (instead of Career RpI) to Median Runs between Dismissals (instead of all innings scores). The results using this are as follows: Richards (50.24/35) 1.44 Gavaskar 1.48 Sir Don 1.49 Lara 1.53 Dravid 1.54 Tendulkar 1.54 Kallis 1.56 Ponting 1.56 Border 1.58

    Border and Kallis lose the most because of their high N.O. instances.

  • Anshu N Jain on April 18, 2012, 14:37 GMT

    I was quick to post the numbers, but should have been quicker to check and re-check the Excel for formula consistency. My apologies. Have redone the numbers and posted below.

    Anyway, the methodology is this: Calculate the ratio of Career RpI to Median Score of all innings (including scores in innings where the batsman was not dismissed). The closer the value to 1, the higher the consistency. Since there are enough failures for all players, this ratio value is more than 1 for all: Border (42.2/31) 1.36 Dravid (46.5/33) 1.41 Richards (46.9/32.5) 1.44 Kallis (48.2/33) 1.46 Tendulkar (49.7/34) 1.46 Lara (51.5/33.5) 1.54 Sir Don (87.5/56.5) 1.55 Ponting (47.4/30) 1.58 Gavaskar (47.3/29) 1.63 contd... [[ The low median scores of Gavaskar and Ponting are intriguing. Let us take Dravid and Ponting. Similar number of innings (286 vs 282). With 3 failures the 30 might even come down to 29.5. What does this difference of 3+ runs indicate. Ananth: ]]

  • Hitesh on April 18, 2012, 14:14 GMT

    Hey Ananth, as my posts are generally long, I have been using gr8 instead of great.... :D I hope you will put our case in front of cricinfo to increase the number of characters so that I can start using proper english spellings... :D [[ In fact I myself is subject to the 1000 character limit. Let me ask them. Ananth: ]] Until then I hope you don't mind me using Gr8... ;) [[ I know gr8 and 4u and 2day are today's lingo. What is wrong with "grt". Ananth: ]]

  • Boll on April 18, 2012, 13:56 GMT

    Doug Walters at 41%...

  • Boll on April 18, 2012, 13:47 GMT

    I have Barrington at 44.4% (55 innings over 50, out of 124, not including 7 not outs under 50).

  • Boll on April 18, 2012, 13:28 GMT

    Ok, just to start the ball rolling, I have Bradman with 42 scores over 50 in 78 innings (not including 2 not outs under 50) at 53.8%. I`d also put a bit of money on Barrington and Walters being right up there as well. [[ No, Boll, it is not that simple. That is only the % of 50s which is a back-of-the-envlope calculation. What happens, theoretically speaking, if Bradman had 42 scores of 50s at the start and finished with 38 scores below 50. He has to be termed totally inconsitent. However if he averaged 4/5 fifty-plus scores in each of the 8-innings slices, he has to be considered to be totally consistent. So this work has to be done within the slices to determine the overall consistency. Ananth: ]]

  • S Ganesh on April 18, 2012, 12:39 GMT

    Ananth,

    I have a feeling that most people are missing the whole point of the article in debating SRT is the greatest and ODIv/s Test cricket. The article is basically stating that a consistent batsman is one who in a 10 inning streak will make mostly RPI*10 . By this standard we can bet even money that in the next 10 innings Kevin peterson will score +(-) 465 . The odds are around 85% in favor ( KP being the only current Batsman hence the example)

    It does not comment anything at all on the utility of the batsman or his greatness or indeed the beauty of batting. Hence it may be that a batsman has an RPI of 21 and is 100% consistent which does not say anything at all. [[ But you must agree that Willis was a revelation as also Lara, Pietersen, Border. It has also put to rest some of the perceived notions which were floating around. Ananth: ]] I would suggest that a alternate measure of consistency is to count the number of innings which are +(-) 20% of RPI against the total number of innings played. Of course this will also not comment anything on the greatness of the Batsman. That is and will be a different point. [[ The 50-based analysis is on these lines. Ananth: ]]

  • Hitesh on April 18, 2012, 12:34 GMT

    This is a reply to Alex on Ponting vs SRT... Sorry Ananth a bit off topic but feel Alex has chosen his data very wisely and tried to do some mud slinging and I need to respond to him...

    Firstly, ur facts for SRT and Lara may be perfect but 4 Punter u hve forgotten to put a lot of filters as u guys have been doing 4 SRT.

    1. vs WI in WI - Avg of 27.43

    2. vs India at home - Avg of 28.33 (SRT 31 vs Oz in Oz)

    3. vs Pak at home - Avg of 25.68

    I am nt sure if u can cal him gr8 if he is unable to take advantage of his home conditions... Also, his overall career SR is much lower than SRT which I think is a gr8 factor in ODI's.

    I do admit I have also now indulged in mud slinging so sorry... I have highest regards 4 Ponting n he is an all time gr8... But u have resorted to selective reporting as per ur convenience n hence i had to respond.

    P.S Anath, this was just to let ppl knw tht selective reportin can bring down the best batsmen as well... Hope u wil publish this... [[ Until now I have trashed 4 comments, out of 115. One abused me, first in English and then, luckily for me, in Hindi which I do not understand well. But enough not to show it to my daughter-in-law who is from UP. Two others were much better but ascribed ulterior motives to me. I have published the other 115 so far. So you need never worry on that count. Only one request. gr8 is most irritating. Use great instead. Ananth: ]]

  • Anshu N Jain on April 18, 2012, 12:07 GMT

    continuing from my previous comment, including the scores in the N.O. innings for calculating Median of all inning scores, a value closer to 1.00 for the ratio (Career Average:Median score) indicates higher consistency. This is how the values for consistency measure turn out to be for these select batsmen:

    Richards - 1.29 Lara - 1.32 Gavaskar - 1.38 Dravid - 1.54 Border - 1.74 Bradman - 1.77 Tendulkar - 2.02 Kallis - 2.47 Ponting - 2.58

    Using scores between dismissals to calculate Median will obviously change the measure. But it is too much effort to extract innings scores and adjust for N.O. innings. Ananth, you think this is something you see value in doing on your dataset? [[ Anshu, I need some more clarifications. How did you arrive at Richards' 1.29. Ananth: ]]

  • KnowWho on April 18, 2012, 10:52 GMT

    Hi Ananth, Once again a great article. Once Again SRT vs BCL debate. Two things i tought turned to be false by this analysis. 1. BCL best was better than sachins best. 2. Sachin was a consistent batsmen. Seems like both are false according to the slices you have provided. Whether runs were scored against bangladesh or australia is not thing in question here. One slice in sachins career shows 899 where as inspite of 688 runs or 375/400 in a single innings lara has not come nowhere close to it.However lara has scored consistantly. One morething about comparing ODI's Ponting is a top notch player played well everywhere unlike his test records. Sachin did in test what ponting did in ODI. Ponting credentials as a ODI great is far better than sachin or lara. he has 3 world cups period. lara does not even come into picture how many world cups has he won. ODI can be won by batsmen while bowlers win test matches. So if u judge lara vs sachin ,it is sachin due to a world cup win . Yes to all. Before the WC win, I would have been tempted to say Ponting might just pip SRT. And both have a Champions' Trophy win (it is significant as long as it lasts). But I will never bring in captaincy or fielding into a batsmen analysis.

  • Anshu N Jain on April 18, 2012, 10:37 GMT

    Good follow-up to the consistency for bowlers piece Anantha.

    Given that you are not a fan of averages (Both bowling and batting!) as a measure of performance, I'll refrain from comments on alternate approaches that use averages to measure consistency, but still hold on to my convictions :D [[ Anshu, I am a great supporter of Bowling averages. I think that is the single most important measure in cricket. Never goes wrong. It encompasses the two essences, accuracy and strike rate. And is valid across 137 years. Batting average I have some problems with. Lara/Gooch/Sehwag have 6 not outs each. Kallis/Waugh/Border have 39 or more. Hence my push for RpI. In fact what I like most is RpI with a tweak, which I use in many of my rating work. The tweak is exclusion of single-digit not outs from the innings tally. Is almost perfect. But I do not use that here. I have also given my reason for not using averages in Bowling consistency. I think, in all situations, 30 wkts at 30 is better than 20 wkts at 25. Ananth: ]] In light of this, as suggested by Ramesh and Shrikanthk, % of career scores above the career average could be a metric for batting (of course, ignoring innings with N.O. scores below the career average while deriving the numerator). Now, would a higher figure necessarily indicate greater consistency? I am not so sure. [[ The 50 is a number suggesting that it moves the innings to a strong degree of significance. I have no problems in setting 50 for all batsmen. Ananth: ]] And if I may suggest another, how about the ratio of Career Average to the Median score between dismissals? It appears simple yet effective in measuring the "spread" of scores over a range.

  • Hitesh on April 18, 2012, 7:50 GMT

    Lot of interesting comments coming in... Like d suggestions of 8 innings and 25% and also the 50+ scores... This wud definitely help in knwing who is more consistent...

    1 more thing abt my previous post, yes 20% seemed harsh to me as well but I feel consistency is like a tight rope on which these batsmen need to balance it out... It is not to throw them out of the consistency zone but to ensure that the tough criteria filters out the best amongst them... So to re-emphasize, if you r having a sample size of 10 then 20% seems O.K to me... But as suggested, if u wud want to increase this % then decrease the no. of innings (8 and 20% is a gr8 suggestion)....

    One more reason y I would like the sample size reduced to 8 is coz though thr were enuff 5 test series played b4, in the past 2 decades we have increasingly seen 3/4 tests being played in a series... Moreover, 8 innings may actually cover the whole 5 test series considering thr wud be few innings wins/bad weather... [[ I think we have a number of good responses which have brought out the balance between the number of innings and the band. In fact what we need are only 3 bands: Less than 0.75, 0.75-1.25, greater than 1.25. Add to this the 8-innings slice and we almost have the perfect basis. But that is for a later day. I have to put to bed my next article and taht takes priority. Ananth: ]]

  • Arjun on April 18, 2012, 7:05 GMT

    Ananth,

    If we consider 50+ innings in Test cricket as succsessful one and put a cap of 50 on all the score of above 50 and then determine the consistency, will it be more correct method ? [[ I think this is not a single correct method type of analysis. Maybe for the first time I should do different types of analysis and put up the results for the readers to comment. Ananth: ]] Similarly for bowling, cap of 5 wkts on all 5 plus wkts captured in a test. [[ 4 seems to be more correct. Ananth: ]]

  • waterbuffalo on April 18, 2012, 6:59 GMT

    As long as Saeed Anwar is in it, I'm okay with it.

  • Ramesh Kumar on April 18, 2012, 6:41 GMT

    Ananth,

    One other dimension on consistency could be the significant scores across elapsed years. ie instead of innings/tests slices, we could also look at year slices. With uneven distribution of tests amongst test nations, year slice would give another view of contribution over time period instead of tests. [[ Problem with years would be that there would be many years, I may put this as high as 25%, during which players play so little that the numbers would become meaningless. It is one thing doing that in a single player's cll-inclusive analysis. Across the board, not sure. Ananth: ]] Alex..On Ponting vs SRT.... I am not big on microanalysis while trying to differentiate greats. I put SRT above others in ODIs given his monstrous runs across many years, has faced many different situations and retaining high average & high strike rate. One other dimension is that I feel that SRT has A game, B game, C game etc for different situations as the game evolved which I didn't see in others. It is subjective and I am not going to lose sleep if any analysis in going to put him lower. I respect all the greats and I love the game of many players. [[ I have always said that for me to appreciate Federer more, I should be able to understand and appreciate Nadal and Djokovic. I appreciated Edberg's magic more only when I could appreciate what Becker could deliver. Or Borg when I could appreciate McEnroe's wonderful game, minus the tantrums. Ananth: ]]

  • Gerry_the_Merry on April 18, 2012, 6:23 GMT

    Ananth, while you are tackling 50+, please remember my suggestion to extract peer comparison restricted to 50+ scores. For instance, a Steve Waugh would bail out his team when in trouble, but may not have cashed in when the going was good. so for such batsmen, e.g. Laxman, including failures penalizes them for failing when the going is easy. So I am, to be clear, suggesting that if Laxman fails, making 20, when the rest of the team makes 143, 200, 75 out of a total of 550/6 declared, ignore that. If Laxman makes 96 out of 210, take that in comparison with average of top 7 batsmen in that innings. But If Laxman makes 25 and the team score is 275, and Dravid makes 103*, then Dravid gains, pushing Laxman down relatively. Do you see merit in this approach? [[ Gerry, that is a different analysis. For the purpose of the consistency analysis this may not be applicable. If we ignore failures completely I am not sure whether we gain. And non-contributions may not be as important as contributions but cannot be ignored. Ananth: ]]

  • Nitin Gautam on April 18, 2012, 5:25 GMT

    So Nitin Gautam thinks that I have taken convenient criteria. Why? I have left out sub-continent pitches

    Exactly Gerry..thats my point..why leave sub continent pitches & do you really believe teams playing there & runs made there are substandard & should not be considered & really would like to know where do you rate the runs made by Richards, Lara, waugh brothers, pontings (120 something in QF WC-11), gilchrist etc & players like Desilva,anwar,jayasurya cos they have truly played some extremely memorable innings in sub continent. why this standard of praising them & criticizing SRT. agreed SRT didnt moved the world as per your criteria but ponting avg 39 (home ground) Lara 32, kallis 29..please read my 2nd comment also I am sure you will support on my suggestions to ICC since Anantha called me "going ballistic" Still instead of asia & out side Asia, it still would have been OK to consider Just Home & away but than many would have come under scanner & not just SRT...

  • Vikram on April 18, 2012, 5:02 GMT

    Shrikanthk, last rejoinder from me on this topic as well - you talked about AG Ganteaume but not Kambli vs RD where the sample size is big enough. Examples exist on both sides. Also, I don't think anyone is saying that ODIs and tests are equal. It's just that you can't discount one format. If SRT played ODIs the same way he did Ranji, I will discount it. However, SRT played a whole career worth of games and so should be evaluated on both, together. While you talk about ODI format issues, there are enough issues in the test format as well, which can similarly be questioned. Bouncers for one. Having said that, I agree with you and as I said it should be a weightage of 2:1 in favour of tests. Back to the topic Ananth, based on the discussions, I believe you are being too lenient with 10 innings and 33% while others are being too strict with 6 innings and 20%. I believe a 8 innings, 25% is the correct criteria. It might just help identify those who were regularly consistent. [[ Vikram, you are very precise with your comments. You have caught the essence of this analysis. In fact I did my initial work with 8 innings but went to 10 because I felt that until 2000, there were so many 5/6 Test series. However since the slices are continuous , it really does not matter. Also I had used 5 Tests for Bowling. If I had continued that, the number of innings would have been around 8/9. I think 8 & 25% makes it tougher on the batsmen but not unfairly so. I will do this later along with the interesting 50-run idea being floated by Ramesh/Shri and Arjun. Ananth: ]]

  • bks123 on April 18, 2012, 4:43 GMT

    10 innings with 9 scores of 10 and a 310* will make it 400. But 10 innings with two 100, two 50, one 40, two 30, two 0 will make it 400 too.If we calculate the average of these two players, the guy with a 310 not out will average ~45 after 10 innings whereas second guy with two 100s and two 50s will average 40 after 10 innings. Now who is more consistent? Let us consider a hypothetical case, suppose someone scores a 310* on his debut and then scores 9 scores of 10 to average 45 in 10 innings. If he scores another 9 scores of 10 from innings 11-19, the selectors may not even give him the opportunity to play the 20th innings to score 310* to maintain his consistency (~ave-45). I always believed that Lara was inconsistent amoung the top batsmen of modern era and it is his big scores like 400, 375, 277 between poor scores that kept his average intact. But after this analysis I went back and looked at his innings by innings scores and they are not as inconsistent as I had imagined.

  • shrikanthk on April 18, 2012, 4:10 GMT

    Back to the topic of batting consistency -

    I am not a huge fan of metrics like Standard deviation in this context.

    I'd like a simple metric - percentage of innings with 50+ scores. That's it.

    To my mind, in most cases a score of less than 50 is a failure of sorts in a test innings. A batsman may put together a string of 30s/40s. But that generally doesn't help the team cause signficantly in test matches.

    If one looks at 50+ percentage, I am sure DGB as always will be at the top. (Think Vinay made the same point).

    While calculating 50+ percentage,we must probably exclude unbeaten sub-fifty (<50 runs) innings from the denominator. [[ Thanks a lot. Very valuable comment and on the lines of Ramesh's suggestion. Will look at it without over-complicating it. That is what Ramesh also wants. Ananth: ]]

  • shrikanthk on April 18, 2012, 3:58 GMT

    If ODIs allow for a gray zone rather than its binary win/loss result, I daresay it will become a better test of the cricketing ability

    My final comment in this regard (since Ananth isn't particularly relishing these digressions) [[ I was only lementing on the absence of specific suggestions from you, that is all. The so called "digressions" are the ones which make this blogspace. Ananth: ]] If ODIs do allow for a draw then it will definitely be a far better format. Yet, for that to happen the ODI needs to be a 6/7-wicket game restricted to 50 overs per side. The MCC should've thought through all this when they introduced this format back in the 60s! We are paying for their mistakes now.

    Even if you do make these tweaks, the ODI will still be a lesser format compared to tests. The reason why test matches last for 5 days is not to satisfy the whim of an idiot. But because history has taught us that you need that much time to bowl sides out twice on reasonable pitches!

  • Gerry_the_Merry on April 18, 2012, 3:54 GMT

    So Nitin Gautam thinks that I have taken convenient criteria. Why? I have left out sub-continent pitches. Except Richards and Chanderpaul, in the list most batsmen are from sub-continent. So it really is sub-continent batsmen in non-subcontinent conditions. So there is no relevance of Ponting failing in India etc.

    If one wants to include NZ, one can, but few matches, little change in results, and not really consistently good quality opposition, except for 2 years in early 2000 decade.

    Shrikanth says ODIs are lousy. I agree for 85% of ODIs. But Waugh 120*, Richards 189*, Maindad 116*, Gilchrist 149, are not innings ordinary batsmen can play. So in ODIs too some effort can be put to extract "tough conditions + opposition". I have done in an uncomplicated way. Another way could be looking at finals / semis / QFs etc. But that makes it very subjective, and many batsmen will not get opportunities. For instance, Imran played in one final, Richards in 3. [[ Not to forget Kapil's 175*, Styris's 141, O'Brien's 113, Symond's 143 and so on. Ananth: ]]

  • Alex on April 18, 2012, 3:07 GMT

    @shrikanthk & @Ananth: I thought a lot on why shrikanthk has a gripe vs ODI's. IMO, since ODI's (& now T20) are legitimate & financially more lucrative components of cricket, those must be used to rate a player. Throughout history, the best performers always optimized their efforts to suit their times, and ratings are always contemporary --- e.g., we rate a Hindustani classical vocalist on how well he sings ragas & semi classical, and not on how well sings dhrupads, even though that is the purer sterner test.

    In all spheres, survival is the true test. Not only does the test cricket test a player more on it (e.g., McGrath can bowl 10+ overs & Lara/Boycott can bat for 200+ balls) but you also have an additional result, viz., a draw. That creates more possibilities to test the survival abilities. If ODIs allow for a gray zone rather than its binary win/loss result, I daresay it will become a better test of the cricketing ability --- albeit in a tournament setting.

  • shrikanthk on April 18, 2012, 3:03 GMT

    Lastly -

    I know several people here have grown up watching the white ball game. So it is probably very painful for them (and me as well) to accept the truth that the format isn't as great as they made it out to be in their childhood.

    But this thread is supposed to encourage hard-headed analysis. Not sentimentality.

    If Tendulkar were reading this thread, he will be embarrassed to know that his fans are propping up his case against Lara by saying he did better than Lara in ODIs! Seriously. Sachin deserves better than this. [[ Very, very unlikely. Knowling the way Tendulkar thinks, he would not place Tests at that high a pedestal. And I am not his "fan" in the way you have used. There is no pushing one up through articifial methods. Ananth: ]]

  • shrikanthk on April 18, 2012, 2:42 GMT

    But, Shri, you have brought in the players who were nobodies in Tests and somebodies in ODIs. There are many who did well in both formats.

    That's the whole point!!! The ODI figures don't discriminate skill levels too well. So it's a boon to journeymen.

    To say that several people do well in both formats is like saying - "Both a Math graduate and a bright high-school student can solve quadratic equations". Yes they can. But what does that tell you about their relative levels of Math knowledge? Nothing. One requires a tougher examination to figure out that the Math graduate indeed knows more.

    Ofcourse we may reach a point in future when ODIs/T20s are all that get played with the odd Test thrown in. In such a situation, use ODI stats by all means because you have nothing better to fall upon! But we haven't reached that state of affairs yet. The greats of the present era have played tests in hundreds. So one needn't look into ODI/T20 stats to bolster their claims. [[ My hands are becoming wary and tired. So I will make my comments short. I have one major gripe with you. You rarely comment on the article on hand. Your comments are almost always on the periphery. They are all Test based and at times comparisons with ODIs. That is all. You talk about Tests, FCs with a level of deep understanding. You respond to readers' comments specifically. We all get a lot of insight in all these areas. But I (and the fellow readers) have no idea what you think of Test batsmen consistency, the methodology used, the pros and cons, is Lara's seeming consistency a mirage, what is better: 100 and 0 or 50 and 50 etc. The other readers might wander off. But they would come back to ask me to take 6 instead of 10 innings, consider series, express surprise about Lara and Pietersen, express happiness at the stamp of approval on Border etc. I am sure you understand what I mean. You have knoweldge in depth and a very keen analytical nous. Address these, at least partly, to the topic at hand. Ananth: ]]

  • Swamy on April 17, 2012, 19:20 GMT

    Hi,

    I agree with Mr.Imhotep. Considering series performance rather than 10 innings will throw more light on the consistency of the batsman under different conditions. For instance a 10 innings average might nor account for variations in a high performing short home series, and a very bad away series performance. Since 2 test and 3 test series are now not that uncommon. [[ This shows taht you have neither ream my response to Imho's comment nor have you taken the trouble of going through the series-based analysis done a few months back. Ananth: ]]

  • shrikanthk on April 17, 2012, 18:55 GMT

    Also, stats like CLint Mckay and Mcgrath can be repeated at test levels as well. If average or RPI is the best measure, then the best batsman in world history has to be AG Ganteaume

    The problem with cases like AG Ganteaume is sample size. Pure and simple.

    In ODIs one can name umpteen cases where the averages don't discriminate skill too well even after a hundred games. A smart bloke like Nick Knight averaged 40 in ODIs by "playing the percentages". The same chap averaged 24 in Tests. A Chris Harris made a reputation for himself as a bowler in ODIs playing 250 games. The same bloke averaged 73 with the ball in Tests! A Xavier Doherty may be just as hard to get away in ODIs as Shane Warne. The same person is not good enough to play even a dozen tests for Aus. [[ But, Shri, you have brought in the players who were nobodies in Tests and somebodies in ODIs. There are many who did well in both formats. Ananth: ]] I can go on and on. The fact is that in ODIs it is often difficult to discriminate between a utility player and a class player. And that shows in figures.

  • Alex on April 17, 2012, 17:48 GMT

    @Ananth: I didn't mean to criticize your 10-innings slice analysis. I only said that supplementing it with Vinay's 6-innings slice suggestion would give a better perception. One other thing to examine would be 10-innings slice analysis for "team contribution" --- i.e., % of runs contributed to the total team runs ... you have done it for Dravid in his tribute article (which, hopefully, that erudite gentleman will read some day). [[ I never take offence, Alex. You know very well when I take offence. The overlying principle is not to look for inconsistency but consistency. 10-innings gives chance for a batsman to go through short difficult periods but still recover. Ananth: ]]

  • Alex on April 17, 2012, 16:42 GMT

    @Ramesh and @Ananth: In many ways, as the following facts show, there is an overwhelmingly strong case for Ponting (& some others) over SRT in ODI's. Let's rule out BD, Holland, Nigeria, etc.

    1. SRT's low averages: 31 in Oz vs Oz, 25 in SA vs SA. This is really bad. Also, he was mediocre is Oz and SA vs good bowling attacks: some truly great innings and a few 50's are exceptions to this norm.

    2. Lara's low averages: 30 in Eng; 32 vs Ind in Ind. Apart from these two bleeps, Lara (esp. until 2004) was uniformly good. Even after 2004, he played well vs top attacks but, surprisingly, failed vs poor attacks.

    3. Ponting's lowest average vs a nation is 36 (Pak) and in a nation is 39 (Ind). He was excellent everywhere and against everybody.

    I think Ananth could do a nice piece on "consistent excellence" by looking at "vs a nation" and "in a nation" splits. [[ Will certainly do. (Hopefully) At least people won't say that I have done that to push up some batsman or other. Ananth: ]]

  • Alex on April 17, 2012, 14:46 GMT

    @Ananth and @Vinay: I think Vinay is correct --- 10 innings slice is too long and potentially covers an entire 6-test series. You might get more relevant results by looking at 6-innings slices: this gives a batsman a chance to play at least 3 big innings ... this is useful since most top batsmen tend to post a 50+ innings in every 3 innings or so, and a 100+ score every 6 innings or so. It might be a good idea to tabulate how the 10-innings slice result compare with the 6-innings slice results. [[ Last time both Arjun and Raghav argued for longer slices. Now you are talking about shorter slices. I am going to stick to 10 for the simple reason is that it covers a complete old-fashioned Test series and the correct method is to wait for end of the series to see how he measured up. With 10 Tests, I can say with total confidence that if a person scores 50% of his career RpI, he IS out of form. I cannot say that with a lower 6-innings slice. Ananth: ]]

  • Hitesh on April 17, 2012, 13:07 GMT

    Thanks 4 d suggestion.. I have wrkd on the file which u had saved n have kept the SPF between 0.8 to 1.2 for being consistent.. The good thing is that none of the batsman gets a 100% score. The highest in the list is Ranatunga (75%) followed by Grenidge... Surprisingly none of the batsman with 100% in ur table tops the list... (Saeed Anwar has only 40% in the 80-120 range)... KP and Cook move up in d list whereas Collingwood moves down significantly... The interesting fact is that Lara, Don and Sobers are tied (37.5%) with Gavaskar and Jayawardene following closely... The other notable shifts r 4 Kallis, Dravid and Tendulkar all of them in mid-40's... For Ind, Dhoni moves to the top spot followed by Ganguly.. I am not sure if this is the best result for consistency but felt 20% on either side wud be the best way to judge a player for consistency (which still gives them 40% bracket to fall within)... I wud like to share the file with u. Let me know how can I send it across.. [[ Hitesh I think 20% on either side is too harsh. It is borne out by the results. The purpose is not to get people out of the consistency zones to prove them inconsistent. I am not sure whether 33% on either side is too high. But one thing I am certain is that 20% on either side is too narrow a band. I think 25% on either side seems to be better. Please do that. I will work out a method for you to send the Excel file. Ananth: ]]

  • Dinesh Reddy on April 17, 2012, 13:04 GMT

    Hi Ananth

    Take a look at the links here. Sachin Scored at an RPI of almost 50 in England and Australia Lara scored at an RPI of 44. So there was a whole difference of Six runs and thats a huge gap. For your cross checking i am providing these Statsguru Sorted out links For Sachin http://stats.espncricinfo.com/ci/engine/player/35320.html?class=1;filter=advanced;host=1;host=2;opposition=1;opposition=2;orderby=start;type=allround;view=match

    For Lara http://stats.espncricinfo.com/ci/engine/player/52337.html?class=1;filter=advanced;host=1;host=2;orderby=start;template=results;type=allround;view=match

    SO this shows that Sachin Scored more runs than lara at more RPI in tougher conditions than him and this despite him having two nightmarish tours(BY his standards) in 2011.

  • Hitesh on April 17, 2012, 10:47 GMT

    Hi Ananth,

    I was just having a look at the tables again and found one thing which was a bit weird... Considering we are talking about consistency over here, don't you think grouping Batsmen who have a SPF between 0.67 to 1.33 as consistent is giving a lot of lee way?? This is 33.33% on either sides (67%).. Ex. a person having an RPI of 50, even if he scores 34, he is deemed consistent...

    I feel if u tighten this a bit and keep the SPF between 0.8 to 1.2 or at the most 0.75 to 1.25 as the ones consistent, then it would be good to see how it changes the chart... I am not sure if this may have a big impact but would like to see the results...

    @Nitin Gautam: Thanks for understanding my perspective and explaining it to Gerry... :) [[ Excellent idea. Why would you not do it yourself using the Excel table and post the results. Quite easy. That is the idea of posting such comprehensive backup tables. Ananth: ]]

  • Vikram on April 17, 2012, 10:36 GMT

    Shrikanthk, I understand your passion about saying that ODI and tests are two different forms of games. So when you analyze these games, results and teams, they can be done completely independently. However, the players, especially batsmen, have to be analyzed as a package. I agree with Ananth's method of giving more weightage to tests but even that can not be more than 2:1 at the max. Also, stats like CLint Mckay and Mcgrath can be repeated at test levels as well. If average or RPI is the best measure, then the best batsman in world history has to be AG Ganteaume and the best Pakistani batsman of all times is Afaq Hussain or Taslim Arif. Or on a like for like comparison to the example you gave, Kambli is better than RD. As for IPL, it will destroy Indian cricket same as EPL did to the English football team. The last great English player was Gascoigne, but the press manages to find atleast 4 world-class players each year who then under-deliver thereafter (Theo being a prime example).

  • Fan on April 17, 2012, 7:35 GMT

    I know you are statistician but I love to see some qualitative analysis from your side. Batsman purely rated on the basis of entertainment value. It will be fun :) [[ First, let me say that I am not a Statistician but a Cricket Analyst. Now how do you define entertainment value. Scoring fast, hitting fours and sixes and then we will see. Today's first session will be one of the most boring ones or one of the most engrossing (note the word used: not entertainment) ones depending on how one views Test cricket. Ananth: ]]

  • Vinay on April 17, 2012, 7:28 GMT

    "The zeroes were given only to show that Bradman has also failed often, more than others."-For failures "Zeores" are not alone, that we all know. So no question of "More than others". There can be any criteria but should be reasonable by common sense e.g. Bradman scored less than 20 in 22 innings out of total of 80 innings.(Around 25%). SRT is 125 out of 311 (Around 40%). Lara is 97 out out 232 (Around 40%). Bradman is distantly far away just like that his batting average. If any statistical analysis is not showing Bradman on top, means either the method is wrong or we are doing unimportant analysis. Best way is, to exclude him as exception and do analysis on others.:-)Still discussion can be open on Bradman like "Lesser opponent teams", "Played mostly against England”,” Not so hi-tech technologies are available" or "not so great fielding standards" etc. His stats are just a pass/fail key for any method of statistical analysis.Take it on lighter vein, not to undermine your hard work:-)

  • FAZIL on April 17, 2012, 7:06 GMT

    Excellent work. It's tough to devise a criteria/ method which perfectly measures consistency. However this process nearly does that. I was going through one of my favorite batsman- S. Ganguly. Though generally termed as an inconsistent player, the statistics proves it wrong to a certain extent :)

  • Vivek on April 17, 2012, 6:35 GMT

    Thanks for your comment Ananth, I was not trying to put one above the other. I just wanted to mention that there were hardly a few players who made 10k runs at such a consistency throughout the world, across 1.5 decades facing quality bowling. If you go by RPI, there are few players who have above 50 - Lara,Sanga,Don, Hammond. Also hardly there would be 10-12 series where Lara went without a century, barring series where he played 1 match..Also he hasnt averaged below 40 not more than 5 times in the above matches....TO me that talks about his consistency. Of course I am a little biased towards him, but I dont completely lack knowledge as you had said. These stats come from heart and notfrom cricinfo as i type. ALso in my original comment, i mentioned "Lara somehow comes to top in every other list you prepare" and not " you bring him to top" . and I am sorry if you felt comment was biased, I agree I have no rights to talk about fellow fans. [[ No problems at all, Vivek. I may be quick on the trigger when I read something but am also equally quick to apologize or forget. Ananth: ]]

  • Vinay on April 17, 2012, 6:33 GMT

    "Consistency is with reference to one's own career numbers"-If really this analysis is on that basis,then no issue at all.Ranatunga also can be better than Bradman.Not sure about Warnie/Kumble who may have considerable good number of runs,but they can be better than some top order batsman.It will be sure different analysis than "Consistency of having good scores".I believe 50+ score in test innings is good score enough without considering match results,conditions etc.He may have ability to score 70+,100+,150+ but still we can consider him as consistent if we agree 50+ is good score.Are you saying,during Bradman days anybody could have scored 50+ easily by saying"Bradman 50 was good as 25 was to Tenfulkar and 15 was to Habibul Bashar".But just checking for zero score,I do not think it sounds logical.If really wanted to show this zero as failure,what is the difference between 0,1,2,3 etc until some reasonable score.In that case,Bradman will still not top that list of more failed innings. [[ The zeroes were given only to show that Bradman has also failed often, more than others. That is all. Where have I used that. Ananth: ]]

  • Ramesh Kumar on April 17, 2012, 5:22 GMT

    Ananth,

    I get your drift on ranking 1,2,3 on innings to measure consistency. But pls. don't bring in % of total runs, 1st/2nd innings weightage etc. They are match contexts and they pad up some inconsistent batsmen. If you come to crease, are you scoring some minimum runs and how consistent you are? That is the question. It will be interesting to arrive at a simple methodology if you don't go by 50s/100s etc. Let us think over it. Also, if you arrive at a methodogy without getting into team share, wins etc, we can extend it to team performance as well.

    Alex...on rating Ponting in ODIs, why are you bringing in Fielding? That way, we can bring in all rounders esp Kallis.IMO, SRT might edge over Ponting, more runs, better average, better strike rate.I agree with you on Richards though I put him along with SRT as the best ODI batsman. IMO, with modern day fielding techniques, Richards would have better strike rate, but lower average, like Sehwag.

  • shrikanthk on April 17, 2012, 5:17 GMT

    In ODI's, SR and RPO are important factors, and McGrath is far better than McKay on these despite averaging the same, as you probably know very well

    He can't be better than McKay on both SR and RPO. That's not possible if they are averaging the same. He's obviously much better than him at RPO. But even that's not exactly comparable because of all the rule-changes in ODIs over the past 6-7 years, with the coming of 20 over powerplays, free hits and what not.

    The fact is that ODIs have ONE overriding objective. To get the crowds in. All rule tweaks are intended to achieve that objective. I've no problems with that. It's an honourable objective. So let's accept that and stop using ODI data to make historical assessments.

  • sandeep on April 17, 2012, 4:58 GMT

    all the players are different in the way they play. Some score more at home and some score more away. Fr example: srilankans play on 3 or 4 stadiums only at home. They score heavily at home. They scored very less comparitively away. Those 10 inngs which u have taken mite have 5 home n 5 away.. those 5 at home make up for 5 away. And should also consider the opposition. Ex: 5 inngs vs ban and 5vs aus mite b overlapped.. Actually its very complicated i guess. And every player has ups n downs in their career.. lara has never scored a century in india n kallis hasnt scored much against aus.. Srt n dravid failed miserably in patches. Kp has been dropped from team twice.. [[ Finally what is your message. 10 innings is sufficiently long enough for a player to fail and recover, score at home and fail away or vice versa, move from dustbowls to greentops and back et al. The point is simple. If a player does not score enough in this 10-innings slice he is inconsistent in a manner which does not contribute to the team at all. If a player scores way above par in this 10-innings slice he is inconsistent, but in a manner which contributes significantly to the teaml. One normally leads to the other. Ananth: ]]

  • Nitin Gautam on April 17, 2012, 4:57 GMT

    Poor guy, he is being paid for the entire season what SRT/MSD get paid PER DAY of IPL.

    I Agree Anantha. But He was touted as "not so good T20 batsman" before IPL auction & that is why he is not getting paid enough. Past glories matters a lot when a corporate invests. See KP, he was auctioned at 1.55mn in 2009 by RCB & bought by DC in 600K in 2012. same goes with YP...let the next auction start in 2014 & Rahane would be star n might be the biggest buy at that time. Anyways, his last 100 against RCB was the best T20 100 I ever saw bcos almost 95% of the shots were top cricketing shots & not mere slogs. He has got talent & he along with pujara should be there in ODI & T20 international team even if means to show the exit door to SRT in ODI & Rohit Sharma in T20. [[ Yes, the only time I really enjoyed a IPL match was to see those quality shots of Rahane. Not the sledge hammer drives of other batsmen but that inside-out six against spin and then the tennis shot for six off the next ball. Ananth: ]]

  • Vinay on April 17, 2012, 4:56 GMT

    10 innings slice is too long. Out of 10, playing 3 to 4 better (30 to 40 percent) innings can maintain the 50+ average. That much big slice always give chance for reasonably good batsman to recover the lean patch especially kind of Lara,Sachin,Ponting etc. Bradman scored 50+ scored in 42 innings out of 80, means more than 50% of his innings where as others are coming only at 30 to 35% of total innings for their good scores. Bradman should be more consistent comparing to any of the batsman in this world. He was failing only once for every other innings to score 50+ score, where as others are failing every 3 innings out of 4 or 2 innings of 3. If Bradman cannot be appeared as more consistent, that method itself something wrong. How much percentage of innings of good scores, should be the basis for consistency. By the way what is the 'core' of being consistent, if not good score? [[ Not really. Obce one considers a certain number of innings, 10 is correct since it covers a long series or two short series. Why this preconcieved notion about Bradman. And where did you get the idea that for Bradman 50+ score was par. Consistency is with reference to one's own career numbers. For Bradman 50 was good as 25 was to Tenfulkar and 15 was to Habibul Bashar. You have not read some recent comments on Bradman. He has 7 zeroes (out of 80). Please compare with 14 out of 311 for SRT, 5 out of 138 for Hutton, 8 out of 286 for Dravid, 16 out of 238 for Ponting, 2 out of 84 for Sutcliffe and so on. I am giving this just to show that Bradman was very consistent was a myth. Recently we got a fantastic insight on Bradman's big scoores and below-100 scores. Pl read earlier comments. Ananth: ]]

  • Nitin Gautam on April 17, 2012, 3:51 GMT

    I Guess ICC should call an immediate meeting to decide future of cricket & should just ban all cricket in Asia because it is anyways insignificant & never considered real cricket. & all the matches should be played in Aus, Eng, & SA on those matches should be unofficial unless atleast 6 out of top 10 bowlers are not playing bcos runs made against lesser bowlers do not deserve anything better than a trash can. Test cricket should only have 3rd & 4th inning as 1st & 2nd dont have any meaning at all. Runs made by winning team only should be published as efforts in a losing cause are anyways not important & BD, NZL, WI, Ind (due to their last year performance) should be just stopped from playing cricket. If all the articles ever written, scorecard ever recorded, analysis ever done have to be kept in a harddisk, im sure it would be lighter by half,analysis would be easy,filters wont be required & world would become free from an enigma called as SRT. NO OFFENCE please [[ Nitin, after a very sane comment, you have decided to go ballistic, like Rahane. Poor guy, he is being paid for the entire season what SRT/MSD get paid PER DAY of IPL. Ananth: ]]

  • Alex on April 17, 2012, 3:41 GMT

    @Ananth: Writing a separate reply on your comment on Bevan. It actually highlights that we should analyze batsmen in ODI's strictly from the batting position standpoint. You had done a super article on it 2 yrs back which showed Hussey's phenomenal record at #5 and #7, and Bevan's at #4 and #6. Bevan was one the greatest ever specialist finishers.

    @shrikanthk & @Ananth: In ODI's, SR and RPO are important factors, and McGrath is far better than McKay on these despite averaging the same, as you probably know very well. I do agree with Ananth on longevity --- it is just that I factor in playing conditions and # years as well while computing it, rather than looking at the bulk aggregates alone. E.g., a hundred in every calendar year 1994-2012 and in every country says a lot more about SRT's outstanding longevity than 18,000 runs in 23 years ... besides Ponting, no one comes even remotely close to him on it.

  • Nitin Gautam on April 17, 2012, 3:41 GMT

    Gerry...Bang on target but isn't it wht HItesh was exactly referring to..Applying filters as per convenience & result came as it was desired. give me one 1 batsman (of course barring DON) who remained numerouno in all countries, against all oppositions, in all conditions, against all bowlers, in all seasons, braving all odds, led his team to victory always & mos imp. no filter can demean him at all. Every player has its shortcomings but proving something n deciding something entirely on them is Hitesh was saying. Tweaking of stats to degrade a player. I dont mean to demean anyone but few filters here & there as per convenience & many cricketing giants would crumble just like that. [[ I am with you, Nitin. In fact I will take a challenge that I can show ANYONE, barring Bradman to have had his weak spot. I thought Sutcliffe might be there. But a casual glance at his career shows that he did not do well against West Indies. Without doing any investigative work, I would venture to say that only SF Barnes might come closest to Bradman in this regard. Ananth: ]]

  • Alex on April 17, 2012, 3:21 GMT

    @Ananth: Thanks for the clarification. We can go back & forth endlessly on such great batsmen. I don't mean Lara & Ponting are 50% better than SRT. It's just that I put SRT only 5% ahead of Lara and within plus/minus 2% of Ponting. Among these 3, Ponting is certainly #1 on consistent brilliance across all playing conditions.

    Also, not many realize that Viv's brilliant ODI record is thanks to the truly great 1979-86 phase. He was really ordinary, at best, in his first 4 and final 3 years. Hence, Ponting (or maybe even SRT) probably deserves the all-time #1 tag in ODI's, esp. given his fielding.

    I well remember choking through SRT's 85 in a Bangalore office: not a soul in our office did any work that day after 2PM. A very important match-winning innings all the same, in stark contrast with Ponting's signature captain-on-a-sinking-deck 100 in the previous match.

  • shrikanthk on April 17, 2012, 2:24 GMT

    I dare say some of the best ODI innings were as great as some of the best test innings

    Ofcourse. I never deny that. Lots of test hundreds are often pedestrian. While several ODI hundreds may be sublime. But that's beside the point.

    We are not discussing knocks on a case-by-case basis here. We are interested in gauging players based on aggregate statistics. When it comes to career statistics, Tests win hands down. ODI stats are full of red herrings in contrast.

    Clint McKay and McGrath are equals in ODIs going by averages. In fact we don't even know what are the right metrics to look at in ODIs. Nobody knows what a good ODI average is. Benchmarks keep changing because of rule tweaks. [[ Shri, I have enough metrics to make sure that, in no analysis of mine would 52 at 21.85 go within a few kilometres of 381 at 22.02. Or 3590 at 50.56 goes within the same few kilometres of 18000 at 44.83. Pl see my response to Alex. Ananth: ]] Consider Tests in contrast. As Aus and WI are discovering in Trinidad, scoring freely at 3+ rpo on a slow turning pitch is just as difficult today as it was in the days of May and Cowdrey! Things don't change much in Tests. The format transcends eras! [[ The last point I agree, Shri. The last two days were a throwback to the tough Test-eras. But the game was no less interesting. And how Australia adopted. The next three days will prove that, rain permitting, Australia would win and 300 would have been a match-winning score. Ananth: ]]

  • shrikanthk on April 16, 2012, 18:47 GMT

    Also, I am not sure why we should not appreciate ODIs and T20s. Even though they may not match Test cricket in terms of skills, temparament, etc. they have their own standards and people enjoy them

    Vimalan: People enjoy them. I enjoy them. In fact I've seen more live ODI cricket than Test cricket in my life (going by minutes watched). I am no ODI hater as Ananth paints me. I love the white ball game.

    But Let's face it. The white ball game is a compromised sport. It is basically an attempt to make cricket more popular. Let's never lose sight of that. The 50 over game is after all a format where Clint McKay, Brett Lee and Glenn McGrath end up with similar bowling averages! Nick Knight, Brian Lara and Mark Waugh end up with similar batting averages!

    Test cricket is about making runs and taking wickets.Simple. The white ball game is about playing percentages. That's a great skill yes. But it's altogether different from the basic cricketing skill of making runs and taking wickets

  • Dinesh on April 16, 2012, 18:27 GMT

    You have made statements without any back-up. Ananth: IN Aus: Tests SRT: 20 1809 AVG: 53.20 RPI: 49 Till 2006 12 1029 AVG: 54.15 RPI: 47 Lara: 19 1469 AVG: 41.97 RPI 41.97 SO Sachin was better here till lara retired.

    IN England: SRT: 17 1575 AVG: 54.31 RPI: 52.5 Till 2007(Included 2007 series as he had a bad series to show he still had better RPI) 13 1302 AVG: 62.00 RPI: 59.18 Lara: 15 1268 AVg: 48.76 RPI: 46.96 Now its upto you do understand it.People wanted stats.I give them in the format(RPI) they wanted.

    P.s: No offences meant.Just wanted to state the stats and show that Lara wasnt better than tendulkar in tough conditions as people pointed out. [[ I am publishing these without verification. That is for the other readers. Ananth: ]]

  • Alex on April 16, 2012, 17:23 GMT

    @Gerry: Based on all your comments, you should simply restrict your sample sets to SRT's failures! Since we know SRT very well, we are acutely aware his deficiencies; however, every batsman has puzzling weaknesses ... why was Lara mediocre in India where he ought to have posted his best records and how come he did well in SA despite obvious discomfort vs fast bowling?

    I am always confused why Ananth puts SRT above everybody in ODI's. Against top opposition in Oz-SA-Eng, SRT was good but rarely that great. As an opener, he is probably the most rounded & consistent in ODI history but that doesn't mean he was better than Lara & Ponting. IMO, Ponting is probably the all-time #1 ODI batsman, or maybe #2 just behind Viv. [[ Alex, in any career-level analysis I do, I give some weight (normally not more than 20-25%) to longevity-related achievements, such as Runs scored or Wickets captured. I will never go off from this. Longevity is often over-emphasized or under-emphasized depending on whose case one pushes. I try and get a middle level approach. One has to be a fool to say that Tendulkar is the greatest because he has scored 18000 runs. Similarly an equally foolish person will say that Bevan is the greatest becaue he averages over 50. There has to be a middle path, governed by objective considerations. As such I see nothing wrong in anointing Tendulkar as the best ODI batsman. Upto about 12000 runs, maybe he found it difficult to offset Richards' explosive and top-quality numbers with the weight of runs scored. But not now. His run tally is enough to overcome anyone. Lara has not played enough top-quality innings and has won but a single title. Ponting would run Tendulkar close, that is all. I would like to remember Tendulkar's ODI career with his innings of 85 against Pakistan at Mohali. Awful innings and a very lucky one. But amongsr the 5 most important innings he has played. But for this innings there was no WC for India. Anything what happened afterwards is a blot on his fine career. Ananth: ]]

  • Vimalan on April 16, 2012, 16:59 GMT

    Ananth,

    I still vividly remember some of Lara's best ODI knocks. However, what I was indicating was Lara's ODI career can not be matched with Sachin's test career. when you had mentioned that Lara had truly outstanding test career and very good ODI career and Sachin had the reverse, atleast thats what I inferred. Also, I am not sure why we should not appreciate ODIs and T20s. Even though they may not match Test cricket in terms of skills, temparament, etc. they have their own standards and people enjoy them. Ultimately, that should be the objective of any professional sport. We should not compare different formats, but while judging a player's overall calibre I don't see any wrong in combining all skills together. Ofcourse its my opinion. Thanks, Vimalan [[ Pl see my response to Vikram. Ananth: ]]

  • Gerry_the_Merry on April 16, 2012, 15:45 GMT

    Hitesh, obviously as the villain of the piece I must answer. My answers will be purely numerical. I will select my criteria, exactly as I want.

    The results of the filter of ODI batsmen in Eng v/s Eng + in Oz v/s OZ + in SA v/s SA exclude batsmen from these three countries, and only include Indian, Pak, SL, NZ and WI batsmen.

    Top Indian batsmen are Azharuddin, Dravid and Gavaskar @ 43, 40, 40. Gavaskar has a strike rate of 53, but Greenidge is 58, Haynes 57, etc. (with poor avg) the low strike rates perhaps typical of the times and the batting position.

    Tendulkar averages 31, with Dhoni, Ganguly and Shastri above him.

    The top batsmen with significant runs are Richards, Chanderpaul, Clive Lloyd, Azharuddin, Attapattu, Dravid.

    http://stats.espncricinfo.com/ci/engine/stats/index.html?class=2;filter=advanced;home_or_away=2;opposition=1;opposition=2;opposition=3;orderby=batting_average;qualmin1=500;qualval1=runs;size=200;team=4;team=5;team=6;team=7;team=8;template=results;type=battin

  • DineshReddy on April 16, 2012, 14:51 GMT

    Ananth:following some comments which talked about tendulkar's ODI runs in australia and england saying they were below par..looking at the scores and records..Lara was better than Sachin in AUstralia and Sachin better than lara in England and both were almost identical in Southafrica. And coming to tests Tendulkar had far better averages and RPI than Lara in both england and Australia(Even after considering last two tours formless).If its only till 2007 he had even better RPI till 2006. and IN SA both had the same averages and RPI. So i would request people to look after the RPI and averages and not just comment for the sake of saying A was better than B and vice versa. Since RPI and averages were the thing in question in this articles.MY comments are purely based on them. [[ You have made statements without any back-up. Ananth: ]]

  • Vikram on April 16, 2012, 14:40 GMT

    Hitesh: I believe this is exactly the kind of focus which has meant that SRT has been pulled down by others. when you make him more than he is, it is easy to pull him down. He is good, if you make him a god, the extra o is what is left and that's what other people see him as. I am a huge fan of SRT, but that doesn't mean others are better or worse. And I would largely support Ananth's statement. No one has run him down for something which he can't be questioned about. Ananth, while it's ok to say that test and ODI are different, the fact that the player entered both games means that it's the entire package which needs to be evaluated. I will ignore T20s for the moment, simply because this game is still in its infancy and hasn't had enough internationals to compare. I would be really surprised if a player has as violent a view as some of the people here. There was demand but these players were the suppliers so they need to be evaluated by how they performed across both formats. [[ Vikram, Shri's views are his own and he is entitled to have those. Mine are exactly similar to yours. I think ODIs are an important segment of the game and should not be ignored. I still think T20s are nowhere near enough to be included for serious consideration while IPL (unlike Harsha, I differentiate between the two) should never enter any serious analysis. Analyis on IPL should be within its own sphere. Ananth: ]]

  • Anand Vanchi on April 16, 2012, 13:08 GMT

    Dear Ananth.. An excellent analysis. I have a suggestion here. Instead of making the individual slices of 10 innings, how would it be if you divide each players career into 10 Slices. So for example a player who had played 100 innings would have 10 slices and another player who has played 200 innings would have 20 slices. I know this would mean a lot more amount of work but it could ensure the factoring of the longevity of individual players as well. I would be happy to know your views on the same. Thanks. [[ Let us see how the others repond. Problem is that for the 3000+ runs batsman population the range will be 7 to 31. I am not sure what will be the impact for Hassett or for Tendulkar. Being consistent over 7 innings as against being consistent over 31 innings. The later, I would assume, would be easier. Ananth: ]]

  • Manu on April 16, 2012, 12:40 GMT

    it gives you an idea of the form players i guess...ponting/dravid/tendulkar.. prolly more of the form players that rode their form longer..but also dropped form for longer periods...lara/KP are more "inspired" players..likely to play an inspired long innings at any time...but also likely to not...the former set are simply likely to score in bunches..and then slump in bunches as well.. making them consistently good..or consistently bad

  • Manu on April 16, 2012, 12:35 GMT

    umm..wouldnt it be simpler to see the percentage of times a batsman scores within 15% of a 2 innings test (since its one match...both innings should count as as single entity..average times 2) so with a batsman with a 50 average....we would look at the percentage of times he has scored between 85-100 in a test.. (why 15%.. i say why 10 innings :p) ..so ones that have a high percentage..are more consistent..having said that...i'll take an inconsistent 45..over a consistent 55 quite easily in 98% of cases... a half century in every innings is great..but most times you need that century to get you the W..who wants sachin/ponting/dravid to score 55 in every game...i'd rather be disappointed by his "low" scores..just to see those long innings...leave the pretty fifties for the dhonis of the world [[ Intriguing idea. First go off average. Use RpI. Then only will you get the right figure. Be assured it will work very well. If a batsman's RpI is 50, why 85-100. Should not that be 85-115. I have no problem with 15%, but should be either way. This will make sense if you are measuring consistency. But if yu are measuring minimum contributions, then above 85 is what you are looking at. Ananth: ]]

  • Hitesh on April 16, 2012, 12:35 GMT

    "If someone said "ABC has not performed well in Australia" don't counter by saying "XYZ did not perform well in India". That is counter-productive."

    I hope I didnt do that at least... ;) [[ No, you did not say that. But the temptation is there always. Ananth: ]]

  • Hitesh on April 16, 2012, 12:19 GMT

    Hi Ananth,

    Thanks for your prompt reply... I do agree with wat u say, but the way stats are being tweaked by the fellow readers is wat compelled me to write the statement that thr is a bias... (I hope u do agree to that at least) I do not say they are using foul language or have written anything bad... But they have tweaked the stats to their convenience...

    This is where I felt it was biased and you didn't say anything... Anyways, my apologies... I do understand that it is an individual view and u cannot be held responsible... :) [[ Hitesh, I am past the normal UK retirement age. The mind is alert but the body generally protests. A totally arm-chair sports enthusiast, I have acquired a slew of sports-related afflictions: Tennis elbow, Rotator-cuff tear et al. The point is that I may not be able to verify all stats provided and cannot take a guess of the intentions. If the comment does not violate the basic tenets of this blogspace, I publish. If you feel that a stat has been wrongly used, please feel free to put in a correction. It could relate to the numbers or the conclusions However avoid one thing. If someone said "ABC has not performed well in Australia" don't counter by saying "XYZ did not perform well in India". That is counter-productive. Ananth: ]]

  • Hitesh on April 16, 2012, 11:27 GMT

    Hi Ananth,

    Firstly thanks for this article but to me it looks a bit flawed in the sense that it takes only brackets of 10 innings (1-10,11-20 and so on)... Like many readers pointed out, I feel it shud be an over lapping one which will show the most consistent... This yardstick may be harsh on few batsman... I know u have suggested that this may be diff for u... [[ More than being difficulty in implentation for me, it has already been done in a slightly different manner and it would take the article beyond the comprehension of the normal reader. And please remember that it is 1-10, 11-20, 21-30... for ALL batsmen, irrespective of the years they played in. In reality that is the problem. In the 1910s, it might have taken 3 years to complete 10 innings, in the 1960s, 1 year and now, probably 2 years. However we seem to be comparing batsmen of same age, so that should not matter. Ananth: ]] Also, I am not sure if you have seen few of the comments here which are in total distaste and targetted towards a particular bastman like the ones from Srikanthk, Gerry the merry and Swarzi.... If I remember in one of ur response u had pointed out to a reader that u wud remove any comment which is derogatory to any player and puts one over the other.... I am not sure if you have applied the logic same here... Sorry to say but I can note some bias here on ur part as well... N yes I am a Sachin fan just in case you are going to point out but I am not here to criticize the other batsman... Hope u dont take it personally... [[ Saying that A is better than B is not wrong, not derogatory to B, not at all in bad taste and that is what everyone does. Saying that ABC averaged only 30 against Australia in Australia is not wrong. Saying that ABC failed in India is not wrong. Saying that ABC let his team down by scoring slowly is not wrong. Saying that ABC should have retired (or not retired) is not wrong. Saying that a young batsman is kept out is not wrong. Saying that XYZ is letting his team down by playing in IPL is not wrong, if the circumstances are considered. Saying that "you make money by doing trash", "you south indian *******", "xyz is a fourth rate batsmen", "abc (reader) is a moron", "xyz should be kicked out", "you are biased", "you have written this article solely to show A is better than B" et al are totally and completely wrong and will not be allowed here. Now please point out one statement of mine which violates the above mentioned guidelines. I will apologize and remove that line. Also point out one statement made by a reader which violates these guidelines, I will remove the sentence and apologize on behalf of the reader. Ananth: ]]

  • Gerry_the_Merry on April 16, 2012, 11:26 GMT

    Shrikanthk, good stuff. well written. Also, for bowlers there are some equally significant distortion factors. 1) A Joel Garner between the 45th and 50th over is a completely different proposition than in the first 10 overs. 2) The stage of the innings at which a top batsman gets out is also very significant. Getting Tendulkar out for 10 is a highly deserved wicket. Getting him out for 135 means that he was gunning for quick runs towards the end in slog overs. Hence measuring wicket quality by the batsman dismissed cannot always be taken as a reliable indicator. 3) the importance of a match varies too much. Hopefully, over time, people will stop mixing up ODI with tests. Till then, one must tolerate.

  • Sudarshan on April 16, 2012, 10:39 GMT

    100 First class hundreds is a elite club dominated by those who played English county cricket and a few others (Bradman, Zaheer etc.) Unfortunately there is no Indian in that list (Probably the closest is Gavaskar or SRT again) as you cannot get more than a couple of hundreds year after year in the Ranji Trophy as the number of games are very less. Let us celebrate Sachin's achievement and also similarly comprehend the vast gulf between Murali and the others. Barnes was great .. but did not have Murali's longevity (or for that matter the average longevity of that era).

  • Ramesh Kumar on April 16, 2012, 10:34 GMT

    Ananth,

    one other alternate view on consistency could be their meaningful contribution when they come to crease. A simple one could be their scores above 50 in slices of ten tests. Insead of comparing with their av.performance, we can look at their mininum meaningful performances across their career. Because in our mind, we associate consistent decent performances rather than deviation from their average levels. What do you say? [[ Once we open that door, Ramesh, the related problems will come in. 50 would be deemed arbitrary. Context will loom in a big way. Hussey's 32 vs Watson's 52. But I like your idea in that, over 10 Tests, how many times has he put his hand up. Makes sense. I would say we must think about further. Why an arbitrary 50. It should be a simple algorithm to assign a "hand-up" value of 1 to 3, based on a simple easy-to-use methodology and then total these. In the last Test, 3 could be assigned to Clarke's 73, Harris's 68, Lyon's 40 and Watson's 52. 2 could be assigned to Hussey's 32 and 48, Warner's 42. 1 to a few other innings.You get the drift. Ananth: ]] Shrikanthk...your views on ODIs are extreme. ODIs have brought new challenges to players. Great ODI players who do not have good test record are definitely looked down. Though may not be in the same scale, if you compare two greats of modern era with not much to seperate in tests, if there is a marked difference in ODIs, I'd tend to use that as differentiator. I suppose it is ok as long as one does not trash the other player. [[ It is an open secret that Shri has no great opinion of ODIs. However he brings to the blogspace a wonderful insight into FCs/Test cricket and we are all enriched by that. It is the same way you guys have to live with my hardline and disciplinarian approach to comments in this blogspace and my obsessive dislike of the greedy attitudes of IPL, despite the often top class cricket which goes on there. Ananth: ]]

  • Sudarshan on April 16, 2012, 10:33 GMT

    Great work as usual Ananth. The mind boggles at the amount of information that you take pains to present in simple, unambiguous fashion. And this, week after week, bringing ever new facets of analysis to the fore. Looking at your meticulous responses coming at all times of the day, I wonder do you sleep at all??

    Players who have succeeded in Test have generally done well in ODIs except during the early era of ODIs - Gavaskar etc. Does form and consistency in one format spill over to the other format. One needs to look closely. A few months/weeks ago, Virat Kohli was in raging form in Australia and in the Asia Cup but has struggled in the IPL immediately after.

    There are many who object to adding up the centuries SRT has scored in ODIs and Tests under the name Intl. centuries. This record is comparable to Bob Beamon's leap into the next century (looking closer home Murali's 5 and 10 wicket hauls in tests are in the same league - probably a notch higher but rarely celebrated) ...Contd

  • Ranga on April 16, 2012, 6:53 GMT

    Some other interesting facets brought about some forgotten players of the recent past, like Saeed Anwar, who in my opinion could have been a worthy match for SRT/BCL in their generation. I dont know why he faded into oblivion, despite having healthy averages, strikerates and a pleasant game in both forms of the game. He had better records when Pak toured. May be he didnt have the defining knocks in test arena as he had in ODIs. He was one more legend who never managed to become one. Kapil Dev is a surprise, at 74%, given that he gave an impression of getting out anytime. Botham had similar figures, to a lesser extent. 2 surprises! And Eng threw up more 60+ names than any country, showing how they played as a team. But sadly, not many from post 1990 generation, which tends to show how people have shorter careers (may be putting SRT's longevity in perspective) [[ Both Kapil and Botham had the benefit of having lower career RpI figures making it easier not to go below 67%. To go below two-thirds of 285 (only 190) and 320 (210) was not that easy for these genuine batsmen. Ananth: ]]

  • Ranga on April 16, 2012, 6:33 GMT

    My Goodness!!!! 162 names, and still just 2 names matter!!! One - A retired genius and the other, a tired genius!!! But as much as Lara's consistency amazed me, I was amazed at the career slices given by Ananth in response to Kiran, SRT had a slice high of 899 while BCL's was just 801. SRT's slice low is lower than BCL. So in a range, the extremities are more stark with SRT than with BCL. SRT could not score 300s and 400s but he could maintain his high scoring form for longer periods than BCL - may be, read as consistency by most people. I simply dont see reason in comparing the two, getting into an all-time XI etc(I still see this alltime XI as a marketing gimmick). Who is better? I think the answer is in the spirit of the article. I would say, SRT of 2008 Aus was better than SRT of 2012 Aus. BCL of 2002 Aus was better than BCL of 1993 Aus. BCL v Aus is better than BCL v Ind. Ponting v Eng was better than Ponting v Ind. Sehwag v Pak was better than Sehwag v Eng. [[ Ha! a variation of my pet theory of peer comparisons, with a stricter definition of "peer". Ananth: ]]

  • Alex on April 16, 2012, 5:55 GMT

    @shrikanthk & @Ananth: I too remember the Don's observation. He actually made that comment in '99 during India's disastrous Oz tour. By that time, Lara had already played his epic series vs Oz. The Don put SRT ahead of Lara on the grounds that he was technically better & compact and although he took fewer risks than Lara, "at the end of the day, he will score as many runs". Bradman was rarely wrong in his assessments but he could afford to voice those straight up --- in contrast, the utterings of people like Harsha Bhogale are governed by their assessment of the market, regardless of what they may actually believe in. [[ Yes, I was extremely disappointed with harsha's recent article. He was writing as an IPL spokesman, but without using the word IPL once in the article. He had used T20 interchangeably for T20-Intls and IPL. Ananth: ]] I dare say some of the best ODI innings were as great as some of the best test innings: e.g., Waugh's spine-breaker 100 vs SA. As Ananth has said elsewhere, ODI is a different ball-game. However, that is not like apples and orangutans ... it's more like novels and short/long stories. Lara was a Tolstoy whereas SRT is more like Chekov/Twain/O'Henry.

  • A. Khan on April 16, 2012, 5:01 GMT

    But all things taken into account, Tendulkar should be right up there in the ODI game, in the top place, comfortably ahead of the next one. Lara might not even be in the top-10. But this was just a comparison so I made it.

    Ananth, Things are not as simple as they look, if you take huge difference in runs scored then your statement is pretty much valid. But if you only concentrate on the quality then the difference is not as much as it looks on the plane paper. Lara's ODI career took a nosedive since 2004. Even at that point, to the naked eyes, Sachin's record was far superior but if you go deep into it, did a pitch/bowling analysis, there is no doubt who will come atop. Even now if you did the similar analysis to what you did for tests, there won't be as big a difference. And during 90's, it was Lara all the way, mostly coz he did equally well all around the world and not just in one part of the world. He certainly is one of the top ten ODIs batsmen. [[ Good, A bhai. I am now taking this approach that I provoke you guys to do the talking (or more appropriately, writing). I am not sure about Lara's place since we do not have proper data. When, where, at what innings point were the innings played. Off who. What was the challenge in front. So may relevant factors. Ananth: ]]

  • shrikanthk on April 16, 2012, 4:40 GMT

    most of these comments/observations are made by these great ex-players, considering Tests and ODIs together. They may not explicitly say so. But it would be clear to allthat they are doing it. And, combining Tests and ODIs, Tendulkar is way ahead of any one else, including Lara

    Perceptive ex-players like Ian Chappell or even Boycott will NEVER combine Test and ODI records while judging a player. That's the sort of thing you expect from B-School educated pundits. Eg: Harsha Bhogle.

    If one considers Tests alone (as one should), there is not a lot to choose between Tendulkar and Lara. Though it might be argued that Lara was a greater match winner. That's it.

    DGB got it spot on in his 1996 interview to Ray Martin. He said something to this effect - "Tendulkar has a very tight technique. His defence is very sound indeed. Lara is the more aggressive of the two and takes more chances. But at the end of the day, both will be just as successful". Spot on.

  • shrikanthk on April 16, 2012, 4:04 GMT

    Another reason why ODI stats are pointless is because it makes no sense to compare the averages of 2 people who bat at different positions. Tendulkar at No 4 is an altogether different story from Tendulkar at No 1/2.

    Which is why it is pointless to compare Tendulkar with say Mahela or a Steve Waugh in ODIs.

    It is only in the Indian subcontinent that we have sanctified this watered down version of cricket. Everywhere else in the world people enjoy ODIs but leave it at that. They continue to judge players based on test records.

    Please don't mistake this for snobbery or condescension. It is pure common-sense. The moment we start saying - "Hey. Sachin is better than Lara because he did a lot better than Lara in ODIs.", we lower the standard of debate. What next? Hussey is better than all of them because he is brilliant even at T20s as well?

    To draw a mathematical analogy, comparing X's test record with Y's ODI record is a bit like comparing Isaac Newton with Shakuntala Devi!!

  • shrikanthk on April 16, 2012, 3:34 GMT

    look at where Kohli is today: 3590 at 50.56. But all things taken into account, Tendulkar should be right up there in the ODI game, in the top place,

    This should help people see ODIs for what they are. A money spinner that makes cricket a richer sport than it would otherwise be.

    You just said what needs to be said. Even Kohli averages 51 in ODIs. Dhoni averages roughly the same. So do Hussey and Bevan. Just goes to show how pointless it is to rate players based on ODIs. In a world where people play only ODIs and T20s(I'm sure that is an eventuality that will be LOVED by marketers like Bhogle), a player as gifted as Lara would be just another top player. Not a legend.

    One loses count of the number of players averaging 40+ in ODIs with SRs of 75+. The shorter the format the more difficult it is to separate wheat from chaff. It levels the playing field.

    Let's accept ODIs for what they are. Accessible commercial entertainment. That's it.

  • Alex on April 16, 2012, 3:31 GMT

    @Ananth: If I am not mistaken, Lara was placed 6th in your ODI ratings article published 2 yrs back. Time to revise it, perhaps. At any rate, I think ODI ratings are best done position-wise: openers (or top 3), 3-5, and 6-7 ... you had done an analysis on this as well.

    If we follow a player live, our opinions are biased by how he finishes. In ODI's, Lara averaged a still decent 35 in his final 4 years but the 100's were all but gone. Until then, i.e., until age 34, he averaged 42 in ODI's. During '93-'97, he was arguably the world's best batsman in ODI's. I will take him in my dream team any day, proviso he bats as an opener. Of course, if we consider the fielding aspect as well, Ponting is the greatest ever ODI batsman, except perhaps Viv.

  • Vikram on April 16, 2012, 3:00 GMT

    While I find your analysis very interesting, Ananth, I find the responses equally insightful. You have said as much through your various responses. People try to push the analysis beyond what it can say to fit their pre-conceptions. The other interesting thing to watch is how most people are never ready to define their criteria before they come to a conclusion or if they do, they never push the analysis far enough to examine the evidence against those criteria. That is one of the reasons why I love this blog. You put the criteria up front. One example is this debate of greatness. To me, greatness is when someone raises his "average" level of performance to a higher level than those around. Therefore, VVS or VS, despite their near mythical innings will never be great but RD and SRT will be. And here I had also fallen in the trap of not having examined the evidence with enough depth. Because, as per this analysis, VVS is more consistent than RD and SRT. So maybe it should be VVS vs BCL. [[ Vikram, I have just finished a book on "Fuzzy Logic" which I had bought some time back but came upon it recently. A real eye-opener. There is a statement there Lotfi says "The logic is not fuzzy. It is the ability to undertstand the situations which are fuzzy and apply precise problem-solving methods". "Greatness" is quite fuzzy. It varies from situation to situation and person to person. And there are different types of greatness. And in IPL matches, if you read the over-paid, hyper-active commentators, especially the Morrison-types, every innings is great and every shot is great. So greatness in IPLs is a day-to-day occurence so much so they find it difficult to commentate effectively on the 8 overs of Rajasthan yesterday. greater than great or super-great or truly great or some such trash. Tomorrow if some other team crosses 100 in the last 5 overs, what then. The lesson is to use the term great carefully . Ananth: ]]

  • Gerry_the_Merry on April 16, 2012, 2:56 GMT

    Ananth, surprised to see so many comments extolling Tendulkar's One Day virtues. In Eng v/s Eng + in Aus v/s Aus + in SA v/s SA he has played around 135 matches, 4 centuries (one of them was a masterly innings of 117* in a final, just to prevent myself from being clawed in the throat by Tendulkar fans), avg around 30. Many Indian batsmen are ahead of him. These were the toughest conditions (big grounds + balanced wickets) to make runs in during the last 20 years. Perhaps earlier West Indies could have been included but in the last 20 years not much chance. I cannot see how such a glaring under-performance can be glossed over. It is however true that he has compensated on smaller Indian grounds and flatter wickets.

  • Vimalan on April 15, 2012, 19:44 GMT

    @Swarzi, How about the elite panel of Don, Benaud, Warne, Donald plus many others who on record have considered Sachin as the best batsman of the modern era ? Also, Ian Chappell from the same panel has also rated Sachin ahead of Lara recently. And you'll also hear from all others in that panel when the great man finally hangs up his boot. @Ananth, Sachin has had a truly amazing test career along with an outstanding ODI career. I am quite surprised that you compare Lara's ODI career with Sachin's Test career. anyway that's your opinion, no one can change that. [[ Vimalan, first please accept my thanks for the measured, constructive and venom-free remark. It makes for good-natured discussions with no malice. But you have always been like that. Let me make a few short points. I think, in this blogspace at least, Bradman's statement about Tendulkar should be given a miss. After all he only said "Tendulkar reminds me of myself. He is the nearest to me in technique" (may not be the exact words). That is a reference to the technique, compact play, the delaying of forward-back decision to the last second et al. That is all. I agree that people will have divergent views. Pl see my response to Alex. Finally you should not say that I should not talk about Tendulkar and Lara in the same sentence when talking about ODIs or someting similar. After all we are talking about one who has scored 10405 runs at 40.49. And if average comes in, then one cannot forget Richards, with 47.00. And, much as I dislike his abrasive in-the-face behaviour, look at where Kohli is today: 3590 at 50.56. But all things taken into account, Tendulkar should be right up there in the ODI game, in the top place, comfortably ahead of the next one. Lara might not even be in the top-10. But this was just a comparison so I made it. Ananth: ]]

  • Alex on April 15, 2012, 18:03 GMT

    @Ananth: Absolutely - and even in his 3-year lull of '96-'98 in tests, Lara was phenomenal in ODI's, averaging 53. What _is_ true about his inconsistency is that, except for '94 (and '04 which was saved by 400*), his record in even years is never great but his record in odd years ('95, '99, '01, '03, '05) is brilliant ... he really should have played in '07!!

    @Swarzi: The last year has taken some sheen off SRT's career but in his most recent article last month, Ian Chappell put him ahead of Lara ... so, there you go!! It is tough to separate these two. SRT was ahead of Lara till the 9000-run milestone and they were neck-to-neck at the 10000-run milestone. Then, Lara surged ahead during SRT's dismal injury-ridden 2004-06 phase. Since big innings are not SRT's strength, he can never feature high on any metric that relies on one-off innings/test/series. His real strength is mental in that he keeps coming back. As Viv says, Lara was more gifted but SRT contributed better to his team. [[ Alex, most of these comments/observations are made by these great ex-players, considering Tests and ODIs together. They may not explicitly say so. But it would be clear to allthat they are doing it. And, combining Tests and ODIs, Tendulkar is way ahead of any one else, including Lara. Ananth: ]]

  • Ram on April 15, 2012, 16:30 GMT

    Nice one. Surprised to Lara at the top here. It presents a nice and indepth alternative to this one - http://www.espncricinfo.com/magazine/content/story/245575.html by S.Rajesh [[ Quite old. 6 years back. But nice of you to remember that. Ananth: ]]

  • Alex on April 15, 2012, 15:43 GMT

    @Ananth: As per your additional table, plenty of batsmen had their final 10-innings SPF of exactly 0.67. Dravid's final 10-innings SPF was, by far, the second worst of his career at 0.42. [[ That is a mirage. That is the adjustment I have talked about for the last slice, If the last slice has fewer than 5 innings, I am limiting the SPFs to between 0.67 and 1.33 so that they fall into the inner groups. Otherwise it is unfair to the players who have played 311. 232, 121, 103 innings et al. In Tendulkar's case, the only innings in the last slice is 13 and this would have pushed the slice to A. He is now protected with 0.67. Boycott's last three innings were 34, 18 and 6 and he is protected with 0.67. Ananth: ]] The myth of Lara's inconsistency is primarily because his SPF was almost consistently less than 1 over innings 60 through 130, which many consider the peak of a batsman's career. Lara compensated for this 70-innings lull through an SPF that exceeded 1 over the next 60 innings or so (this came at the cost of sub-par ODI performances), giving him a low SD across his career. I think outside India, and even in Indian players, there is little doubt that Lara was the best batsman of the last 20, or even 40, yrs (see http://www.youtube.com/watch?v=ONXZJ_8xSac&feature=relmfu). [[ Alex, even during this period, Lara's numbers are 479(0.93), 309(0.60), 401(0.78), 385(0.75), 585(1.14), 334(0.65), 340(0.66). The total is 2833 over 70 innings, a RpI of 40.5. Ananth: ]]

  • SWARZI on April 15, 2012, 15:27 GMT

    Ananth, I don't know why when you sit down and produce work of this excellent quality, there is this group of Tendulkar vs Lara fanatics who always think that you do it purposely to show the comparison between those two great men! Nonsense! Your work reveals so much more than that myopic view in cricket to us, that I think those few who are suffering from cricket myopia should cry shame on themselves. As far as the Lara vs Tendulkar issue is concerned, the verdict was out since 2006 after Lara had retired. In November of that said year, in a ESPN/Cricinfo Round Table discussion among one of the most elite panels of international cricket experts you could find: Tony Greig, Ian Chappell, Sanjay Manjrekar, John Wright, and Ravi Shastri, (note the presence of TWO leading Indian experts), in which they talked about 'How Good is the Modern Batsman’, 'THE ENTIRE PANEL' (with the exception of John Wright who abstained from the vote) agreed that when the myriad of factors pertaining to great batsmanship are taken into consideration, Brian Lara is the best of the modern greats - Chappell reiterated the said sentiment in one of his articles quite recently. In addition, every qualitative statistical batting analysis that has been carried out by all the great cricket statisticians including you, have shown that Lara is second only to Sir Don Bradman where great quality batting is concerned. Both men have executed some of the most extraordinary and awesome performances not ever to be seen in test cricket again! On the other hand, from a qualitative perspective, when the statistics are done, Sachin Tendulkar does not show up on the radar most of the times (not that he does not have great quality). But he has been way out front in terms of quantitative achievements. And, his quantitative accomplishments are more related to performances associated with his significantly long career (which is the most awesome aspect of his greatness) rather than the unimaginable innings that Bradman or Lara played with such great consistency throughout their careers - I am talking about very very long innings of prolonged dominance, skill and excitement that only they could have done - they never bored the spectator during their long careers. Tendulkar's achievements are very important, but playing so many more test matches, batting in so many more test innings, facing so many more overs and balls, etc, than either Bradman or Lara should result in 'THE OBVIOUS' for a man of his calibre: more runs, more hundreds, etc. Remember that 'Lara owned all these records before he retired - even though he had plated less test matches, batted in less innings and faced less overs and balls, etc than Tendulkar'. But what you do Ananth, is to use statistics to bring intellectual transparency and real meaning to the records of our greats. I think that we should all just commend you for the onerous task that you have taken upon yourself - especially when we look at the unbiased way that you do it. And, some of us should stop manufacturing the unwarranted scepticims that we invoke, when we are dissatisfied by your reveleation of the truth. [[ Oh! I had to take multiple deep breaths while reading the long piece. Your fingers must be crying out for some rest. Many thanks for the kind words. I think Tendulkar has had a truly outstanding ODI career and a very good Test career. In Lara's case, it is the other way around. On balance, I would rate Lara ahead of Tendulkar in Test cricket and Tendulkar ahead of Lara in ODI cricket. That is all. But I would stop there. Finally both have been truly great players and one should enjoy reading about, admiring and watching their performances, with one still active, and stop there.. Ananth: ]]

  • njr1330 on April 15, 2012, 15:01 GMT

    Can't say I even begin to get my head around the statistical stuff!! ...nor am I surprised by Greig's consistency...but the fact that KP is one of the most consistent batsmen, is truly stunning news and deserves a wider audience.

  • Anand on April 15, 2012, 14:40 GMT

    I hope I understood correctly that you looked for consistency across innings. How about looking for consistency across tests, i.e., consider the two innings of a test together as one entity and then recompute the scores? [[ Anand, I think you have not read the two article(s) completely. For Bowling analysis I had used 5 Tests since the bowlers get opportunities in both innings. However for Batting analysis I have used 10-innings and the reasons have been explained right at the top of the article. Ananth: ]]

  • Som on April 15, 2012, 7:55 GMT

    Still recollect those days of my childhood with great fondness, when Allan Border came to the crease to bat. He had this sure footedness and assuredness in a very non chalant way, which differed much from the likes of other fighters and saviors like S Waugh, Chanderpaul, Laxman. His pedigree showed. When they were selecting Cricinfo All time XI for Australia and included Border over Ponting, there was much hue and cry. But I had found it a little bizarre. Ananth, your analysis vindicates that feeling. Thanks much. And believe me, I had felt, he deserved to be a strong contender for the combined All time XI, and should have narrowly be beaten by SRT (owing to the sheer volume, tenure, comprehensiveness and that irresistible X factor). Don't know if the jury had felt that way.

  • wake_up_india on April 15, 2012, 4:27 GMT

    Dissatisfied with the average as an adequate metric of a batsman's true quality, I have performed some analyses using other statistics. The results are quite surprising.

    Bradman Net=6996; Median=56.5; Prob(<10)=0.18; Prob(>100)=0.36 Tendulkar Net=15470; Median=34; Prob(<10)=0.23; Prob(>100)=0.16 Sehwag Net=7821; Median=31; Prob(<10)=0.24; Prob(>100)=0.13 Dravid Net=13288; Median=33; Prob(<10)=0.24; Prob(>100)=0.13 Ponting Net=13200; Median=31; Prob(<10)=0.24; Prob(>100)=0.15

    Note that the median is the score the batsman failed to exceed 50% of time. Isn't is amazing how close these statistics are for all of these very different types of players, except Bradman who stands in a class by himself. [[ Your nice analysis would be complete if you did the same for Kallis & Lara and added to the table. Ananth: ]]

  • shrikanthk on April 15, 2012, 3:12 GMT

    ""Don did suffer a slump in form ( albeit brief ) and was contemplating retirement but had the benefit of the doubt or rather benefitted from an error by the Umpire and went on to score a double century and the rest , as they say is history"" You are referring to his first test innings post War. That was at the fag end of his career. He was adjudged not out on 28 (a debatable decision) and went on to score 180 odd. It was the Brisbane test of 1946-47. [[ No third umpire or Hot spot (for that matter there are no "Hot spots" even today in many matches). Anyhow there are differeing versions on Ikin's xcatch. So the benefit of doubt, rightly, went to the batsman. Fingleton and Hammond, no friends of Bradman, might have made disparaging statements. However there was a doubt. Shri, I have an interesting photograph in the pictorial Ashes book, edited by David Frith. Around the time when Ikin caught the ball, Bradman is nonchalantly looking forward. His head is not turned back. The cynic might say he did it deliberately. The same cynic who says that England's bowling in the 1930s was ordinary. I have scanned the picture and uploaded the same. The link is given below. http://dl.dropbox.com/u/39210851/Bradman_Catch.jpg Ananth: ]] So that wasn't a form slump. It was a simple question of whether Don wanted to carry on post World War II. Even if he had retired after that innings he would still have over 5000 test runs with an average of 90+. [[ If we assume that Bradman was given out at 28 (and that Australia still had enough firepower to win by an innings: after all they won by inns & 332 runs), Bradman would have finished with 5121 runs with an average of 96.62. A few metres below, but still a few kilo-metres ahead of the next batsman. Ananth: ]] The only real form slump he had was in the first half of the English season of 1934. He failed in the first 3 tests of the series and wasn't very consistent in the first-class games either. He set it right with a 309 in the 4th test and 244 in the 5th test. [[ The scores Shri is referring to are 29, 25, 36, 13 and 30: a total of 187 in 5 innings. Ananth: ]]

  • Sudhi Sant on April 15, 2012, 3:03 GMT

    Ananth

    Wonderful work and analysis. Your patience and objective approach is highly commendable. The reason I say this is the results throw up some surprises and that is the best part. Have you thought about publishing this in a research journal (statistical analysis)?

    Cheers

    Sudhi [[ The problem, Sudhi, is that my work gets outdated faster than the current fashion trends. The tables I prepared last year are not valid today. However the ideas are still valid. Ananth: ]]

  • rizwan on April 15, 2012, 1:10 GMT

    Initially I found it strange that Border has a lower SD than Bradman and recalled the old chestnut , lies , damn lies and statistics. But , on closer inspection found that , the reason Border 'beat ' Bradman is because the Don had set such high standards from the start of his career and his average reflected that throughout his career . I recall reading that the Don did suffer a slump in form ( albeit brief ) and was contemplating retirement but had the benefit of the doubt or rather benefitted from an error by the Umpire and went on to score a double century and the rest , as they say is history . [[ As I have pointed out to Boll & Som, I think there is enough in these specific insights to do a piece on the reasons why Bradman was great. Boll has pointed out that he had more than his share of failures. Ananth: ]]

  • Som on April 14, 2012, 23:03 GMT

    Cont...

    --And then the question of under 100 scores and RPI's in such innings. As you pointed out, would love to see an article on that. But here is what I think, when one looks at Don's average and compares it with others (30+) range. Don played a very small % of innings where he scored under 100, compared to others who played many such innings. Since Don had an average of 99.94 and an average of 50, is greatness...almost everyone else's under 100 performances should compare with Don's under 200 performances...and then we can see how far ahead he was. [[ Som,. once I open that door, I will go deep to see what really clicked for Bradman. On the surface it is a combination of 186 (average hundred valee) AND 29/80. One by itself would not be enough. (e-g) Sehwag (176) but 22/131. The other side is Headley: only 148 but 10/40. Let us have fun. Ananth: ]] --When someone says, I am a student of the game, I often wonder, what it means...but there is some insight here. Sanga, Hayden are clearly students of the game...Lara was not.

  • Som on April 14, 2012, 22:52 GMT

    Ananth...This is a great piece of work. A few observations though: --A curve fitting to see how on an average great batsmen across generations ramp up their average (or RPI) and how they eventually decline will bring in a greater perspective to this discussion. --While consistency is definitely brought out in this piece, but more than that, it separates players who debuted with all their prowess and managed to keep their standards at that level from others who had a more traditional rise, plateau and decline. In this analysis, there are some whose RPI was 40 or so, and a genius like Lara who crossed 50. Believe me, this article, even though without any such intent, just shows how great Lara was. Cont... [[ The comi-tragedy is that one misguided reader accused that I had used RpI to push L over T. He does not realize that it was T who benefited from this change. If I had used Batting average instead, T's higher batting average would have meant a GREATER variance for those below-par perfiormances. Ananth: ]]

  • IPSY on April 14, 2012, 16:36 GMT

    Anantha, your patience is unparalleled! If you start to yield to the wishes of people like 'gaurav banodah', whom I suspect wants you to do an an analysis whose findings synchonizes with his own subjective views of his favourite player, your work will never be completed! You have almost exhausted all the statistical scenarios to present a comparison of all the batsmen who should be assessed on this particular subject of 'batsman consistency' - and, everybody including 'gaurav banodah' agrees that it is an excellent piece of work; but, it would appear that the findings brings him some sort of disappointment with respect to a particular player of his preference; so, he wants you to generate some bias statistical acrobatics in order to make a satisfactory point. However, your original challenge to him/her is a good one - the data is there on the Excel spread sheet, and he could do it to satisfisy him/herself. [[ I have recently started sending most of my back-up tables in the form of Excel sheets to allow readers to do their own analysis. What is needed is only basic Excel knowledge. I have no problem with gaurav. He is courteous and at least does not abuse me, as one or two others recently have started doing. That too one in Hindi, which I understand only 10%. Ananth: ]]

  • SWARZI on April 14, 2012, 15:43 GMT

    Ananth, I wonder where you get the time to manufacture this excellent quality of work so frequently - congratulations on another super presentation. You may remember that I once reminded you that sweeping statements cannot be made about the consistency of batsmen by just looking at their records at face value. I did agree with one of your contributors that Standard Deviation should be used to reveal the facts pertaining to consistency. You've proven once again that nothing magnifies the truth better than honest scientific analysis. [[ Thank you, Swarzi, for the kind words. Ananth: ]]

  • Harsh Thakor on April 14, 2012, 15:37 GMT

    Great work,Ananth and I particularly admire the selection of Barrington,Sutcliffe and Hobbs in the RPI who were the ultimate stalwarts in consistency.Everton Weekes too deserves a place with his mammoth consistency.

    However I consider Gavaskar,Border,Dravid,Hutton, and Boycott amongst the greatest batsman in terms of consistency.It is also suprising that Brian Lara has been rated more consistent than Tendulkar.I feel Sachin has been more consistent than any middle-order batsman in the last 6 decades. [[ Pl look at the slice runs shown in an earlier comment and also part of the current Excel file. You will find the reason for this seemingly increased consistency. Ananth: ]] I feel what has to be evaluate is the consistency against the more powerful attacks in the more demanding conditions.This may well promote the likes of Rohan Kanhai,Ian Chappell and Clive Lloyd.It would automatically promote Border and Dravid ,who championed the best bowling in all conditions,consistently.

  • gaurav banodha on April 14, 2012, 15:29 GMT

    It is amazing to see Brian Lara who had a batting style of his own ahead of likes of Sachin, Dravid, Kallis, Ponting etc who are renowned for their great technique in both RPI and consistency(SD). Gary Sobers has rightly said that batting is 90% about hand and eye coordination and 10% about how the bat comes down. [[ I hope you downloaded the most recently uploaded file containg the tables, ordered by innings played. Ananth: ]]

    5

  • Karan Mamgain on April 14, 2012, 13:36 GMT

    Hey Ananth, yet another great piece. I've been keenly following your posts and find them very interesting. Thanks for the hard work you put in them. I have a question slightly unrelated to the this post. I wanted some data from T20I. More specifically, I want to look at over 11-15 analysis of each and every T20I innings because one of the IPL commentators have been repeating, this very period often decides the game, so it would nice to see some statistics about that. Do you know how I can find these data?

    Thanks, again. [[ Karan Thanks. I have responded to you separately. Ananth: ]]

  • gaurav banodha on April 14, 2012, 12:37 GMT

    ha ha ha...you got me bro... embarrassing...i wanted to write 200 but typed 300 instead. but i hope the point i am trying to make is not all that funny. [[ All good-natured fun. So we can laugh. I also thought so, but could not resist. No offence meant. There are 19 batsmen with 200 innings+ careers. If I have to do an analysis, based on the same crtiteria, then it is only a matter of extracting the concerned batsmen from the Excel file. In any case, I will do that and add an additional worksheet to the Excel file. Pl download the same and you can view the same and draw your inference. I do not want to do that since people will pounce on these inferences and attribute motives. Ananth: ]]

  • Azaz Ahmed on April 14, 2012, 11:32 GMT

    hi.....i am really impressed by your hard work on the stats but i belive u have missed younis khan and thilan samaraveera both have above 50 average and have performed consistently... [[ I have not missed out, Azaz. Please download the Excel sheet and you will find both in rows 51 and 83 respectively. I cannot very well show all 162 players in the table in the main article. Ananth: ]]

  • Narinder on April 14, 2012, 10:07 GMT

    [[ Your comment will not be published since you have made an insinuation that I have done this analysis to put one player over the other. It is an insult to the other 160 batsmen who are featured. It shows your complete bias and lack of knowledge of the game. Kindly understand that greatness will be embellished if shared but will lost its value if usurped. No more such comments please. Ananth: ]]

  • vivek on April 14, 2012, 8:10 GMT

    [[ Your comment will not be published since you have made an insinuation that I have done this analysis to put one player over the other. It is an insult to the other 160 batsmen who are featured. It shows your complete bias and lack of knowledge of the game. Kindly understand that greatness will be embellished if shared but will lost its value if usurped. No more such comments please. Ananth: ]]

  • Boll on April 14, 2012, 6:09 GMT

    Interesting to note the relative lack of consistency apparent in the careers of players such as Javed, Ponting, Jayawardene, S.Waugh and Hayden - all of whom have excellent records at home, less good away.

    In contrast, there is Border`s excellent record (ave.57) away from home and v.high consistency. Fleming and Gower, perhaps more remembered for fluent, attractive stroke-making than consistency, also fare v.well in this analysis.

  • gaurav banodha on April 14, 2012, 6:05 GMT

    it is very difficult to imagine likes of ponting, kallis dravid not among the most consistent batsmen in the world. surely consistency over 100-150 innings is different from consistency over 300+ innings. so, to make the comparison fair, either there should be a separate analysis for batsmen with 300+ innings or if all batsmen are to be compared, then we should take into account only first 100 innings and not more. [[ I do not know whether to cry or laugh. There is only one batsman with innings over 300. Tendulkar. So do I do an analysis comnparing Tendulkar with himself. Ananth: ]]

  • Rajesh on April 14, 2012, 5:50 GMT

    Congratulations. As usual excellant piece. Although it is an analysis of "most consistent" batsman, it is the "most consistenly good" batsmen who had or will have greater impact in the team. Can we get an answer to this by combining the Group E to mid 3 groups? There is a chance that I have overlooked. If so, pl. ignore [[ Can I suggest that you download the excellent (if I may say so myself !!!) Excel table I have uploaded and do this yourself. You could also try a few other options. Ananth: ]]

  • Boll on April 14, 2012, 5:24 GMT

    Ah, The Don in Superman undies... damn you Ananth - I just woke up my daughter with a roar of laughter! [[ One day, when she is old enough, explain to her the difference between normal person and Superman. One wears inside and the other, outside: the jocks. But Don was Superman, irrespective of what he did in the dressing room. Ananth: ]]

  • Boll on April 14, 2012, 5:16 GMT

    Bradman`s RPI for scores over 50 is an astonishing 150 (his average for those innings is 185).

    So, he was no better than other greats at getting off to start, but when he did he almost always made it count. He was never dismissed in the 90s, and scored almost as many double centuries (12) as he did scores under 10 (14) - I think Ananth`s analysis gives us an important insight into his unexceptional consistency, but also why he remains so far ahead of any batsman to have played the game. [[ I think this particlular piece of insight and the Tendulkar/Lara figures mentioned by me in another response are the way these aricles should be looked at. Unfortunately readers are looking for statistical shortcomings instead of coming out with and/or understanding such insights. Ananth: ]]

  • Boll on April 14, 2012, 5:03 GMT

    @CricketLover/Ananth.

    I think Ananth makes a very important point re.Bradman`s RPI for centuries/other innings. Sure, when he passed 100 he went on to make score more than anyone else (186 on average) but not significantly more than players such as Lara (173) or Sehwag. When he didn`t, he averaged 31, slightly more than Lara, slightly less than Border. Excellent stats, but not otherworldly. [[ Boll I had the shock of my life when I found that the RpI for below-100 scores for Bradman, Lara and Border were almost identical. I get the feeling that this number will almost be the same for all top batsmen. That means Bradman's other-world performance is solely due to a) the frequency of hundreds and b) the size of hundreds. He was a normal and average batsman in white flannels below 100. Above 100, he wore the jocks outside and became Superman. There lies the material for a near-future article. Since then I have seen that of the top 39 run-makers (Bradman being the last in this list), the RpI below-100 varies between 27.7 (Langer) to 32.0 (Dravid). So that article is on. Ananth: ]] Bradman also `scored` 7 ducks in his 70 completed innings - a significantly higher percentage than modern greats such as Sanga (7 ducks in 171 dismissals) or Dravid (8 ducks in 254) and 7 other times he was dismissed for less than 10. Again, not too bad, but not extraordinary.

    Where Bradman truly separates himself from any other batsman, was in his ability to convert fifties into centuries. I`m fairly sure that only 3 other players (with 10 centuries or more) have a positive centuries to fifties ratio - Walcott(15/14), Hayden (30/29) and Azha (22/21). Bradman`s 29/13 is freakish.

  • Ananth on April 14, 2012, 4:07 GMT

    As I have mentioned in my response to "west indies follower", I have incorporated the Runs scored in each 10-innings slice for each batsmen and incorporated this worksheet in the same Excel table. Readers can download the same afresh. The link is the same, given below. http://dl.dropbox.com/u/39210851/TestConsistency_Batsmen.xls Ananth

  • Victor on April 13, 2012, 22:20 GMT

    Hi Ananth, I would like to say that I love reading your blog articles, they are quite literally an inspiration in numbers. In fact they have spurred me to try and do my own such analysis based on the consistency of teams in global sports! I have one question: Would employing the Co-efficient of Variation (st dev/arithmetic mean) be a worthwhile means to measure consistency across a data sample such as this one? [[ I did the Coefficient of Variation also. Then I found that the mean was a very narrow band and ranged around 5% on either side of 1.00. So it would not have made too much of a difference. Ananth: ]] I've been toying with the idea of using that (along with the standard deviation) in my spreadsheet and considering the analytical depth you achieve with these posts I thought it would be worthwhile asking you. Thanks, Victor

  • Nick on April 13, 2012, 22:06 GMT

    The problem with the analysis i find is that you are measuring a time series without appropriate time series methodology. You are treating each segment as independent samples in a test to try and show a measure of non-independence between subsequent innings. What could perhaps be done is to use rolling segments and to consider the correlation between sequential blocks. [[ I have already talked of rolling segments and the problems in presenting the data. Anyhow that is what Gabriel Rogers did couple of years back and it would be stupid of me to re-do that, not having the credentials as a Statistician nor the statistics-related educational qualifications. Ananth: ]] To eliminate the discrepancies caused by innings wins and not outs simply consider runs per dismissal. I have previously written about the statistical theory in considering this stat, mostly to do with exponential distribution of scores. Also, so far your method makes no allowance for a player being constantly poor and a player being consistently good. [[ Runs per dismissal is the Batting average and I don't want wrong treatment of situations such as 10-4-600-100.00 and 10-0-600-60.00. As far I am concerned both are equal. And if both have the same RpI of 50, they have even performed the same way. So I have used Runs per innings. There is nothing wrong in that. Ananth: ]] Perhaps consider the traditional statistical tests for time series and remove the cyclical effects and take necessary differences to create independently distributed error terms, and then use those to determine the level of consistency. [[ I have only one objective. My article should be understood, appreciated, applaued/criticised by at least 90% of the readers who come in. When I do not do that I deem that article a failure. If I do a complex statistical analysis, and anyhow I cannot really do that, it would be passed over by the majority of the readers. This presentation might lack statistical pedigree and may very well not pass the true-statistical-analysis litmus test but that is not my objective. In the bargain if I fall short, that would not be the first time, nor the last. Ananth: ]]

  • kiran on April 13, 2012, 21:29 GMT

    I might be completely misunderstanding but wouldn't this analysis suffer when calculating the slice performance factor. Is calculating the SD of the SPF, as a measure of consistency, not flawed as it creates a comparison using average as a reference - despite having been previously deemed inaccurate for several reasons? [[ Previously deemed inaccurate: by who. Ananth: ]] For a batsmen with (284)runs in 10 innings vs. Career Expected RPI of 40 (in each of the 10 innings)this would create an output ratio of 71% based on a fundamentally flawed calculation of the expected average - in this way making it just a more elaborate misrepresentation of batsmen performance? There are too many variables at play with comparisons of these all inclusive averages that dont do justice to some players. A disaggregated regional analysis of players using a similiar procedure might present a more accurate portrait? [[ I fon't pretend to understand all what you say. The last sentence makes some sense but not in total. If a batsman scored 284 in a slice of 10 innings against the expected 400, his SPF is 0.71 indicating that, in the selected 10-innings slice, he has been below par, to the extent of 29%. THAT IS ALL. If he scored 500 runs in the next 10-innings slice he has performed at 125%. If he had scored 380 in one slice and 420 in another slice, he has performed almost at his par right throughout. One has to draw cricketing observations from these, not difficult-to-understand statistical inferences. In response to Rajesh, I did a special analysis. I did the career summary of Tendulkar and Lara. It is presented below. Tendulkar 332(0.67), 334(0.67), 419(0.84), 640(1.29), 690(1.39), 496(1.00), 195(0.39), 511(1.03), 727(1.46), 618(1.24), 635(1.28), 419(0.84), 687(1.38), 497(1.00), 669(1.34), 536(1.08), 414(0.83), 446(0.90), 298(0.60), 702(1.41), 170(0.34), 487(0.98), 429(0.86), 470(0.94), 331(0.67), 612(1.23), 577(1.16), 899(1.81), 452(0.91), 356(0.72), 409(0.82), 13(0.67) Lara 573(1.11), 368(0.71), 754(1.46), 523(1.02), 631(1.22), 423(0.82), 479(0.93), 309(0.60), 401(0.78), 385(0.75), 585(1.14), 334(0.65), 340(0.66), 414(0.80), 801(1.55), 479(0.93), 635(1.23), 723(1.40), 684(1.33), 629(1.22), 433(0.84), 409(0.79), 592(1.15), 49(0.67) A perusal of the two sequences will indicate that Tendulkar had two 10-innings slices in which he scored fewer than 200 runs (lt 40%) and once below 300 runs (lt 60%). Compare this with Lara who had NO 10-innings slice of below 300 runs. Maybe that might explain why Lara is seemingly more consistent in this analysis. Why I have brought this point up is to tell this is a cricketing insight and the purpose of this blogspace is to bring out such insights. And I am sure someone would say, Lara scored only 239 runs in the innings group 143-152 and you are wrong, that might be correct but only indicates a deliberate attempt to find fault. Ananth: ]]

  • Rahul Bose on April 13, 2012, 20:55 GMT

    Well great stats analysis, but the answer is quite simple. One D Bradman is the most consistent batsman to ever play the game. He might have deviated more from his mean. But considering his mean was some distance above the rest, its a no contest. [[ No problems at all. The article says nothing to the contrary. It may not be a bad idea, then, to exclude Bradman from all Test batsmen analysis so that true comparisons can be made amongst the rest. Ananth: ]]

  • west indies follower on April 13, 2012, 20:52 GMT

    Hi Ananth, great piece, and a really topical cricketing, if not all sporting analysis. Can we measure a players ability and value to a team through consistency? Is the higher risk, higher reward player (I would put S.Broad in that category pre 2011) more valuable than the dependable, less explosive player? One thing I was wondering was instead of using just career average as the benchmark, could you group certain parts of players careers by their average at the time. For example, Viv Richards in the 1970s averaged 58, and throughout the 1980s 49, so there are two distinct periods of play for him, on which you can measure his consistency. Thus, we can potentially value the impact of certain players on a team, for explosive, through dependable etc. [[ I have already mentioned this a few times. I have uploaded a comprehensive Excel file which can be downloaded by any readers and conclusions drawn. The problem is that the career split will be different for different batsmen and I cannot very well generalize. I will also do one more thing. I have already uploaded an Excel file which contains the slice ratios. I will complete the work and upload a file containing the actual runs scored in the slice. This will facilitate further analysis. Ananth: ]]

  • Sri on April 13, 2012, 19:57 GMT

    Ananth,

    Thanks for the meticulous analysis. I am unable to see the bigger picture of this analyis, please enlighten me. A key parameter that affects consistency is the different conditions (pitch, bowlers, environment ...) that one plays throughout their career in my view. For example, players in the earlier era never had to venture beyond certain countries and those in modern era have traversed all across the world. Wet vs. dry vs. well-rolled pitches. Can you provide your thoughts how can you account for these effects in the models that you have made. [[ I have done multiple analysis where I have taken care of all the variables you have mentioned. Right from the beginning I have mentioned that this is purely to see how batsmen perform in fixed 10-innings slices, irrespective of other conditions. That is the way you should take it. And every analysis cannot contain all relevant factors. Then what would be the difference. When you see Tendulkar's batting average is 55.45, every time do you wonder what were the conditions. That is a single figure across 23 years in every possible conditions. It is a single figure encapsulating the career. Take this analysis like that. Ananth: ]]

  • Zak on April 13, 2012, 18:10 GMT

    Great work!

    What does this tell us? The greatest batsmen are the ones who add the most value and help you win matches. In other words - five consecutive fifties will rarely win you a match, but most likely two big hundreds will.

    In other words, the great batsmen make their good form count by scoring big to counter their low scores when they are in bad form.

    That is why the Pontings and the Dravids are greater than the Tony Greigs.

    Keep up the good work.

  • CricketLover on April 13, 2012, 16:20 GMT

    Ananth,

    I respect your work, but one thing what bothering me is; no matter which way you look;Don B. played less then 50 test and with 29 centuries and some fifties how any analysis can call him less consistence then Boarder or Lara?

    Just doesn't make sense to me, but again not trying to take away credit from you....( and I am sure you not claiming that every time you apply logic for analysis, is the most perfect!) :)

    Cheers [[ Bradman's RpI value is 87.5. So when he scored 334, he had to play almost 3 ducks. Hope you understand what I mean. He had 12 double centuries. So he has to have played many smaller innings to compensate for these runs. His 29 centuries totalled 5394 runs. The RpI value for these 29 innings was 186.0. The other 51 innings grossed 1602 runs leading to a RpI of 31.41. Startling figures indeed. For Lara the figures are 34 centuries at 173.2. 198 other innings at 30.63. For Border, the figures are 34 centuries at 132.6. The other innings are 238 at 31.90. You still maintain that Bradman was very consistent. Do not confuse greatness with consistency. Ananth: ]]

  • Jonas on April 13, 2012, 15:23 GMT

    Given two batsmen of the same average and total number of runs, who would help his team win more games -- the one who is less consistent or the one who is more? [[ Your guess is as good as mine. Let us throw a coin. The whole thing is contextual. There are times when two 50s would win a match while a 0 and 100 might lose if the 0 came at the wrong time. There are times when a 100 and 0 would win if the 100 was a crucial innings. How can there ever be a single answer. For that matter which of the following two innings is more valuable. Agarkar's 109 or Ishant Sharma's 31. Again you can only answer only if you know the context. Ananth: ]]

  • Imhotep on April 13, 2012, 15:11 GMT

    Dear Ananth,

    Another thought provoking analysis. However, I have a little objection to the arbitary 10 innings formula as it might over lap different series and different conditions and like in the case of Lara, one big home /away century would skew the result...

    would it be possible to test consistency by using series runs and averages as a barometer rather than 10 innings...the problem might be with single tests and also in case of SriLanka, a multitude of 2 test series....

    In the above cases, we can see if we can add the single test / 2 test series to the nearest Home / away series depending on where the single test / two test series were played so that we can find out how the batsmen fared series over series.. [[ A few points. The batsmen, bowlers and all-rounders analysis by series have already been done a few months back. If you go to 2011 archives you will see that. In fact taking 10 innings is the perfect way in which ALL the batsmen are treated similarly. Previous eras, these 10 innings might be covered in a single Test series. Nowadays these would be split over even three series. No point in bringing series into this. Then the batsmen will have different yardsticks and the whole thing will go haywire. The 10 innings has been fixed for all 2700 players. If there are minor problems with the number, they are applicable across 137 years. Ananth: ]]

  • Mustafa on April 13, 2012, 13:54 GMT

    Long time reader, don't really comment a lot. The numbers are perfect, they make sense. Border seem to be the perfect example, he was a fighter, not as gifted as many, but always did his job in some shape or form, his numbers prove his contribution to the building of the Australian base that lasted for nearly 15 years. [[ Border has been more consistent. But others have contributed in equal measure. My feeling is taht there are 80 batsmen who average above 45.0. These have all contributed significantly to the team cause. And some of the others, too. Ananth: ]] Lara was Lara, lets leave it at that.

  • dodgeywinger on April 13, 2012, 13:47 GMT

    The number 10, why not look at all the inings a player has played? Is it not intutive, the longer a player plays, the higher the S.D? [[ What do you mean. Of course all the innings the batsman has played have been considered, in 10-innings slices. Ananth: ]] It does not really make sense to compare the S.D of Lara to that of Tendulkar, as Lara stopped playing 6 years (I am guessing here) ago. Your graph says this empahtically, the three best, Tendulkar, Kallis, Dravid, Ponting have their S.D so close together, but higher than the rest, cause they have had some of the longest careers. If Kevin Pie can keep his SD to .21 after another 140 ininnings,I will consider him in the same class as the above 4. [[ I fail to understand why the fact taht Lara retired 5 years back should preclude a comparison with Tendulkar. And Lara has a SD of 0.78 after 232 innings and Border has a SD of 0.241 after 265 innings. These are not short careers. Ananth: ]]

  • Tom on April 13, 2012, 12:59 GMT

    Great stuff as ever Ananth. I suspect that Pietersen, like Gower, goes through a number of innings making middling or low scores and then gets a big one, creating the impression of inconsistency without ever having a prolonged period out of form. It's also interesting to see that Strauss's figures are impressively consistent across his career despite frequent mutterings that he's out of form.

    Also very interesting to see Tony Greig's high level of consistency with the bat considering that in the bowling analysis he was all over the place. [[ Possibly Lara and Pietersen do that. This means that when they crossed 100 they were most likely to play big innings. Lara's average hundred value is 173.2. Only Bradman (186.0) and Sehwag (179.8) are ahead of him., amongst those with 20 hundreds. Pietersen is tied with Tendulkar at 145.8 Ananth: ]]

  • Gerry_the_Merry on April 13, 2012, 12:19 GMT

    Ananth, I believe asolitarywave is also talking about non-overlapping only. just shifting the starting point by 5 innings. [[ I think, Gerry, that was a only the basis for the argument that if we had a different starting point, the results would be different. He is not suggestimng 5. Then why not 11, ignoring the first 10. Okay Azhar, Walters and Rowe would complain ??? Anyhow I have to analyze the complete players' career. Ananth: ]]

  • Ananth on April 13, 2012, 12:18 GMT

    Mr.Sujoy/Champu, I do not know how much time you took to type your comment. But at the end of the 10th second of reading, I came to the personal attack on me and instantly trashed it, sending it to the deserved place. It is amazing that you guys still do not know the rules of this blogspace, quite different to the others. Some of you make good points, maybe negative, no problems, but stop there. Your comment will be published. The minute (no, make it the second) you make a personal attack on me, another reader or a player, your comment will be trashed and would only have a 10-second viewing life by me. I have taken the trouble of junking your comment but responding since I think the message has to be passed and understood. The requirements are inviolable, at least as far as this blogspace is concerned. Ananth

  • asolitarywave on April 13, 2012, 11:36 GMT

    Ananth, I'm not sure I am reading too much into one example. you've gone for the first ten innings, then the nest ten and so on. Fine, makes sense. But if you you ignored the first five, say, and then repeated the analysis for innings 5-15, 16-25 etc I think you would get a different result (simply because for most batsmen, they score a large number of below average scores with a relatively small number of big scores - and which cohort these hundreds fall into is the significant factor) [[ Richard, I get what you mean. In fact that was what Gabriel had done earlier. However then we will have 301 values for Tendulkar, 276 for Dravid, 268 for Ponting and so on. There is no way these numbers can be handled other than by pure statistical methodology and graphs. It will go over 90% of the readers then. By having non-overlapping discreet 10-innings slices I might not have satisfied the purists but have been able to limit the maximum number of slices to 32 and made the results meaningful and visible, without resorting to p-factors and the like, to almost all the readers. Anyhow thanks for the valuable insights. Ananth: ]]

  • gaurav banodha on April 13, 2012, 11:35 GMT

    perhaps the reason sachin, ponting & dravid are not there is the large no. of matches they have played. when you play so many matches over such a long period of time, ups and downs do happen.

  • Yash Rungta on April 13, 2012, 11:31 GMT

    @Ananth(regarding my earlier comment): So if there is a player like Ponting who's career can be divided in 3 stages. 1) Youth but inexperienced 2) Balance of youth and experience which is generally the peak period for any player 3) Experienced but old

    Now the case with Ponting I think was that he was consistent in those periods. Like he probably averaged around 40 in the first period, about 70-75 in the 2nd period and about 35 in the last period. So if you calculate deviation from his career average of 53, his deviation would be a lot. But if you divide it into 3 stages, actually there isn't much deviation. Obviously I know you're taking into account runs(and not average) but they go hand-in-hand.

    I'm not expecting you to divide his(or anybody's career) in 3 stages but what I'm telling you is that players like Ponting were consistent within those 3 periods but taking his career as a whole, he was inconsistent. Similar issue with Sangakkara(and to a lesser extent even Sachin).. [[ Yash, as I have already replied to Nitin, you can download the Excel file and analyze to your heart's content. Ananth: ]]

  • Nitin Gautam on April 13, 2012, 10:41 GMT

    Anantha

    very comprehensive & detailed analysis. for sure its quiet surprising to see KP's credentials as a consistent batsman owing to lots of articles written about his fall from the stupendous heights he reached early in his career & so called inconsistency. Regarding Lara, I am sure he would not have failed in any stretch of 10 innings in his entire career & wherevere i failed for, say, consecutive 5-6 innings he compensated with a mammoth score n that is why it was not a surprise. Ponting & Dravid's relative inconsistency is not surprising since their phenomenal record is nore due to their inhuman effors between 02-06. SRT actually surprised me. Thought of him as the most consistent player but i guess post 2002, he started failing more often n periods like going from 100 no. 29 to 30, no.34-35 & finally 99-100 took its heavy toll. Not to start any debate at all, but in modern batsman till 2002, I am sure SRT would be the most consistent player. [[ Nitin, download the massive Excel file, which comtains the complete SPF values for all 162 batsmen and you can prove your conslusion yourself. You have to relate the slice number to the year. Impossible for me to show that also. Yes, Lara's penchant for big scores have certainly kept him in good shape in this analysis. Ananth: ]]

  • Milpand on April 13, 2012, 10:25 GMT

    Thanks for the recent posts. In order to respect your article with a proper comment, it is important to go through the tables and reread the article a couple of times. While reading on the move is easy understanding tables require a proper sit down, typically after tucking only son to bed. Last few weeks, after reading the Amar Chitra Katha version of Shakuntala to him, I too have fulfilled my desire to read the original. I decided to get back to you in a rush to let you know that your work is highly appreciated even if it is sometimes not feasible to write back in time. [[ The sentiment is appreciated. Amar Chitra Katha came out way after my childhood days. But our son used to read and all of us also. Now the same bound books have been passed on to our grandson. These books cost Rs.5 each (now Rs.50). Anant Pai was a wonderful old man. Probably has a special place amongst the gods and Goddesses he wrote so beautifully about. I am sure on your part you must have read Kalidasa's original Sanskrit version. Ananth: ]]

  • Charles on April 13, 2012, 10:20 GMT

    Excellent article! I firmly believe that consistency does not necessarily make batsmen great. Think of what the experts always say: You can forgive a good batsman for getting out for less than 10, but once he gets past 20 or 30, he should go on and get a big score. This will obviously lead to larger standard deviations from good batsmen. Your analysis seems to support this view. What is the correlation between your measures of consistency and measures of the "greatness" of a batsman? A quick caluclation from the limited figures here seems to show that that there is generally a negative correlation between "greatness" and consistency. Well done once again! [[ Charles, "greatness" is a very difficult to define across career. So you have to draw your own conclusions. However see Lara's figures. For all the so called perceptions of inconsistency and "flawed genius" type of putting downs, he and Border are the most consistent amongst the top-20 run scoring batsmen. So your conclusion is partly true. The bowler analysis, if you had followed the same, supported your axiom in a clearer manner. Bob Willis ??? Ananth: ]]

  • asolitarywave on April 13, 2012, 10:17 GMT

    Maybe I have misunderstood the analysis, but surely the blocks you take have a significant effect on the consistency measured here? To take your example of the english winter / summer - this appears as a an inconsistency if you choose all the summer test as one block, and the winter tests as a different block. If however you block divisions end up using the last two tests of the summer and the first three of the winter, then the inconsistency will be much more muted. This would be an argument for using a rolling analysis, for all that freak scores would then be counted ten times. [[ You are reading too much into the example. That was given only to drive home the point that a great series followed by a terrible series do not make two good series. It was also relevant to use that example for 2011. Replace the words "English summer" with slice no 12 consisting of innings 111-120 and "English winter" with slice no 13 consisting of innings 121-130, what I say will make sense. It does not matter whether there are overlaps between home/away Tests, strong/weak teams et al. I do not really like rolling periods. I know that people might say the 10-innings slpit is arbitrary. It is no more arbitrary than all analysis which split a player's career into years. MCG in 2011and SCG in 2012 ??? The point is that 10 innings, across all varying conditions, gives a batsman enough opportunities to reach his career RpI x 10. Ananth: ]]

  • Yash Rungta on April 13, 2012, 9:02 GMT

    Hi,

    I appreciate the statistical analysis gone into this piece! :) But if one batsman scores a fifty in 2 innings and the other scores a duck and a century in 2 innings, to me both as equally as valuable. Being inconsistent means that a batsman has had very bad patches but also equally very good patches which evens it out! If 2 batsmen average the same, they're equally as valuable. If he disappointed in some series, he must have single handedly won some too..

    And in terms of players like Ponting, the analysis above shows he is a highly inconsistent player which wasn't the case if you ask me. He started off his career as a decent batsman then had a very good patch that ran about 6-7 years and is now back to being a decent batsman. This is because he didn't have a lot experience prior to 2000 and after 2008 he starting ageing. In between he had a good balance between the two elements. Something similar to Sangakkara. Its a way age(youth and experience) has shown off in their careers!! [[ Yash, you have used imprecise and subjective words such as "decent", good patch", "decent again", "ageing" to define Ponting's career and concluded that he is a consistent batsman. I have used precise 10-innings slices, determined the runs scored, derived a ratio between these and a clearly defined career measure, grouped these, got the SDs and concluded that amongst the top-10 run-scoring batsmen, Ponting is probably the least consistent. This, without one negative comment oh his undisputed greatness. You have to decide which method is acceptable. If you say your method, I would not argue at all. Ananth: ]]

  • Sunil Mamgain on April 13, 2012, 8:47 GMT

    Great work of statistics buddy, you might have spent lot of time doing it.

  • No featured comments at the moment.

  • Sunil Mamgain on April 13, 2012, 8:47 GMT

    Great work of statistics buddy, you might have spent lot of time doing it.

  • Yash Rungta on April 13, 2012, 9:02 GMT

    Hi,

    I appreciate the statistical analysis gone into this piece! :) But if one batsman scores a fifty in 2 innings and the other scores a duck and a century in 2 innings, to me both as equally as valuable. Being inconsistent means that a batsman has had very bad patches but also equally very good patches which evens it out! If 2 batsmen average the same, they're equally as valuable. If he disappointed in some series, he must have single handedly won some too..

    And in terms of players like Ponting, the analysis above shows he is a highly inconsistent player which wasn't the case if you ask me. He started off his career as a decent batsman then had a very good patch that ran about 6-7 years and is now back to being a decent batsman. This is because he didn't have a lot experience prior to 2000 and after 2008 he starting ageing. In between he had a good balance between the two elements. Something similar to Sangakkara. Its a way age(youth and experience) has shown off in their careers!! [[ Yash, you have used imprecise and subjective words such as "decent", good patch", "decent again", "ageing" to define Ponting's career and concluded that he is a consistent batsman. I have used precise 10-innings slices, determined the runs scored, derived a ratio between these and a clearly defined career measure, grouped these, got the SDs and concluded that amongst the top-10 run-scoring batsmen, Ponting is probably the least consistent. This, without one negative comment oh his undisputed greatness. You have to decide which method is acceptable. If you say your method, I would not argue at all. Ananth: ]]

  • asolitarywave on April 13, 2012, 10:17 GMT

    Maybe I have misunderstood the analysis, but surely the blocks you take have a significant effect on the consistency measured here? To take your example of the english winter / summer - this appears as a an inconsistency if you choose all the summer test as one block, and the winter tests as a different block. If however you block divisions end up using the last two tests of the summer and the first three of the winter, then the inconsistency will be much more muted. This would be an argument for using a rolling analysis, for all that freak scores would then be counted ten times. [[ You are reading too much into the example. That was given only to drive home the point that a great series followed by a terrible series do not make two good series. It was also relevant to use that example for 2011. Replace the words "English summer" with slice no 12 consisting of innings 111-120 and "English winter" with slice no 13 consisting of innings 121-130, what I say will make sense. It does not matter whether there are overlaps between home/away Tests, strong/weak teams et al. I do not really like rolling periods. I know that people might say the 10-innings slpit is arbitrary. It is no more arbitrary than all analysis which split a player's career into years. MCG in 2011and SCG in 2012 ??? The point is that 10 innings, across all varying conditions, gives a batsman enough opportunities to reach his career RpI x 10. Ananth: ]]

  • Charles on April 13, 2012, 10:20 GMT

    Excellent article! I firmly believe that consistency does not necessarily make batsmen great. Think of what the experts always say: You can forgive a good batsman for getting out for less than 10, but once he gets past 20 or 30, he should go on and get a big score. This will obviously lead to larger standard deviations from good batsmen. Your analysis seems to support this view. What is the correlation between your measures of consistency and measures of the "greatness" of a batsman? A quick caluclation from the limited figures here seems to show that that there is generally a negative correlation between "greatness" and consistency. Well done once again! [[ Charles, "greatness" is a very difficult to define across career. So you have to draw your own conclusions. However see Lara's figures. For all the so called perceptions of inconsistency and "flawed genius" type of putting downs, he and Border are the most consistent amongst the top-20 run scoring batsmen. So your conclusion is partly true. The bowler analysis, if you had followed the same, supported your axiom in a clearer manner. Bob Willis ??? Ananth: ]]

  • Milpand on April 13, 2012, 10:25 GMT

    Thanks for the recent posts. In order to respect your article with a proper comment, it is important to go through the tables and reread the article a couple of times. While reading on the move is easy understanding tables require a proper sit down, typically after tucking only son to bed. Last few weeks, after reading the Amar Chitra Katha version of Shakuntala to him, I too have fulfilled my desire to read the original. I decided to get back to you in a rush to let you know that your work is highly appreciated even if it is sometimes not feasible to write back in time. [[ The sentiment is appreciated. Amar Chitra Katha came out way after my childhood days. But our son used to read and all of us also. Now the same bound books have been passed on to our grandson. These books cost Rs.5 each (now Rs.50). Anant Pai was a wonderful old man. Probably has a special place amongst the gods and Goddesses he wrote so beautifully about. I am sure on your part you must have read Kalidasa's original Sanskrit version. Ananth: ]]

  • Nitin Gautam on April 13, 2012, 10:41 GMT

    Anantha

    very comprehensive & detailed analysis. for sure its quiet surprising to see KP's credentials as a consistent batsman owing to lots of articles written about his fall from the stupendous heights he reached early in his career & so called inconsistency. Regarding Lara, I am sure he would not have failed in any stretch of 10 innings in his entire career & wherevere i failed for, say, consecutive 5-6 innings he compensated with a mammoth score n that is why it was not a surprise. Ponting & Dravid's relative inconsistency is not surprising since their phenomenal record is nore due to their inhuman effors between 02-06. SRT actually surprised me. Thought of him as the most consistent player but i guess post 2002, he started failing more often n periods like going from 100 no. 29 to 30, no.34-35 & finally 99-100 took its heavy toll. Not to start any debate at all, but in modern batsman till 2002, I am sure SRT would be the most consistent player. [[ Nitin, download the massive Excel file, which comtains the complete SPF values for all 162 batsmen and you can prove your conslusion yourself. You have to relate the slice number to the year. Impossible for me to show that also. Yes, Lara's penchant for big scores have certainly kept him in good shape in this analysis. Ananth: ]]

  • Yash Rungta on April 13, 2012, 11:31 GMT

    @Ananth(regarding my earlier comment): So if there is a player like Ponting who's career can be divided in 3 stages. 1) Youth but inexperienced 2) Balance of youth and experience which is generally the peak period for any player 3) Experienced but old

    Now the case with Ponting I think was that he was consistent in those periods. Like he probably averaged around 40 in the first period, about 70-75 in the 2nd period and about 35 in the last period. So if you calculate deviation from his career average of 53, his deviation would be a lot. But if you divide it into 3 stages, actually there isn't much deviation. Obviously I know you're taking into account runs(and not average) but they go hand-in-hand.

    I'm not expecting you to divide his(or anybody's career) in 3 stages but what I'm telling you is that players like Ponting were consistent within those 3 periods but taking his career as a whole, he was inconsistent. Similar issue with Sangakkara(and to a lesser extent even Sachin).. [[ Yash, as I have already replied to Nitin, you can download the Excel file and analyze to your heart's content. Ananth: ]]

  • gaurav banodha on April 13, 2012, 11:35 GMT

    perhaps the reason sachin, ponting & dravid are not there is the large no. of matches they have played. when you play so many matches over such a long period of time, ups and downs do happen.

  • asolitarywave on April 13, 2012, 11:36 GMT

    Ananth, I'm not sure I am reading too much into one example. you've gone for the first ten innings, then the nest ten and so on. Fine, makes sense. But if you you ignored the first five, say, and then repeated the analysis for innings 5-15, 16-25 etc I think you would get a different result (simply because for most batsmen, they score a large number of below average scores with a relatively small number of big scores - and which cohort these hundreds fall into is the significant factor) [[ Richard, I get what you mean. In fact that was what Gabriel had done earlier. However then we will have 301 values for Tendulkar, 276 for Dravid, 268 for Ponting and so on. There is no way these numbers can be handled other than by pure statistical methodology and graphs. It will go over 90% of the readers then. By having non-overlapping discreet 10-innings slices I might not have satisfied the purists but have been able to limit the maximum number of slices to 32 and made the results meaningful and visible, without resorting to p-factors and the like, to almost all the readers. Anyhow thanks for the valuable insights. Ananth: ]]

  • Ananth on April 13, 2012, 12:18 GMT

    Mr.Sujoy/Champu, I do not know how much time you took to type your comment. But at the end of the 10th second of reading, I came to the personal attack on me and instantly trashed it, sending it to the deserved place. It is amazing that you guys still do not know the rules of this blogspace, quite different to the others. Some of you make good points, maybe negative, no problems, but stop there. Your comment will be published. The minute (no, make it the second) you make a personal attack on me, another reader or a player, your comment will be trashed and would only have a 10-second viewing life by me. I have taken the trouble of junking your comment but responding since I think the message has to be passed and understood. The requirements are inviolable, at least as far as this blogspace is concerned. Ananth