THE CORDON HOME

BLOGS ARCHIVES
SELECT BLOG
March 17, 2013

Statistics

The ODI batting giants: part 1

Anantha Narayanan
MS Dhoni has developed into one of the finest finishers in one-day internationals, and has revelled in the difficult task of batting in the lower middle order  © BCCI
Enlarge

At the outset let me apologise to all the readers for my inability to read and respond to the comments sent for my previous article. Even though the reasons were beyond my control, it is my corner of "the Cordon" and it was my responsibility to keep this small area working, accessible and free of trash. My thanks to Cricinfo technical team for addressing this sticky problem and I will have access to the comments for my articles, which will henceforth be published, interjected with my responses.

The year 2012 witnessed retirement of the two giants of ODI format. Two pillars of the game, neither of whom could ever be denied a place among the top four of the ODI game. They scored tons of runs and scored these in a magnificent manner, scored these when their team needed them and pulverised top-quality attacks. Sachin Tendulkar and Ricky Ponting played like kings and retired in style. Towards the end of the year another player retired from Test cricket, in the opinion of many, when he was still at the top. Unfortunately this also meant an exclusion from a farewell ODI series. However, this cannot deny the stellar contributions Michael Hussey made towards his team's successes. The retirement of these three giants has prompted me to do a very exhaustive analysis of the ODI batting giants. This is a two-part article due to the number of areas covered. At the end of this article I have outlined the second article so that reader requests can be incorporated wherever possible.

How many batsmen do I include in this analysis and how do I decide on the specific batsmen? A very difficult task indeed. I have used a combination of numbers, the team achievements, their contributions to the ODI format in general and their team in particular. I have tried to ensure as wide a representation across countries and as broad-based representation across the years as possible. Over the past 40 odd years, 10 World Cups and 6 Champions Trophy tournaments have been held. Australia has won 6, West Indies 3, India 2.5, Sri Lanka 1.5 and Pakistan, South Africa and New Zealand, one each. The selection of players reflects, to some extent, this level of success of their teams.

Taking all above into consideration, I have selected following 15 batsmen, listed in chronological order. There may be minor disagreements among readers but there is no denying that these 15 represent the very cream of ODI batting. They have contributed over 130,000 runs, just short of 10% of the total runs scored in ODI matches. Most of these batsmen select themselves.

Richards      1975 1991
Miandad       1975 1996
M Crowe       1982 1995
M Waugh       1988 2002
Tendulkar     1989 2012
Jayasuriya    1989 2011
Lara          1990 2007
Inzamam       1991 2007
A Flower      1992 2003
Bevan         1994 2004
Ponting       1995 2012
Kallis        1996 2012 Active
Gilchrist     1996 2008
Pietersen     2004 2013 Active
Dhoni         2004 2013 Active

I included Hussey in the beginning but towards the end felt that Mark Waugh could not be ignored. Since both are among my favourite batsmen, it was a wrench to exclude either of them. Martin Crowe played in tough times and just about edged out Fleming. Jayasuriya transformed the game itself. Miandad and Inzamam selected themselves. So did Richards and Lara. The Tendulkar lily need not be gilded. Dhoni's finishing exploits are legendary giving him the edge ahead of Ganguly. Pietersen and Andy Flower are, by far, their respective country's best batsmen.

My sincere apologies to Lloyd, Gayle, Haynes, Dean Jones, Hussey, Saeed Anwar, Mohammad Yousuf, de Silva, Sangakkara, Jayawardene, Ganguly, Azharuddin, Sehwag, Graeme Smith, Gibbs, Fleming et al, and their supporters. Any of them would have graced this list with distinction. Shakib Al Hasan will certainly qualify into an allrounder's list but has not done enough in batting.

In view of the number of tables and the huge amount of information, I have decided to do this mammoth analysis in two parts. The first part is based on available data and I would venture to say that most of this data could be extracted using statsguru. However, here it is available in a single place in easily understandable tables for the crème de la crème of ODI batsmen. The second part will be slightly different where I have done a lot of extraction and grouping for those tables. Many of the ideas in the second part are unique and will not be available anywhere. Readers would also have the opportunity to suggest any new tables which could be developed. I would be glad to create these if I can.

Here is one important factor about the presentation of these tables. I have decided not to order the tables on any data field. It will invite unnecessary discussions. These are all great players and this exercise is not to determine who the best is. Hence all tables will be presented in strictly chronological order. The highlighting would be done in the commentary after the tables.

1. Career Summary
BatsmanInngsNOsRunsBallsAvgeS/RRpIIndex
Richards 16724 6721 758147.00 88.740.235.7
Miandad 21841 73811097941.70 67.233.922.8
M Crowe 14119 4704 646438.56 72.833.424.3
M Waugh 23620 85001106339.35 76.836.027.7
Tendulkar 45241184262137144.83 86.240.835.1
Jayasuriya 43318134301473632.36 91.131.028.3
Lara 28932104061305640.49 79.736.028.7
Inzamam 35053117391582739.53 74.233.524.9
Flower 20816 6785 914435.34 74.232.624.2
Bevan 19667 6914 929953.60 74.435.326.2
Ponting 36539137031706742.03 80.337.530.1
Kallis 30753114991575645.27 73.037.527.3
Gilchrist 27911 9619 992335.89 96.934.533.4
Pietersen 12116 4369 503541.61 86.836.131.3
Dhoni 19656 7259 822851.85 88.237.032.7

These are numbers which any cricket aficionado would reel off if woken up at 3am. The key columns here are RpI and the Index. The Batting average in ODIs is far more skewed than in Test matches because of the limited number of overs, non-completion of innings and high percentage of not-outs. The number of not-outs varies between 67 for the middle-order stalwart like Bevan to 18 for an attacking opener like Jayasuriya. Hence RpI (Runs per Innings) is a very important measure and reflects the relative positioning of batsmen far more accurately. Hence almost all these tables will have both Average and RpI.

Now for the Index. This has been a favourite combination measure of mine over the past 15 years. Realizing the importance of Average/RpI and the Strike Rate, I had multiplied the two measures, thus giving them equal importance. It allows players to compensate shortcomings in one measure with higher level performances in the other. Earlier I used this Index based on Batting average but now I always use RpI in all Index calculations.

The batting averages which stand out are Richards' 47.00, Bevan's 53.60 and Dhoni's 51.85. Jayasuriya's 90+ strike rate and the near-90 strike rate of Richards at 88.7 and Dhoni at 88.2 stand out. The 35+ Index values of Richards and Tendulkar sets them apart. Readers can note how Jayasuriya's low RpI value is partly compensated by the high strike rate. Similarly Kallis' average strike rate is offset by a good RpI value. The Index represents a clear value to the team: non-contextual of course.

2. Opening runs
BatsmanInnsNOsRunsBallsAvgeS/RRpIIndexOPP RunsOPP Avge
Richards 0 0 0 0 0.00 0.0 0.0 0.0 0 0.0
J Miandad 4 0 85 8621.25 98.821.221.0 10225.5
Crowe 22 1 814 117338.76 69.437.025.7 100845.8
M Waugh 14111 5729 747344.07 76.740.631.1 595942.3
Tendulkar 34023153101739748.30 88.045.039.61415441.6
Jayasuriya 38315127401378834.62 92.433.330.71287133.6
Lara 52 5 2166 287146.09 75.441.731.4 159730.7
Inzamam 12 0 516 72243.00 71.543.030.7 49341.1
Flower 44 2 1352 192832.19 70.130.721.5 167638.1
Bevan 1 1 40 62 0.00 64.540.025.8 1717.0
Ponting 6 1 272 35054.40 77.745.335.2 29549.2
Kallis 7 0 135 20619.29 65.519.312.6 24334.7
Gilchrist 259 7 9200 938636.51 98.035.534.81109242.8
Pietersen 8 1 412 46958.86 87.851.545.2 56370.4
Dhoni 2 0 98 11349.00 86.749.042.5 2110.5

The opening position in ODIs has undergone sea-changes across the years. Steady opening partnerships with both batsmen striking at around 60, with the objective of scoring a match-winning 250, were the order of the day around the 1980s. Mark Greatbatch was the first of the attacking openers who set the 1992 World Cup alight. This was followed by the Sri Lankan blitz with both openers, Jayasuriya and Kaluwitharana, blazing away, in the next edition. They won a World Cup with these attacking methods although the final two matches were won by conventional batsmen. Over the next 15 years the opening batting has changed in accordance with rule changes and power play implementations. Today it would be difficult to find a Greenidge-Haynes combination opening, both striking at below 65. This is a specialist position and various top batsmen have occupied it, with telling effect. Hence a special view of this anchor position is necessary.

The table contents are standard for most of these analyses. Richards never opened, Miandad once in three years and Martin Crowe, very rarely. Many of the others like Mark Waugh, Tendulkar, Jayasuriya and Gilchrist, spent most of their careers opening the batting. Mark Waugh, less than the other three. Gilchrist was the true opening-zone batsman, striking at 98. Jayasuriya was nearly as good, at 92 and Tendulkar did his scoring at 88. Mark Waugh was more sedate, probably content with watching Gilchrist's striking at the other end. Still a respectable 77. Tendulkar's RpI of 45 is mind-blowing. This also means that he has the highest Index value of nearly 40. It can be seen that all the four had Index values exceeding 30. Lara's opening stints were more effective than the rest of his batting efforts. Andy Flower, on the other hand, was less successful opening the batting.

I have also determined the average opening partnerships when the concerned batsman opened: mostly with some other batsman, other than Mark Waugh and Gilchrist who opened the innings regularly together. Gilchrist leads this quartet for an average opening partnership of 45.8 and is closely followed by Mark Waugh, with 45.3: a testament to the fact that they opened together often. Tendulkar, often with Ganguly, and the rest of the time with different batsmen, averages 41.6. Jayasuriya's average is only 33.6, a reflection of the fact that he and his partner were in attacking mood almost throughout their partnerships.

3. Late order runs (6 onwards)
BatsmanInnsNOsRunsBallsAvgeS/RRpIIndexTeamRunsBatRuns %
Richards 6 2 80 10020.00 80.013.310.7 69111.6
Miandad 9 1 175 26721.88 65.519.412.7 40043.8
M Crowe 3 1 138 17669.00 78.446.036.1 50727.2
M Waugh 4 2 51 6025.50 85.012.810.8 15333.3
Tendulkar 5 1 168 13542.00124.433.641.8 20781.2
Jayasuriya 31 2 405 55013.97 73.613.1 9.6 172023.5
Lara 8 3 263 27452.60 96.032.931.6 51551.1
Inzamam 25 3 542 75424.64 71.921.715.6 176230.8
Flower 8 3 197 29739.40 66.324.616.3 68928.6
Bevan 10645 3350 428654.92 78.231.624.7 824340.6
Ponting 6 3 75 9025.00 83.312.510.4 42417.7
Kallis 7 0 120 22017.14 54.517.1 9.4 49424.3
Gilchrist 19 4 390 49326.00 79.120.516.2 121432.1
Pietersen 5 1 128 12532.00102.425.626.2 49825.7
Dhoni 11334 3375 403242.72 83.729.925.0 947935.6

Just as opening the batting is important, finishing an innings, whether setting a target or chasing is important. Unfortunately in this analysis the required data, which is the information on when a batsman was dismissed, is available only over the past 20 years, for two-thirds of the matches. So I have to take a view based on when the batsman entered to start his innings. So this may not be a complete analysis. But we can draw a few insights. For this analysis I have compiled the runs scored where a batsman came in at no.5/6 or afterwards. At no.6, only two of these batsmen, viz., Bevan and Dhoni have played enough innings and that is the reason why I have considered that the role of a finisher can be at No.5 or No.6. Those at 1 and 2 are openers, 3/4/5 are consolidators and 5/6 onwards are finishers. I know this is not perfect but it cannot be helped.

Examining the table reveals that only two batsmen have played enough innings at No.6 onwards. Bevan has played 106 very effective innings at an RpI of 31.6 and Index of 24.7. Dhoni has also excelled in this position with 113 innings at 29.9 and Index of 25.0. It should be noted that the RpI will necessarily be lower since these two batsmen have remained not out 45 and 34 times respectively. These values are very important for the team since these are scored at the end of the innings in crunch situations, whether the team was batting first or second.

Here I have looked for a different type of additional analysis. I compiled the sum of the runs added by the team after the entry of the concerned batsman and looked at what % was scored by the concerned batsman. I think I struck pay dirt since Bevan's 40.6% of all the runs scored by the team after his entry is a testament to the great impact on the Australian batting. Inclusive of Bevan's own failures this figure of 40+% is the essence of finishing a match. Dhoni is slightly below, at 35.6% possibly because he often had top-order batsmen batting with him.

4. Late-order runs (5 onwards)
BatsmanInnsNOsRunsBallsAvgeS/RRpIIndexTeamRunsBatRuns %
Richards 35 3 930 109729.06 84.826.622.5 282832.9
Miandad 36 6 887 132129.57 67.124.616.5 306229.0
M Crowe 8 3 219 29643.80 74.027.420.3 125717.4
M Waugh 41 5 985 119027.36 82.824.019.9 281335.0
Tendulkar 41 9 965 111830.16 86.323.520.3 309731.2
Jayasuriya 38 2 438 62212.17 70.411.5 8.1 218420.1
Lara 47 9 1277 166533.61 76.727.220.8 407431.3
Inzamam 13025 4015 531338.24 75.630.923.31282031.3
Flower 56 7 1658 238433.84 69.529.620.6 478734.6
Bevan 13950 4515 584850.73 77.232.525.11111740.6
Ponting 12 4 124 16415.50 75.610.3 7.8 113510.9
Kallis 30 6 950 137439.58 69.131.721.9 356326.7
Gilchrist 19 4 390 49326.00 79.120.516.2 121432.1
Pietersen 17 7 826 83082.60 99.548.648.4 181645.5
Dhoni 16047 5258 623946.53 84.332.927.71383138.0

This is a slight variation to the previous table. I have looked at batting efforts from No.5 onwards. A lot more batting efforts now find a place. Bevan improves slightly while Dhoni has a significant move upwards. Inzamam is the other batsman who has played many important innings at No.5 onwards. His Index value is a reasonable 23.3. Note how badly Jayasuriya and Ponting performed on the few occasions when they were moved out of their comfort zones of opening and No.3 respectively.

Bevan remains at 40.6% in the % team runs measure. Dhoni's figure improves to 38.0%. The new serious entrant to this table, Inzamam scored 34.6% of his team runs.

5. Against Best Team (based on runs scored)
BatsmanTeamInnsNOsRunsBallsAvgeS/RRpIIndex
Richards Aus 50 7 2187 258250.86 84.743.737.0
Miandad Win 64 7 1930 300533.86 64.230.219.4
M Crowe Aus 34 3 1096 158135.35 69.332.222.3
M Waugh Win 45 2 1708 222739.72 76.738.029.1
Tendulkar Slk 80 9 3113 355843.85 87.538.934.0
Jayasuriya Ind 85 5 2899 298936.24 97.034.133.1
Lara Aus 50 3 1858 242939.53 76.537.228.4
Inzamam Ind 64 9 2403 306043.69 78.537.529.5
Flower Ind 35 3 1297 172440.53 75.237.127.9
Bevan Saf 31 5 1163 149044.73 78.137.529.3
Ponting Ind 59 5 2164 265840.07 81.436.729.9
Kallis Win 40 7 1666 214450.48 77.741.632.4
Gilchrist Ind 45 1 1622 163436.86 99.336.035.8
Pietersen Ind 28 3 1138 132345.52 86.040.635.0
Dhoni Slk 4511 2041 225260.03 90.645.441.1

Who were the favourite opponents of these wonderful batsmen? I have gone on runs scored rather than an average since scoring 2000 at 40 is more significant than scoring 500 at 50. The table is self-explanatory. The numbers which stand out are Dhoni's outstanding Index value of 41 against Sri Lanka, Richards' 37 against Australia and Gilchrist's 35+ against India. Tendulkar's 3000+ runs against Sri Lanka is the only instance of a batsman scoring over 3000 runs against a single country. I must confess that this is an educated guess and I will be glad to be disproved.

6. Best Year (based on runs scored)
BatsmanYearInnsNOsRunsBallsAvgeS/RRpIIndex
Richards 1985 25 5 1231 133261.55 92.449.245.5
J Miandad 1987 22 6 1084 154267.75 70.349.334.6
Crowe 1990 20 1 810 117742.63 68.840.527.9
M Waugh 1999 36 3 1468 194244.48 75.640.830.8
Tendulkar 1998 33 4 1894 185465.31102.257.458.6
Jayasuriya 2001 33 1 1202 144437.56 83.236.430.3
Lara 1993 30 3 1349 185749.96 72.645.032.7
Inzamam 1999 28 4 1106 157146.08 70.439.527.8
Flower 2001 33 3 1060 130135.33 81.532.126.2
Bevan 1998 22 8 959 117468.50 81.743.635.6
Ponting 2007 24 6 1424 155379.11 91.759.354.4
Kallis 2000 38 9 1300 195244.83 66.634.222.8
Gilchrist 1999 37 0 1241 139333.54 89.133.529.9
Pietersen 2007 25 4 889 113042.33 78.735.628.0
Dhoni 2009 24 7 1198 140070.47 85.649.942.7

Measuring performance over a year is a good analysis since the year is a sufficiently long time to give due respect to high level of performances. It is possible that during the 1980s a player might have only played 20 matches while a player might have played 30 matches during the 2000s. However since this table uses runs scored as a basis the lean years do not figure in the same.

Tendulkar's 1998 must rank among the greatest achievement by a player across formats. Even though he played only 33 matches, he aggregated 1894 runs at better than run-a-ball. The Index value of 58 is certainly Bradmanesque. Ponting's 2007 performance, the year during which Australia won the World Cup, is second to Tendulkar's and the Index value of 54 defies description. Richards, during 1985, suffers only in comparison with these two giants. It is nice to note that, arguably the best three ODI batsmen of all time, feature in the top positions in this table.

7. Best/Worst streaks: of 15 matches
BatsmanBest StreakWorst Streak
StODIYearRunsBallsRpIS/RIndexStODIYearRunsBallsRpIS/RIndex
Richards 19840253 994 108166.27 92.060.919880485 291 37419.40 77.815.1
J Miandad 19870417 906 125360.40 72.343.719770041 250 41616.67 60.110.0
Crowe 19920734 725 88248.33 82.239.719850305 337 50722.47 66.514.9
M Waugh 20001620 846 101556.40 83.347.019890549 271 34618.07 78.314.2
Tendulkar 19981323 1105 110473.67100.173.719930795 209 30913.93 67.6 9.4
Jayasuriya 19971207 922 74561.47123.876.119900623 112 163 7.47 68.7 5.1
Lara 19940947 934 98262.27 95.159.219981364 258 35717.20 72.312.4
Inzamam 20001580 744 103549.60 71.935.719961095 251 39416.73 63.710.7
Flower 20021814 719 85447.93 84.240.419950982 248 38016.53 65.310.8
Bevan 19981300 740 92949.33 79.739.320021802 309 43420.60 71.214.7
Ponting 20072473 925 100861.67 91.856.620082687 342 45222.80 75.717.3
Kallis 20032029 893 109759.53 81.448.520052244 315 55521.00 56.811.9
Gilchrist 20032052 860 74657.33115.366.120062434 306 34220.40 89.518.3
Pietersen 20042193 786 79052.40 99.552.120092827 268 35217.87 76.113.6
Dhoni 20092815 773 85851.53 90.146.420103030 311 50120.73 62.112.9

This is an analysis based on 15 consecutive matches. 15 represents about 4-6 months of cricket and is a very clear indication of short-term form. And I have not even bothered with not-outs and average for this short period since three not-outs out of 15 will distort the figures considerably. So it is only RpI and Index based on RpI.

Tendulkar during 1998 features again. A golden run, rather platinum run, starting on 20 April, 1998 and ending on 28 October, 1998 produced an unforgettable sequence of 38, 143, 134, 33, 18, 100*, 65, 53, 17, 128*, 77, 127*, 29, 2 and 141: 1105 runs in 15 innings at a strike rate of almost exactly 100. The first two centuries were those famous Sharjah blitzes against Australia. Richards' golden run is equally noteworthy considering the run-scarcity during the 1980s. He scored 67, 189*(!!!), 3, 84*, 47, 98, 49, 103*, 30, 74, 51, 46, 68, 9 and 76. A total of 994 runs off 1081 balls, striking at 92. The 189, considered by many as the greatest ODI innings ever played, is the jewel in this crown. Ponting's sequence only suffers by comparison. His 2007 sequence was 82*, 10, 5, 51*, 111, 104, 75, 7, 113, 23, 91, 35, 86, 66* and 66. The aggregate was 925 runs at a strike rate exceeding 90. The later part of this golden run was during the 2007 World Cup.

Jayasuriya's worst streak during 1990, a run of 4, 1. 23, 5, 0, 4, 1, 26, 0, 0, 5, 3, 32, 5, 3 is something to behold. A mere 112 runs in 15 innings. Let me also add that he compiled 271 runs in 30 innings during this disastrous period. The next lowest is Tendulkar during the barren period of 1993. Note the consistency of Ponting. Even during his worst streak period, he aggregated 342 runs. Martin Crowe, against better bowlers, was nearly as consistent, aggregating 337 runs. Dhoni is in a surprise third position with 315 runs.

It is interesting to note that three other batsmen have exceeded 1000 runs in 15 innings. Hayden, with 1101 runs off 1108 balls during 2007, Amla, with 1083 runs off 1070 balls during 2009-10 and Kohli, with 1003 runs off 1054 balls during 2012. There are three instances of competent batsmen scoring below 112 runs in a 15-innings streak. Samuels had a nightmare run of 90 runs in 15 innings during 2006, Pollock compiled only 105 runs during 2002 and Boucher scored 109 runs during 1999. It may seem obvious but it is clear that when a batsman is in great form he scores much faster than when he is in miserable form.

A shot-in-the-dark measure, not based on any scientific reasoning, is the difference between the best 15-innings RpI and worst 15-innings RpI figures. Tendulkar leads in this measure with a difference of around 60, followed by Jayasuriya, with a near-55 figure. Richards has a value of around 46. Readers can make what they can of these numbers.

In Part-2 I will be covering the following areas of analyses. If any reader comes out with a good suggestion it can be incorporated. I request that the readers do not ask for changes in players. I have completed a part of this analysis with one set of players and cannot abruptly follow-up with another set. The next part covers mostly performance oriented measures.

8. World Cup SF-F Champions Trophy Finals / Significant / Early matches
9. Batting Position / Boundary analysis
10. First & Second Innings analysis
11. Home / Neutral / Away analysis
12. Won / Lost matches analysis
13. Impact Inns / High Scoring Index / MOM analysis
14. Team share of runs/balls

Since I see many new readers I have to make my usual pitch regarding this specific blogspace, a small corner of "The Cordon". Please bear with me for this, just once.

I am always welcome to criticism and negative comments. In fact I value these a lot since new ideas flow and I can correct any blinkered views I have. But there is a clear line of propriety drawn for this particular blogspace. I had established that with "It Figures" and I had the best collection of informed readers who I respected and whose respect was my reward. When you come into this blogspace you are expected to follow these ground rules since you are here by invitation. If anyone makes a rude or insulting comment it only reflects his own shortcomings, not mine.

You will not insult me, any fellow reader or any player. That is the single commandment here. I have no problems with the following observations:

- Your analysis is based on a wrong premise.
- Your computations are wrong.
- Your analysis is statistically weak.
- You have over-complicated a simple issue.
- You have over-simplified a complex issue.
- You are not being fair to *.
- You have tried to solve a non-existing problem.
- Reader * is wrong/has not understood.
- Player * should retire.

I have problems with the following statements (and the like) and your comment will be junked instantly.

- You are stupid. Even if you have a Ph.D. in Statistics, you have no right to say that. I could, but never will, counter with a retort in far more colourful language.
- You favour players from *. Because, to me all countries are same.
- You favour *. Because I do not, and always leave my personal preferences aside.
- Reader * is stupid. Because he is not.
- Player * is selfish/greedy. Not in this forum.
- You are biased. I may publish this if I feel I can provide additional insight by answering this comment.

Finally let me say this unequivocally. I write for the average and interested cricket enthusiast who may or may not possess any serious statistical/mathematical qualifications. I do not write for the statistical/mathematical experts. Although I possess graduate-level statistical knowledge, I limit myself to Mean, Median, Mode, Standard Deviation, Quartiles, Normal Distribution and the like. I know that I will lose 75% of my readers the minute I bring in p-Value or z-Score.

I have only one avowed objective. I want my article to be understood by 95% of the visiting readers. I am not writing to get my articles published in the conferences of RSC or ISI or ASA. I am certain they would not be. Common sense is the cornerstone of my articles and I am proud of that. If I do not meet your expectations, my apologies and if you feel very strongly about it, au revoir.

Let me close with a short story. I had an excellent reader during the early years of "It Figures". Sometime back he stopped following the blog for various reasons. But he was never rude or insulting to me. I thanked him and said that the doors of this blogspace would always remain open since I valued his insights. I see that he has made a comment for the first article - to the point and that too, a point well-made. I never have any problems with readers like him.

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

RSS Feeds: Anantha Narayanan

© ESPN Sports Media Ltd.

Posted by Buldozer9 on (May 13, 2013, 13:13 GMT)

Dear Sir,

Very good insight into the game as such, one common thing I want to bring about is the batsmen mentioned would have nothing other than the team in their mind while performing on the pitch, great guys. Can you please tell us also about Bridjesh Patel and Krish Sreekant, who basically triggered the one day habit for India. After 1979 world cup (India placed last) the Srilanka tour of India saw the emergence of Cheeka and Roger Binny into one dayers, and the comment of Roy Dias -"Srikant beat us not India".

Please.

Suresh Venkataraman

Posted by Anshu.N.Jain on (April 12, 2013, 6:58 GMT)

It is interesting to see that the two batsmen with the highest Index value in their best 15-match streaks have the lowest in their worst 15-match streak. While Jayasuriya could well be termed inconsistent, the same could hardly be said of Tendulkar. Yet, this is what the numbers throw up!

Posted by Anshu.N.Jain on (April 12, 2013, 6:53 GMT)

Hi Ananth,

On the vexed question of Not-outs, and by extension the measure of RpI: The problem, if it exists, clearly impacts ODIs more where over resources are limited. I have the following suggestion to make to deal with this:

1. Substitute each not out score with the mean/median of all dismissals at scores above the not out score (I prefer the median) 2. If the highest score is a not out score, take it as is. 3. Add all such scores to the total runs scored in dismissals. 4. Derive the Adj RpI using value in 3. above as the numerator, and all innings played in as denominator.

Moving on to reading part 2 of the article now :-)

Posted by Rohan1 on (April 8, 2013, 4:06 GMT)

Contd... These 17 Grand Slams have been attained by dominance over a period of several years. Unlike cricket these years are forever "locked in"…Completely opposite to what you claim - Federer's "final average" simply does not matter - not one whit.

Federer's "final average" will not have the slightest impact whatsoever on his extended glory years - the same emphatically cannot be said about Tendulkar and other cricketers who's overall average simply doesn't tell the tale.

Ahmer Raza's point is very ,very valid and cannot simply be shrugged off.

Posted by Rohan1 on (April 8, 2013, 4:05 GMT)

Ananth, I couldn't possibly disagree more. The reason the number of GS titles in tennis is the ultimate mark of greatness is because it effectively = "Number of years of extended dominance over peers".

Since, there are only 4 GS per year to collect many titles you have to be better than your peers for YEARS. Odd flashes of brilliance won't do.

THIS is effectively what Ahmer Raza is pointing out. Tendulkar had an extended period of dominance over his peers for an extended period of a decade.

This is something which very very few sportsmen can claim.

If Federer's "average" GS performance is now "Quarter final" - this is 100% sure to decline if he goes on since he is past his peak. If he goes for another say 9 years ala Connors he may end up with an average of "3 ½ or Fourth round"…Another player still at his peak ( assume Djokovic for now ) may at that point have a higher "average" …But ,unlike Federer, he doesn't have 17 grand slams. ...Contd...

Posted by Meety on (April 7, 2013, 12:08 GMT)

Hi Ananth, I think this has been said before, but I think Not Outs should not go to the RpI, if they haven't had sufficient chance to bat a full innings, (think this is what viksingh was saying). So in the case of Bevan, whilst I know his Not Outs skew his raw average, but maybe not outs where he didn't face his his average deliveries get excluded? I just think it is harsh to count an innings as "out" - when the batsmen could of scored 35 runs (say) & guided his side to victory. == == == On Bevan, I think people will look at his stats & say he was selfish - batting so low with a s/r so low. It won't be until they would see a highlights package, to realise how many tight situations he pulled Oz out of. == == == Glad you included Mark Waugh, IMO - he was a better ODI opener than Hayden for Oz. I know Indian fans would put up the Sachin/Ganguly combo as the greatest opening duo's of all time, & Grenidge/Haynes fans would have case, put IMO it'll always be Gilly/M Waugh!

Posted by   on (April 4, 2013, 8:07 GMT)

Dear Ananth

Will the rankings change if we consider the scores of these batsmen while chasing the large totals say in excess of 250 plus?

Thanks Nagesh

Posted by   on (April 2, 2013, 10:22 GMT)

Hi Ananth,

Small Correction in fourth table you mentioned it as "The new serious entrant to this table, Inzamam scored 34.6% of his team runs." But from the table 34.6 is Andy Flower but not Inzamam.

You idea of stats is good as you are talking about other players when they excelled in particular section I would love to see comparing them with Dravid, Ganguly and Gavaskar so that we will have an idea where these guys have excelled too. Am not saying to include them in the current player list but when certain instances the above three players excelled well just let us know.

Posted by   on (April 2, 2013, 9:14 GMT)

Above all Zaheer Abbass missing iin the list.

Posted by   on (March 26, 2013, 10:57 GMT)

...On the contrary, the N.O high scores indicate a batsman playing well. For eg. using this method eliminates Lara's 400*.

The best way to go is to use average...with another column showing the number or N.Os (for whatever they may denote). Readers may then use their own judgement.

Another idea , instead of using best 80 innings or best 70 dismissed innings may be to use percentages .

This is the only way some sort of comparison may be made with the incomparable Bradman without it being too contrived such as the 80 innings figure. (Bradman had 7 ducks in his 80 innings- so not fair)

The percentage method would be such as to use the best 10%, 20%. 30% etc innings...And yes use the regular average ( no need to complicate things)

So, We can then compare the top 10% innings of say Richards with the Top 10% of Bradman...with the least amount of distortion as compared to other methods.
[[
Once the box is opened many new ideas will come through. Your idea is one such good one.
Ananth
: ]]

Posted by   on (March 26, 2013, 10:51 GMT)

Frankly, the "Not Out" issue is being blown out of all proportion primarily because of one batsman - Lara. Lara and Sehwag have very ,very few purely defensive innings in their entire careers. Their greatest innings are a result of their refusal to do anything but attack. This attitude produced spectacular results when things clicked, but the very same led to their more frequent dismissals than more cautious batsmen. It is odd to then try to artificially reduce the gap between them and other batsmen by some extrapolated means. Lara retired at perhaps the perfect time as well, perhaps the push for Sehwag may also have been right. SRT s averags is sure to only decline as he goes on and on and on...

Averages as currently used are a reasonably fair measure. The problem arises when we use the average as the holy grail and only important stat.
[[
Problem does not start with Lara. It starts with batsmen like Bevan, Hussey, Dhoni (in ODIs) who have well over 25% of their innings as not outs. In Tests there are middle order players who have well over 15% as not outs. that is the problem.
Ananth
: ]]

As regards the fine new ideas by JeffG - I don't agree at all that the 70 dismissed innings remove the "vexed" issue of Not outs....contd

Posted by   on (March 25, 2013, 22:29 GMT)

I think something which should be added should be dot ball percentage, as per the entire innings. It should be shown for scores between ranges, eg 200-220, 250 upwards. The reason is that for example, Kallis has always been know to bat very slowly, placing pressure on other batsmen. It works well in small totals, but big totals it causes too much pressure on other batsnen, as witnessed in the 2007 world cup match against austrailia. There is an artile about this match, on cricinfo, where it says Kallis faced 33 % of dotballs. Dotballs percentage could also be used to show which batsmen with lower strike rates dont actually waste too much time, they might just get out sooner, with lower averages. I wrote a blog trying to get it published by cricinfo about averages. My feeling based on some minimal stats ive done, is that first of all, by inspection we should remove outliers. Included also the strike rate which is abnormal. for the not outs we could have some weighted value used.
[[
Data not available.
Ananth
: ]]

Posted by JeffG on (March 25, 2013, 12:03 GMT)

Sangakkara's best streak was 4870, placing him 5th all-time behind Bradman and the aforementioned three players. Mohammed Yousuf also had a great run of over 4600 runs. Chanderpaul has the best current streak - 4600 runs over his last 70 dismissals.

Posted by JeffG on (March 25, 2013, 10:00 GMT)

Ananth, I think something went wrong with the previous message I tried to send, so I'll have another go.

My suggestion to get around the problem of not outs is to use dismissals rather than innings. For example, Bradman suffered 70 dismissals in his career for 6996 runs. What is the most runs that any other player has scored in their best ever sequence of 70 dismissals?

I've looked at a number of players and have found just three who have scored more than 5000 runs over the course of a 70 dimissals streak - Ponting, Sobers & Kalis - and Ponting has the most at 5217 (ave 74.53). Tendulkar's best run is 4763 (ave 68.04). I think there's only about 20 players in history who have scored more than 4200 (ie averaged more than 60) over a streak of 70 dismissals.

You can also look at a batsman's worst streak of 70 to see how it compares to their best. For example, Ponting's worst streak was 2556 (ave 36.51) - less than half of his best streak.
[[
If we start looking at it we could very well get a few such gems. For what it is worth "dismissals" at least takes away the vexed not outs.Your work seems correct. If asked without showing these three I would have suggested the same. Tendulkar and Lara did not have such test streaks. What about Sangakkara in 2007.
Ananth
: ]]

Posted by MJ-freehit on (March 25, 2013, 7:55 GMT)

Very well written article and good analysis.

However, when calculation index, you use SR and simply multiply it Average/RpI. I have my reservations about the same. Let me try an explain with an example. Tendulkar scored at a SR of 86 in an era where there were few others also scored at the same pace and some even faster. While, Viv Richards scored at 88 at an era where 5 run per over was considered fast scoring. So, Richards' SR of 88 in his era is far more valuable in comparison to Tendulkar's 86.

Therefore, if there is anyway that the average SR of all batsmen during a player's tenure and/or SRs of all batsmen involved in matches considered for a batsmen are/is also taken into account, then it would be a fairer analysis.
[[
On the anvil is an exhaustive Peer analysis. From the first match Richards plaued to the last match he played I would look at the strike rates across the world and in West Indies. Woud put his 90 in true perspective.
Ananth
: ]]

Posted by   on (March 21, 2013, 13:51 GMT)

@Ananth. Yes, I completely agree - why bring fielding into an analysis of the batting greats? However, by opening up the the Pandora`s Box of team selection, as I think I did, it became something to consider. Unfortunately, one of the biggest fielding dilemmas came in the choice of Gilchrist or Dhoni (an opener or a No.5,6,7). Thanks to JeffG for offering me Gilly, but I might take both. I can`t remember seeing either of them field without the gloves...do they have the stereotypical `keeper "custard arm"? Perhaps I could exchange Bevan for Hussey as well - I do like my top 6 with 3 lefties and 3 righties though. Either way, pretty happy to back my Top 6 , originally chosen, against all comers. Love from Nippon ps. bring on Delhi...hopefully we can finally see a less than completely lopsided match between these teams.
[[
Dave, 15 minutes more of Starc and Doherty might well have saved the Test. My feeling is that many straight-forward deliveries were blocked keeping the scoring rate quite low. In these situations runs matter as much as wickets. But I must say that Hughes' dismissal was one of the worst. Not just that the ball might have missed leg stump, the ball hit the pad outside the leg stump and all three stumps were clearly visible. I am not saying the result would have been different.
To their credit, the Australians bowled reasonably quickly. They could easily have cut out 2/3 overs before 3.30. Ananth
: ]]

Posted by JeffG on (March 21, 2013, 8:57 GMT)

@ Mainak_Sen - at the risk of contuing "off-topic" for this article and of re-opening the debate from Ananth's last artcile, it could be argued that the not outs don't really affect the averages, or might even negatively impact them.

Anyway your comment on Kallis & Sobers inrtigued me so I took a look at their best 80 inns and the results were surprising! Sobers does actually average over 100 - he has 6944 runs at an average of 109. And Kallis best 80 inns average is almost identical to Tendulkar 8784 runs @ 154.

Chanderpaul is another who averages over 150 for his best 80 inns - in fact, the more I look at Chanderpaul's stats, the more my admiration for him grows - of the 27 test match tons he has scored, in 16 of them he remained unbeaten (and probably most of them because he ran out of partners) - how many more runs would he have scored if given the opportunity to finish his innings?

Apologies for moving the focus to test matches in an article about ODIs.
[[
Jeff, Surely for this mini-exercise, we need not differentiate between 400* & 400, 365* & 365 and 248* & 248. Frannkly a better idea, since the 80 is a fixed number, that too, almost all 2+ hour innings, is only to talk of the total runs scored.
Ananth
: ]]

Posted by STNS on (March 21, 2013, 7:56 GMT)

Great Work Ananth!!

Just a bit surprised with the inclusion of Sanath & Flower.

Though you have apologised for the exclusion of some players plus a few others like Saurav, Kirsten, Greenidge, Hayden etc (probably et al includes these names).

Had this blog been on the selection of an all-time ODI world X1, then it would have been fine. Because then i would have considered other abilities of these 2 players as well. Like Sanath was a pretty competent bowler & an oustanding fielder, & Flower was a WK. Both have captained their sides as well.

But since this blog is on ODI BATTING GIANTS, these 2 certainly dont fall into this category.
[[
I wanted this to represent periods and teams. There was also a need to reward team successes. Sri Lanka surely needed to be represented. Jayasuriya revolutionized ODI batting. Okay, not-so-great numbers. Only possibility was de Silva for Jayasuriya. I am sure as a batsman, Ganguly was streets ahead of Flower. But Flower performed admirably well playing with average players right through his career.
Ananth
: ]]

Posted by Mainak_Sen on (March 20, 2013, 19:49 GMT)

(...contd. from previous). Coming to think of the "best 80" numbers for Tendulkar, Lara, and Steve Waugh that JeffG has crunched for us, I guess it is skewed in their favor by the number of Not Outs. I will suggest that if Ananth does this analysis he also uses RPI apart from the standard average. (naturally also for Bradman, 87.4). The respective figures (based on the runs as provided by JeffG) for the three are 124, 115, and 98. So they still "better" the Don, but the gap is reduced. RPI to me is the more appealing measure in this case as we are only concerned with the "best 80". Getting back to Ananth's earlier "streak analysis", I do not recall if he had used something like an 80 innings moving average. If not then that is surely something that might be interesting and illuminating. But in this case the Don will certainly be supreme. Thank you again for your time and comments. I felt nice to participate in the conversation.
[[
Just to give you the numbers Tendulkar has aggregated 9924 runs in his best 80 innings. I would not even fleetingly think of using the average for this. What does it matter whether the score was 248 or 248*. So the RpI for this is 124. The minimum (80th) score is 74. So if you take off this base of 73, the RpI for the 80 innings 50. Bradman's equivalent figure is 87.45.>br> So you can slice and dice whichever way you want. The difference is between 30 and 40.
Ananth
: ]]

Posted by Mainak_Sen on (March 20, 2013, 19:22 GMT)

Ananth and JeffG, thanks for your time and comments. As I said, I know that the "best 80" is a contrived measure. I also understand that "theoretically" you can have someone playing 300 tests and scoring 80 hundreds. My point is (and I guess that is the fun of the exercise) is how many *actually* have managed to beat Bradman? And how long did they take? It is true that the average of the "best 80" is likely to correlate (for specialist batsmen) with the total number of innings (career length), but I do not see that as *necessarily* implying that it would be better than 99.94. The crucial point here is that the Don is clear 40 runs ahead of the next guy to start with! I expected Tendulkar to "beat" the Don, but I did not expect it to be 155. (But more than 4 times the no. of innings). The interesting case is Lara, and he is *only* 119! Even without looking at data I suspect that Kallis will perhaps not beat the Don, and surely Sobers will not! (...contd.)

Posted by FitzroyMarsupial on (March 20, 2013, 17:34 GMT)

Ananth - great article. A couple of comments (1) would love to see the Bat% stats across entire careers (i.e. what %age of the runs scored whilst they were at the crease did they score). I think this would put SR into some context. (2) i do think you've been a bit harsh on Gilly :-) - (e.g. noting strike rate of Jaya, Viv and Dhoni when AG's is superior); (3) any chance of stats for setting total v chasing? Would be particularly interesting to see how the 'finishers' like Bevan and Dhoni fare and (4) On the streak stats (and the variation) - you asked what we thought. I found them interesting as a cricket nerd, but not sure how it 'fits' into the theme of the analysis - short-term form is surely the role of the ICC rankings etc - you have grouped together a list of the finest 15 ODI batsmen over long careers. As you say yourself, in terms of hot streaks there are those outside the 'top 15' who would feature (Amla and Kohli may of course one day merit inclusion). Thanks again. Amit.
[[
Part 2 might bring provide most answers.
Ananth
: ]]

Posted by   on (March 20, 2013, 14:45 GMT)

Hi Anantha,

Thank you very much for this analysis. It really helps point out the variance between the good and the greats in terms the variance between scores.

The one point I would like to include as part of any analysis to be built around the shorter versions of the game is the required run rate which impacts a players innings to a large degree. 2 clearly distinct innings that come to mind are Sachin's innings in Sharjah and the 2003 final. While during his innings in the 2003 final did have an impact, the pressure of the 350 run chase led to him playing a rash shot. In the 1998 innings though, despite the pressure he was able to guide the team.

My inclination to this analysis would be to weight he total runs scored, rate at which they were scored and the asking rate. Not sure how much of this information would be readily available
[[
One, the data is not completely available. When was the dismissed is a key data which is not available for over a third of the matches. Second is that that becomes contextual and the analysis is innings by innings. It does not mean anything at career level.
Ananth
: ]]

Posted by JeffG on (March 20, 2013, 10:18 GMT)

@ Mainak_Sen - Tendulkar has made 118 scores of 50 or more in 325 innings, so just based on those numbers, you'd expect his best 80 inns to have a huge average.

When you look at the data you see that his best 80 inns are all over 74 and include 16 not outs. He scored 9923 runs in those 80 inns for an average of 155.

By way of comparison, Lara had 82 inns over 50 (out of 232) and scored 9201 runs in his best 80 inns at an ave of 119.

Obviously, the more inns you have, the higher your best 80 will be - I would imagine there is a very strong correlation (amongst specialist batsmen) between number of innings in total and runs scored in the best 80 inns.

I haven't checked every player but I did have a hunch that Steve Waugh's best 80 inns would have a high average.

He also had 82 inns of 50 or more (out of 260) and while he scored far less than Tendulkar or Lara in those inns - 7869 - his average was nearly as good as Tendulkar - 148.
[[
In fact I was also thinking about it. Theoretically if a batsman plays 300 Tests he could have 80 hundreds. So this is a non-starter. But a consecutive 80-inns streak has a lot of possibilities.
Ananth
: ]]

Posted by JeffG on (March 20, 2013, 9:30 GMT)

Ananth, I actually had Gilchrist in both teams but i'm happy to let Boll's team have him and i'll have Dhoni in mine - by my reckoning that will cost my team about 2 runs per match on average, but I think my team can cope with that and still win ;-) In fact, I also considered bringing in McCullum instead of Dhoni, to see if McCullum's higher SR makes up for his lower average, however I think my team would score 1 less run with McCullum vs Dhoni.

I also looked at making De Villiers my keeper and bringing in another batsman, but no one really added anything extra (Symonds was the best but his SR is not enough to make a difference to my team.)

As for my teams ability to survive 240 balls - if you look at the Balls/Dismissal data for my original team you get: Afrid (22.1) Sehwag (34.8) Gilchrist (36.4) De Villiers (52.6) Amla (63.8) Jayasuriya (38.9) which makes 249, so on average they would last for more than 40 overs.

Posted by Mainak_Sen on (March 20, 2013, 6:06 GMT)

Ananth, this is following up on my earlier comment from 19Mar. Thanks for your reply despite of it being off-topic. I can recall some of that "streak analysis". Yes, I agree fully that consecutive innings is the more rational comparison. But, perhaps I should clarify: I am one of those to whom Bradman is supreme...I do not think that anyone can even remotely approach him. Having said that part of the fun is to come up with any measure, even a little contrived, to show just how superior he is (may be a little pinch to those Tendulkar apologists :-)). May be just as an extension to your earlier analysis, If we do rank on the basis of the best (not-consecutive) 80 innings, and find a few (I suspect not more than 3-4) manage to surpass him, it will only ADD to the legend of Bradman...for those few would have those 80 innings picked from careers in excess of 200 innings. Thanks for your patience.
[[
Yes, I get your point. Will keep this in mind. It is too interesting a topic to just warrant a response. Probably needs a full article since we can think of other analytical snippets also. Also I have to create an ordered list for each batsman and then sum the top-80. A good thing is that someone like Weekes will qualify since he has played 81 innings.
And let us not forget our dear Chris Martin who has crossed 100 innings or Chandrasekhar who has played exactly 80 innings. For the record Chandrasekhar has scored 2.38% of the runs scored by Bradman.
Come to think of it there is a lot going for this analysis. It has gone to my Inbox.
Ananth
: ]]

Posted by Mainak_Sen on (March 19, 2013, 17:14 GMT)

Hi Ananth, I am a regular reader of your blog, but as far as I can recall this is my first comment. So let me begin with a big thank you, for providing us with such informative and illuminating analysis of aspects of the game we love. This comment is really not related to this current article, but is rather a request for a future analysis. if you think it is worth the effort and you get time. To put it simply, can we do an analysis of all the "great" test batsmen, based on their best 80 innings? (need not be consecutive) I shall call it the "Bradman Measure" How many would be better than the Don? I suspect Tendulkar might be, but then by how far? And who else? I recall you did something similar for consecutive innings, but cannot remember exactly now. 80 seems to me to be a nice number (even apart from its significance associated with Bradman): It is not too small nor too large. And we can use this as some kind of standard for comparing batsmen.
[[
Mayank, when I did the Test streak analysis I did the very thing. Ponting came on top with a 52-Test streak with an average of 74.09. This is a better method. Taking the best 80 innings would be wrong. Then you should change the criteria to the best-50 innings and do the same for Bradman also. We cannot saddle Bradman with 14 failures and not the others.
Ananth
: ]]

Posted by JeffG on (March 19, 2013, 13:18 GMT)

I'm not so sure my team is that brave - in test matches definitely, but not in ODIs where I only have to bat for 50 overs.

Boll's top 6 is great and, based on averages and SRs, they would score 255 in 50 overs and you would probably want to adjust that up a bit to reflect the eras in which some of them played.

My top 6 would score 240 but would do it in only 42 overs, leaving 8 overs for my 5 bowlers to swing the bat - and I bet they would score more than 15 runs in those 8 overs, especially as I might get lucky and have Flintoff at 7, Pollock at 8 etc etc !! Even with a set of bowlers who were only averagely good at batting, i'd bank on them to score at least 30 in 8 overs

Of course, Boll's team would also have Flintoff and Pollock but they wouldn't get to the crease because his top 6 would have used up all of the balls !!!

I'd back my team to beat any other over the course of 100 matches...
[[
I SHOULD do a simulation run of 100 matches between your and Boll's teams. When? I am not sure. Did you take Gilchrist for yourself or gave to Boll when you did the computing. Replacing Gilchrist with Dhoni may not be bad.
Your top six averaging 40 balls per match is something I am not sure of. More often than not they might be 180 for 6 in 32 overs.
Ananth
: ]]

Posted by JeffG on (March 19, 2013, 11:19 GMT)

Ignoring Richards because, while he is my all-time favourite batsman, I haven't looked at the pre-2000 era in this analysis, if I could choose 6 batsmen to get me the highest score, I would choose: Afridi, Sehwag, Gilchrist, De Villiers, Amla & Jayasuriya. And I would bat them in that order - put the fastest scoring players up front.

I definitely wouldn't have Bevan in this team - he scored far too slowly, even adjusted for his era - great if you need to dig your team out of a hole, but no point if you have 5 other world-beaters...

And one final (I promise!) comment on Afridi vs Kallis - while you are right that Afridi would fail more often (41% of his inns have been in single figures vs 28% for Kallis), at least when he does he gets out quickly. Too often when Kallis fails he uses up precious resources - his SR in single figure inns is just 30 - to the detriment of the team - you have to remember that both balls & wkts are resources to use or waste...
[[
That is one brave team. If I remember Boll's team, was it Tendulkar, Gilchrist, Richards, Lara, Ponting and Bevan, That would be a match to remember, I can say. One of you has to give up Gilchrist!!!
Ananth
: ]]

Posted by Charith99 on (March 19, 2013, 11:17 GMT)

its interesting to note that 6 batsmen out of 15 whom you have selected loves the indian bowling attack maybe it shows the poor standards in indian bowling or that you remember them more because they scored heavily against the indians.
[[
News to me. I never think about these things. Anyhow how can anyone remember matches out of 3000 or so that one has followed or watched (mostly on tv) or read about. Probably 100 matches, if someone has good memory, may stay in memory.
Ananth
: ]]

Posted by JeffG on (March 19, 2013, 9:57 GMT)

Ananth - one final comment on the benefits of an Afridi type player...

if you take 2 hypothetical teams - both containing "average" batsmen in positions 7-11 but one has the top 6 made up of 6 Afridi's and the other having 6 Jacques Kallis's - what would those teams score in an average ODI inns?

The team of 6 Kallis's would score on average 223-4 in 50 overs (based on Kallis ave & SR since 2000)

The team of 6 Afridi's would be 143-6 after 22 overs and then the lower order would get them to 216 all out after 39 overs - so the Kallis team would score more.

But what if we replaced 1 Kallis with 1 Afridi? That team would score 230-5.

A team of 2 Afridi/4 Kallis would score 238-6. A team of 3 & 3 woudld score 247-7 A team of 4 Afridi/2 Kallis would score 254-9 A team of 5 Afridi/1 Kallis would score 239 a/o in 46 overs.

So it would seem that to get the best score you need a mix of fast, risky players and solid accumulators - with a slight bias towards the faster type
[[
If we factor in off-days or "which Afridi would walk in today" then the Kallis-centric teams might have a distinct edge. If I simulate 100 matches with 6 Afridis and 6 Kallis's I can assure you that the 6 Kallis' teams would fail less often.
Again this is Afridi opening and dominating the top-order positions. I would be more than happy with him at no.7.
Let us agree that we neither need Kallis-clones nor Afridi-clones. We need a perfect mix.
Let us settle on Boll's team.
Ananth
: ]]

Posted by jhabib on (March 19, 2013, 0:38 GMT)

Mr. Narayanan, I appreciate that you've taken the time to write this detailed and impressive analysis. It shows that a lot of time and effort was spent in carefully compiling these statistics.

I have no issue with the premise or results of your analysis. I however object to the rhetoric on reader's feedback and what is accepted behavior. Insults notwithstanding, I personally believe that readers should feel welcome to state their opinions no matter how crude, ill-educated or badly stated they may be. Sometimes even simpletons like me can have a valid point but not the vocabulary nor the mastery of language to state it elegantly.

Please take this only for what it is, an opinion. I would still value your insights as much as I do now even if you did not have a Ph.D in statistics. I would however still state my opinion without regard to "guidelines".
[[
Without guidelines if someone called me **** or called you **** would you accept it. For your information out of 80 or so comments received, 3 have been refused permission. Let me also assure you simple language is fine and wonderful. If you have an innate respect for me and my work, I am certain you will never cross the line. I am not looking at colourful language but acceptable language.
Ananth
: ]]

Posted by JeffG on (March 18, 2013, 22:48 GMT)

Sorry Ananth, I obviously didn't make myself clear in my previous comment.

I actually think that Afridi does make a good opener (if you take his career figures rather than recent form.) In fact, given an "average" ODI team, I would always have an Afridi type opener who scores 25 at a run a ball, rather than another player who scores 40 off 50 balls.

In your example of 6 players who score 40 off 50 balls for a total of 240 in 50 overs, I would rather replace one of them with Afridi. These 5+Afridi would then score 225 from 275 balls and then the lower order would have 25 balls to bat - and on average they would score more than 15 runs and so my team score for 50 overs would be higher than the 240 scored in the original scenario. They will have lost an extra wkt, but that means nothing at the end of the day - 241-7 beats 240-6 every time in ODIs.

Posted by nilbadwe on (March 18, 2013, 19:36 GMT)

I really love your articles... They give me inner peace.. However was surprised to see here that you missed Kallis in worst streak stating Dhoni to be 3rd. Kallis has 2 runs more than Dhoni during his worst 15 match streak. Please rectify that.
[[
Since I no longer have the mechanism to correct the article, please accept my apologies and correction here. Kallis' worst streak is 315, just ahead of Dhoni's 311.
Ananth
: ]]

Posted by   on (March 18, 2013, 18:45 GMT)

Afridi has won less MOM awards for just his batting than Inzamam and I mentioned Afrida THE BATSMEN. The Index metric only deals with batting. Also the fact that you are "happy" with RPI doesn't mean its correct.

Also Afridi's MOM awards could be exagerated because he played for a weaker team than lets say Wasim who had Waqar and Inzamam sharing the effort.

This index measure is not offering any insight. If anything it is skewing the numbers. The Average and Strike Rate offer insight. Mulitplying them doesn't offer any additional insight.

Posted by JeffG on (March 18, 2013, 14:24 GMT)

Ananth, I slightly disagree on what an opener (or indeed any batsman) should contribute. It's all about using the 11 players to take best advantage of the limited resources of 300 balls and 10 wkts.

Given that the average ODI innings only sees 7 or 8 wickets lost after 300 balls there is usually some resource (wkts) unused at the end.

If we take an opener who averages 40 at a SR of 80, then he costs his team the resources of 50 balls and 1 wkt and gains his team 40 runs.

Now, if we take an "Afridi" type opener who averages 25 at a SR of 100, then he costs his team fewer resources (25 balls & 1 wicket) but gains fewer runs (25) However, this leaves more resources for other players to use (ie 25 balls). Now, assuming an average inns, it will be number 9/10/11 who will have to face those 25 balls. They score at an average SR of about 70, so will score about 17 runs. So the net gain for the same resources over all is 25+17=42 runs - in other words, on average, the team scores more
[[
Jeff, I am not sure why you should disagree with me when I have concurred with your suggestion. I have clearly told that Afridi does not fit in with my idea of an ODI opener for the very reasons you have suggested. My suggestion of a 40/80 opener will work well and that is what you have suggested. A 40/80 opener (or two) will occupy 50 balls (each) and score 40 runs. 6 such bastmen will score 240 runs. If add, say 15-20%, for power plays and wicket availability at the end you get 270-280 innings which is an above-average innings.
Ananth
: ]]

Posted by   on (March 18, 2013, 12:07 GMT)

I don't understand why Average and Strike rate are being given equal importance. Please explain. This is what I have issues with. You can't just multiply two metrics and say this is a new metric which is very valuable. Alot more thought has to go into this. Afridi has a better index value than Martin Crowe, Miandad, Inzamam and Flower.Does this mean Afridi as a batsmen in a non context "clearly" more valuable to Pakistan than Miandad and Inzamam? No way!!..Inzamam could definitely score at an Afridi like strike rate (World Cup 92). Afridi can't play an Inzamam like cool and collected innings. I still don't agree that RPI is a better metric than average (last article). When a player goes Not Out it shouldn't count in the denominator. A Not Out player loses the potential of scoring more runs. The fact that Tendulkar went Not Out when he scored his double hundred doesn't mean that innings "doesn't exist in his average calculation"..it definitely exists in the numerator.
[[
Andrew, Thanks for asking cricketing questions. i never have any problems with those. Let us see Afrid's value to the team. He has won 30 MOM awards, unscientific and subjective these may be, but surely reflect a player's value to the team. He wins a MOM award every 11.6 matches, more frequently than Ponting, Inzamam, Wasim Akram, Shaun Pollock, Miandad to name a few. Maybe some of these awards might have been for bbowling but surely his place as a valuable player cannot be questioned. Afridi has won 30 MOM awards, Inzamam, 23, Miandad, 19. Where did I say that the higher the Index the greater the player. These are measures which offer you additional insights. Jeff might look only at averages, somone might only look at strike rates, you might look at a combination of the two, not necessarliy equally weighted. And you are the one who has determined Afridi's index, found out that it is hgher than a, b, and c and asking me who is more valuable. I never said that.
I have used RpI here because of the high number of not outs in ODIs. It may be unfair to the late order players but probably favours the batsmen 1-5. I am happy with that. Ananth
: ]]

Posted by JeffG on (March 18, 2013, 11:43 GMT)

Cont… however, when bats at number 2, Afridi scores on average 25.52 and takes 25 balls to do it.

The average opener scores 32.73 and takes 42.5 balls to do it.

So, by Afridi opening, he exposes the tail to 17.5 extra balls, in which case they would score about 12 runs.

BTW, I'm using the average ODI score since 2000 of 230-7 as my estimate of which position would have to face those "extra" balls - generally numbers 9 and 10.

So Afridi's team value as a number 2 above the average opener is (25.52+11.88)-32.73=4.67 runs.

So, on average, a team would score more runs with Afridi opening than with an "average" opener, but he is more valuable to the team when batting at number 7.

Of course, this is purely based on batting and ignores any bowling value.

Generally speaking, with this method, any player who is 10+ runs above average is a "great" player
[[
Jeff, I am very clear that Afridi should bat low down since a 20-ball-25 is far more valuable at no.7 than the same at 1/2. In my opinion openers should be 40+ average batsmen but should have s/r greater than 75. There is no place for a 40+/65- batsman today.So Afridi does not fit in the opening slot..
Ananth
: ]]

Posted by JeffG on (March 18, 2013, 11:29 GMT)

Ananth, I totally take your point that sometimes a slow 50 is what is needed, whereas on other occasions a quickfire 30 is the ideal score, but, over time, things should even out and looking at "average" data should give us a clear picture.

Taking Afridi as an example and looking at the data since 2000 where his 2 most frequent batting positions have been number 2 and number 7.

The average number 7 scores 22.85 runs per dismissal and takes 29.5 balls to do it. Afridi at number 7 scores 23.72 and takes 18.1 balls to do it. So, by getting out 11 balls earlier than the average, this means that the number 9 or 10 batsman generally has to face about 11 extra balls.

Now, the average 9/10 lasts 18 balls and scores at 0.71 runs per ball. So in 11 balls they would score about 8 runs.

So Afridi's team value as a number 7 above the average no 7 is (23.72+7.83)-22.85=8.7 runs. In other words, when Afridi bats at number 7, his team scores nearly 9 more runs than it would without him

Posted by JeffG on (March 18, 2013, 9:11 GMT)

Cont… the method is something that i'd picked up from baseball - a measure called "value above replacement". I thought this was a good way of measuring value in ODIs, given the time constraints and the fact that a slow scoring player may actually be costing his team runs, even as his own average increased.

The basic premise is that you measure the difference in runs a team would score if you replaced him with an "average" player (adjusted for his era if you want to.) Take too long over scoring and you use up resources that other players could exploit better. Get out too quickly and you risk exposing players of lesser ability too early. You have to find the sweetspot.

I last looked at this in 2010. Without going into detail, this method produced a top 3 at the time of Sehwag, Gilchrist & Dhoni - suggesting that SR was relatively more important than average (Richards, Tendulkar, Hussey et al did well but not top 3 - also Afridi (at the time) did well too...
[[
Jeff, my take is that an analysis like what you have suggested can only be truly valuable if done match by match, taking in context. Take the last two ODI matches played. In yesterday's one, Afridi's innings gave Pakistan some hope. If he had continued for 3/4 more overs who knows what could have happened. However in the previous match, it needed a calm and collected innings from Misbah to win the match. So the context is very important. Even today low totals have to be chased.
Ananth
: ]]

Posted by JeffG on (March 18, 2013, 8:40 GMT)

Hi Ananth,

Firstly, many thanks for the kind personal words (if i've assumed correctly...) - as you know, i have always valued your articles even if i've not had the time to respond. We might not agree on everything (not outs being one!) but how dull would it be if everyone thought the same? The important thing is that everyone follows the golden rule - "play the ball, not the man".

Now, as for ODI batting... it's an area that i've personally spent lots of time on. I briefly touched on it in one of my comments to your last article, but the key for me in determining greatness is getting the balance right between average and strike rate. In tests, where time is less of a concern, average is all-important and I put little or no weight on SR. In T20, it's the opposite - i'd rather a player score a very quick 20 than waste time over scoring 30 so SR becomes king.

In ODIs, there is a fine balance between them...and I haven't cracked what the exact balance is yet.. TBC

Posted by   on (March 18, 2013, 6:32 GMT)

where is lance klusener???

Posted by   on (March 18, 2013, 3:31 GMT)

Hi Anantha, iam a graduate of ISI but I am not going to challenge the analysis :-). They are great. Just want you to look at a few more tables. One, I want you to normalise the RPI and Index based on the scoring rate and average of the era in which the player played. The scoring rates have increased drastically from the time Richards played. I am guessing player averages have improved too. Two, I want to see how these players fared against the best teams and best bowlers of their time. Three, I also want to see a comparison of these guys with the next best player of their generation as well as the next best player on their team. Four, the NBA uses a metric where they judge a player based on the point differential for his team with him and without him. For cricket we could look at the run rate differential for the batsman's team while he was out in the middle and when he was in the dressing room. Don't know if this data is available though.

Looking forward to part 2.

Cheers
[[
What you are suggesting is possibly a full-blown all-batsman peer analysis. That sort of intuitive tweak will be wasted on a limited-batsmen analysis. Many thanks.
Ananth
: ]]

Posted by Mitcher on (March 18, 2013, 2:40 GMT)

Anantha, I almost want to insult you just to elicit one of your brilliant return barbs. Love your work! Appreciate the analysis. For mine, the stats just confirm the absolute greatness of Viv. To be the dominant force across eras in most of the key areas when he played at a time with lower Avgs/SR is just astonishing. Also found Gilchrist's effort to maintain a near-90 SR during his worst streak interesting. In or out of form, you could always count on Gilly to come out and flash the blade.
[[
There are others ready with their own attention-grabbers. And if they do not cross the line, you will get your wish. For the present stay with me, as you are now.
Ananth
: ]]

Posted by   on (March 18, 2013, 1:42 GMT)

I'm a little puzzled Bevan fares relatively poorly in the RPI and Index stats given that he was critical in many Australian performances. Might his figures stand out if the analysis is restricted to victories when batting second?

I hope you will excuse this question, he is from my home town and we all think he was pretty good.
[[ Bevan's RpI figures, as those of Dhoni's and Hussey's, take a beating because of the very high % of not outs. And finishing an innings has relevance even in first innings. But at no.6, these three are a class apart. Ananth: ]]

Posted by   on (March 18, 2013, 0:59 GMT)

Ganguly and Anwar should have been there.

Posted by   on (March 18, 2013, 0:38 GMT)

Awesome Anantha. I am an infrequent reader of your blog, but you have a regular one now. Fantastic work, thanks for the analysis. I think it may be worthwhile to further assess value of batting for above players, in an ODI, from further two points: (1) setting vs chasing a target, to indicate the impact of changing condition and temperament and (2) individual runs as a percentage of the total and index its frequency (e.g. >33% percentage as significant or something similar) to indicate "consistency" in player's value to the team. For this if you consider every innings as a percentage of runs scored while the player was at the crease, it should be comparable between top order and late order batsmen. I guess this may also allow players across eras to be compared as well. I do agree Haynes, Greenidge and Hashim Amla are significant misses, but take your point about selecting players. Thanks a lot.
[[
Most of what you ask is in Part 2. I have a nifty little idea called "impact innings" which is something new incorporating % of Team score and Relative S/R.
Ananth
: ]]

Posted by   on (March 18, 2013, 0:21 GMT)

Thought this was a great analysis, Richards in particular stood out considering how much lower batsmen's strike rates were in his era.

I was curious about other players performances using your index, in particular 3 batsmen who in my opinion could become greats based on their ODI careers so far: Amla - 64 Innings, 3411 Runs, 53.30 RPI, 92.18 SR, Index 49.13

De Villiers - 131 Innings, 5372 Runs, 41.01 RPI, 93.71 SR, Index 38.43

Kohli - 95 Innings, 4054 Runs, 42.67 RPI, 85.96 SR, Index 36.68

Posted by ravi.m on (March 17, 2013, 23:13 GMT)

Considering Hussey's WC failures, I like the fact that you picked Mark Waugh ahead as much as that man's on-drives!

Moving on, I know you don't like to leave out weak teams, but could you do something about non-Test playing nations. I guess it's kind of variable with BD, ZIM & SL. While I acknowledge Kenya making the semis in 2003, Ireland beating England in 2011 etc., stats are heavily skewed in certain batsmen's favour when you include minnows.

Posted by   on (March 17, 2013, 22:59 GMT)

Anantha, good try. As with all your articles this is a mish mash with no real value addition. I genuinely believe you lack understanding of Statistics.
[[ Thanks, Andrew, for the compliment. If you had endorsed my article, I would have been worried since that means I would have lost a majority of the readers. For once try and undertsnd that every article need not be a statistician's delight. The purpose could be to bring out new cricketing insights. Take off the coloured glasses. Ananth: ]]

Posted by PadMarley on (March 17, 2013, 22:37 GMT)

Brilliant analysis. The only factor I would be careful making Jayasuriya including in batting statistics is that big part of his early career he was a lower middle order batsman [A more of a bowling allrounder] coming to bat expecting to slog with only few overs left. All the other chosen names were occupied proper batting slots in line-ups [Dhoni and Bevan often came in as the last recognized batsmen].

Posted by KiQa55 on (March 17, 2013, 22:32 GMT)

This is not a comment about making you change your player selection, but one seeking an explanation for your selection. From the little explanation provided, I cannot tell what the criteria for selection were. Statements like "Andy Flower [is his] country's best batsman" says you have gone for variety, trying in include as many countries as possible; but this clashes with the objective stated, to pick the 15 best ever. Surely, instead of purposely discounting Hussey or Ganguly, they could have been included instead of Flower (nothing against Flower, but Hussey/Ganguly/others I'm not thinking about are better than him).
[[ The title is "Odi Giants" not "The best ODI batsmen of all time". If and when I do the later, I will work solely on figures and England and Zimbabwe might very well go unrepresented. I keep on coming back to Amla. If he continures like this for a couple of years more, he could upstage the big-3. There is sufficient information provided on the selction criteria. Ananth: ]]

Posted by   on (March 17, 2013, 21:56 GMT)

The last review .. the best 15 performances and worst 15 performances seems a bit unfair to these players. Any player would undergo a lean period where they just cannot get runs under their belt. And for opening batsmen like Sachin and Jayasuriya, they are facing the opposition's best bowlers as well. Rather, it would be better to find the top 10 "Best 15 performances" and "Worst 15 performances" for each player. This would give a measure of how often they have been in good form or bad form and how it has affected their performances in turn depicting their value in the side..
[[ Life is a mix of the bitter and the sweet. If we are able to appreciate the mother-of-all-years or periods that Tendulkar had during 1998, we should also be interested in the lows of 2003. Also understand that these are facts presented with no put-downs. I have only referred to jayasuriya's nightmare run. Someone has given an explanation. Also it lets us know that even when he was playing poorly, Ponting did reasonably well. Same with Martin Crowe. These are new insights. Ananth: ]]

Posted by   on (March 17, 2013, 20:57 GMT)

I think the scope of the article is a little wide. There are too many fantastic batsmen who played in different positions and with different roles for the same statistical parameters to be applied to them. Would love to see an article dedicated to blitz opening batsmen (the Jayasuriyas, Gilchrists, Gayles, Swhwags), one to mainstays (Rischards, Miandad, Crowe, Tendulkar, Ponting et al) and one to the finishers (Bevan, Dhoni, Hussey etc.).

There are so many wonderful nuances that can be explored. For instance: (a) Opening Blitzers: the Run rate when they got out (provided they had scored a minimum number of runs) (b) Mainstays: the number of runs score dby the team by the time they got out (c) Finishers: The number of runs already scored when they got in, and what the team ended up with.

Of course, ther emay not be enough information and/or this may be too complex to do, and this may just be wishful thinking on my part
[[ As I have already mentioned, this is a broad coverage of a selected set of batsmen. At a later date I may very well go through an analysis of all the batsmen in a few selected areas. That is the way the articles evolve in this blogspace. Ananth: ]]

Posted by   on (March 17, 2013, 20:52 GMT)

Good Analysis. I would agree with almost all picks except that i would have liked to see aravinda, AB De Villiers and huss there. But again am sure you would have calculated there stats and hence narrowed down to the top 15

I have a couple of questions. Cricket has seen a lot of changes as you pointed out in a passing manner. Earlier 200 was thought to be a good total and hence the strike rate was much lower. Also the bats, pitches and powerplay rules etc were different. So don't you think that the players who played pre 92 except Sir Viv should be given a leverage or a subsidising %age be added to there index? Also, Bevan and Mahi both played in the end overs so they have/had higher strike rates and averages then a the consolidators as in case of early wickets the consolidators like Kallis had to to settle down and lay a platform. How do you suggest one should provide for that? Something like a value for players as per there role in the team. Fantastic article nonetheless. Cheers
[[ In a way, the Index is a solution. Dhoni's higher S/R will be compebsated by his low RpI. An opener's high RpI will be compensated by the relatively lower S/R. The idea is to look for additional insights by going through the individual tables. Ananth: ]]

Posted by Bonehead_maz on (March 17, 2013, 20:50 GMT)

@ Boll " Sobers and Bradman show potential but have yet to impress... " ROFL Yep only a blob between them ! Suspect B Richards would have gone quite well in this format too.

Having said I was never much a fan of ODI's is not quite correct. I quickly lost interest is probably more accurate, so I looked up Clive Lloyd and David Gower just to see how they went.......... their numbers make Richards' numbers look even better than they already did !!

@ Ahmer Raza

"arvinda de silva should be in the list if inzamam is, arvinda was amongst the best batsman in the mid 90s and single-handedly won the 1996 wc semi-final and final while inzaman was never the top guy in odi cricket. "

Inzaman single-handedly ? won the sf and final of the 1992 WC (although he was batted beautifully into position to do so by Miandad and Imran)

Posted by bored_iam on (March 17, 2013, 20:17 GMT)

Hello Sir. Old time reader, first time posting a comment. I had an additional request. Like you have done percentage analysis for while the batsman was at the crease, could you do one as a percentage of total team runs. I suspect that the openers should have higher percentages, as expected. However, I would really like to see the contribution & dependence that players such as Lara, Tendulkar had to their team's cause vs say a Mark Waugh or Kallis, who had good batsman to surround them. Another aspect of the team balance based analysis.
[[ In Part 2. Ananth: ]]
Additionally, could you please do one for bowlers too?
[[ Will do. Ananth: ]]
Thank you again for the wonderful article! Look forward to part 2!

Posted by   on (March 17, 2013, 19:34 GMT)

@Ananth: I see that you've already responded regarding Zaheer Abbas and your reason is too few matches. I would argue that number of matches is the wrong metric is a wrong metric to disqualify someone as "statistically insignificant". Otherwise someone can put min ODIs at 200 and remove Vivian Richards and Martin Crowe! :)

If a player has played atleast 8 years and has played in > 5% of the ODI matches worldwide, their stats should count - it means they've consistently maintained a high level of performance - its not their fault that the in later years our idea of a "minimum number of ODI matches" got skewed.

Surely Zaheer Abbas was an ODI giant for most the 11 years he played!

Posted by   on (March 17, 2013, 19:32 GMT)

Its best statistical judgement I have read so far.As per my knowledge Ganguly may have a place in Opening statistics.Still its a great effort By Ananth.Thanks a lot for this nice collection.

Posted by SyedAbidHussain on (March 17, 2013, 19:32 GMT)

What about Afridi ?? He has been a big big match winner in ODIs as an all-rounder. Only 9 players have more Man of the match awards than Afridi. You should have mentioned him atleast in the "apologies-to" list :)
[[ My apologies. I am a great fan of Afridi. Yesterday he single-handedly took Pakistan close to an improbable win. But these days are few and far between. Unfortunately i do not have access to the article now. So here I am sending my apologies to the mercurial Afridi for not considering. Ananth: ]]

Posted by   on (March 17, 2013, 19:20 GMT)

@Ananth: Surely anyone with RpI of *42.8*/Avg of *47.6*/Strike Rate of *84.9* should have made the cut ? Don't know who I am talking about ?? Hint: That person played in a whopping 20% of all the ODI matches played in the world in his ODI career spanning over a decade! (So not a flash in the pan, surely).
[[ I have already responded on Zaheer. If Amla/deVilliers/Kohli have not been considered for not meeting my own cut-off, surely Zaheer's fans should understand. But I agree that a strike rate exceeding that of Gayle, Lara, Hayden, Clarke, Gibbs, Smith, achieved at a time when 70 was the order of the day, surely needs recognition in this response as an apology for not including. Ananth: ]]

Posted by   on (March 17, 2013, 19:19 GMT)

Firstly , congrats on your article ,a well thought out one and a more diplomatic one ,including mark waugh is defenitely a master stroke . But as a suggestion ,you could have introduced another parameter to your index by involving the ICC rank of the team (may be a moving average) . Any ways a good article ,keep up the good work.

Posted by   on (March 17, 2013, 18:47 GMT)

Ananth, thanks for responding to my comment earlier...I appreciate it.keep up the good work! E.Shai

Posted by   on (March 17, 2013, 18:39 GMT)

where is aravinda ? what is craw and kp doing there ?

Posted by   on (March 17, 2013, 18:35 GMT)

I am not a great fan of Sehwag and Dravid, but if you look at the contribution of these players for the team's success and results is exemplary. especially Dravid. during the chases Dravid was a silent horse, holding the wicket intact. I am not sure of the records, but he did feature a lot in the chase and in winning cause. In my opinion Dravid could have replaced Crowe/Andy on any other day. just my thoughts.

Posted by Deuce03 on (March 17, 2013, 18:10 GMT)

Something I find rather depressing about this analysis is that only one England batsman (Pietersen) appears in the reckoning at any stage. I say this not because I feel you have any bias against English players, Ananth, but rather because it reflects how bad England have been in the ODI format for so long, particularly in batting. Even during the period where England were consistently pretty successful in ODIs (1975-1992, say), they didn't produce a single batsman worthy of serious consideration among the all-time greats.

However, I do notice that the vast majority of those players listed are from the 1990s onwards. From the 1980s we only really have Richards, Crowe and Miandad (even if a few others were considered). I wonder quite what the x-factor was that changed the game so completely and led to what is presumably a dramatic increase in scoring rates?
[[ 1990s: Greatbatch/Jayasuriya/Kalu amd their fearless attacking methods. Also many rule changes which allowed the batsmen to flourish. Ananth: ]]

Posted by red_forever on (March 17, 2013, 18:09 GMT)

Hello Ananth,

Good to the blog up and running. Another fine article from your stable.

Since you would be doing a follow up article, can you consider dot ball percentage of these players, as in ODI's dot balls are verycritical and their importance varies from position to position, as an opener has the liberty to play a hign number of dot balls and still make up for that by staying long, but the same liberty cant be extended to lower order guys, say 5/6/7.

Some recent examples are say India vs RSA in 2011 world cup. We had an outstanding partnerships for the 1st two wickets, but once Sachin fell around 40th over, then the people coming next went after the bowling without trying to get their eye in as they dint have the liberty of using up balls, and i think this is what sets apart Dhoni from others, that he has a high percentage of dot balls and but still makes up for it by the end.
[[ The absence of ball-by-ball data prevents me from doing this analysis. Even now, with complete data available, dot balls do not find a place in scorecards. Ananth: ]]

Posted by   on (March 17, 2013, 18:03 GMT)

Excellent analysis, eagerly waiting for the second part -though I am not sure one article could capture all the points you wish to cover-.
[[ I an always do a third article some time in the future. Ananth: ]]

Posted by   on (March 17, 2013, 17:32 GMT)

I would have Haynes, Ganguly and Anwar in my top 15 in the place of Crowe, Flower and Pietersen. Trescothick and Astle were in my opinion better than some of the players to miss out, I wish they had been mentioned.

Posted by   on (March 17, 2013, 17:26 GMT)

@ananth Not exactly top 15...maybe the bottom 15th.:) And if it were me, i would put symonds in the top 15.

Posted by   on (March 17, 2013, 17:06 GMT)

@Ananth,I think Ganguly should have been there in place of Mark Waugh.. Wait have you thaught of giving equal representation to all countries..! could be possible..otherwise ganguly should feature in.

Posted by its.rachit on (March 17, 2013, 16:56 GMT)

hi ananth .. nice analysis ... but since you have stuck to selecting the best from each team and the next best from some teams (dhoni, gilly, lara, inzi), you had to leave out some definite giants of the same .. surely greenidge, anwar, ganguly, de silva, dean jones (i somehow rate him as one of the very best) are better ODI bastmen then flower, crowe, pieterson (a tad over rated in ODI, averaging only 41, when every contemporary great is averaging close to 45) .. so i dont know what conclusion to draw from this list .... i cant say that these are the 15 greatest ODI batsmen ... so whats the inference ...

Posted by   on (March 17, 2013, 16:40 GMT)

honestly kallis and andy flower did nothing significant in odi cricket( they are amongst the all time greats in test cricket). i remember kallis failing to score boundries towards the end of the innings and sauth africa suffering because of it. andy flower was out batted by grant for a good part of his career and he won more maches. arvinda de silva should be in the list if inzamam is, arvinda was amongst the best batsman in the mid 90s and single-handedly won the 1996 wc semi-final and final while inzaman was never the top guy in odi cricket. also pieterson cannot be picked ahead of symonds, hussey or even yuvraj. he had a great start but his performance dropped singnificantly after the first 18 months.

Posted by   on (March 17, 2013, 16:32 GMT)

Ananth - Great article, awesome research. Full credit to you for that. And I do agree with your assessments. I would recommend adding on some measure of contribution in winning the game (MOM/Runs scored of total team runs/WinLoss ration etc). One thing that most user point out though (and nothing against your article) is we would never know how great these greats really were when you cannot measure that during Viv Richards era, we never had fielding limitations, no bouncer limitations, when wides/no-balls were more lenient. We would never be able to factor in the fact that Sachin probably is the only guy on this list who has faced all the bowling greats in the history of the game (Murali, McGrath, Warne, Arkam, Waqar, Walsh) with the exception of Kumble.
[[ Part 2 will cover all what you ask for and more. Ananth: ]]

Posted by   on (March 17, 2013, 16:20 GMT)

I am not quite sure about this one, but can you prepare a table showing number of innings/ runs scored by batsman when during his stay on crease three or more wicket fell. It can show how much responsibility a batsman like sachin in nineties or martin had to take.
[[ This is probably the 73rd time I am making this statement. But you are probably new to my blogs. Hence I do not mind. The data on when the batsman got out is availablke only during the last 15 years or so. Hence the duration of the dismissed player's innings is uncertain. Ananth: ]]

Posted by   on (March 17, 2013, 15:05 GMT)

Opening batsmen in ODI's, once they're good enough, always have a certain advantage. More balls, lses pressure, time to get set etc. A lot of cricketers become overrated because they're given the opening position in ODIs. The 3, 4 and 5 positions in ODIs are a thankless position, whilst the opening spots allow you to pad your averages and stats. Much of Tendulkar's good ODI stats, for example, are due to him being an opener.
[[ I will add the 3/4 analysis as requested by a few readers. It is possible that the opening batting position carries some benefits. But there is no denying that the team would benefit by having their best batsmen bat in the positions 1-3. Ananth: ]]

Posted by   on (March 17, 2013, 14:29 GMT)

these stats are bullshit, there are a thousands error in this report. like- richards played 187 matches not 167,miandad played 233,mark played 244.
[[ The only mistake is that the title in Table 1 should read "Inns" and not "Odis". I will get it corrected. However I suggest you moderate your language which betrays a lack of everything: courtesy, cricket knowledge, common sense and correct behaviour. If you had looked at the tables for a minute, you would have known that what I referred to was Inns. This is the last time you will use such language. Ananth: ]]

Posted by   on (March 17, 2013, 14:26 GMT)

Andrew Symonds, Scott Styris and Paul Collingwood could've been included.These trio may not have the numbers to back their abilities, but most of their knocks in which they scored 50+ have been match turning. Also all of them were pretty decent allrounders.
[[ In the top 15 ??? Ananth: ]]

Posted by   on (March 17, 2013, 14:12 GMT)

Nice article,Ananth,thanks! My freind David told me about your articles,so I decided to see,and I wasnt dissapointed! One request-can you please make an article about the best players from the non test nations?I very much would like to see how obrien stirling and hamid hassan compare to the zimbabwe and SL players berore they got test status/// Thanks again!
[[ I think this is a rare request. I will try and do it sometime. However my database is limited to the international matches involving teams. Hence some of the matches involving associates which are not considered internationals will be excluded. One request. Your Facebook name is in Hebrew, if I am not wrong. I would not even know you. You could at least put in your name in the comment. Ananth: ]]

Posted by   on (March 17, 2013, 13:33 GMT)

your analysis is quite good and shows why the great batsmans are called great.

But you considered rpi for your analysis. And i think that hampered the figures of plyers batting in lower middle order like dhoni who some times come to bat after 45th over or even in 49th as they bat so low. so shouldn't you omit the not out innings where player played few balls(say 10 15 ) or few runs which ever data is available to you
[[ I have written enough on this. Pl also refer to the earlier responses. The fact that Dhoni's RpI may have come down because of non-consideration of his not outs has not prevented him from being included in this elite list of "Odi Giants" !!! Ananth: ]]

Posted by   on (March 17, 2013, 13:30 GMT)

What about Sourav Ganguly?

Posted by   on (March 17, 2013, 13:26 GMT)

i think ab devillers kohli and amla all should have been consider before peiterson other then that the list you cant argue and i really believe dhoni is giant of ODI one of the top 5 odi batsman ever

Posted by   on (March 17, 2013, 13:07 GMT)

@Chetan Asher. I completely agree that fielding should be seriously taken into account when considering great ODI players, although this ananthalyis is purely looking looking at batsmanship. Richards and Ponting would be the standouts for me - M Waugh brilliant as well. Not many bad ones< a few average ones. "ODI BATTING GIANTS" IS THE TITLE. SO WHY BRING IN FIELDING. ANANTH Sent from mobile phone.

Posted by   on (March 17, 2013, 12:49 GMT)

RPI and index calculated based on it is inferior to Avg and index calculated on it. It's true that averages favor those batting down the order. But the fact is that those batting at the top of the order are able to take complete advantage of getting set. The lower half guys get set and the innings truncates at 30-35 not out or so. Also, when they get more overs to bat (say 35-40 overs) they are walking into a crisis. Their strike rates in such innings are bound to be lower. Those extra not outs are fair compensation.
[[ This is a matter of opinion. You could refer to my earlier article. What you say makes sense when the not outs are within acceptable range, say, 15%. But these favour the middle order batsmen in an out of proportionate manner when the not out % is high, say, Dhoni 29%, Bevan 34%, Hussey 28%. Ananth: ]]
Your cut-off criteria (was the list limited to people with 100 plus appearances) missed out a giant called Zaheer Abbas - avg 47.6, strike rate 85. He outperforms most in your list.
[[ Too few matches. Then what about Amla. His numbers are superior to anyone else. 3229 at 58.71/52.93/50.09. Ananth: ]]

Posted by   on (March 17, 2013, 12:46 GMT)

Tendulkar, Gilchrist, Richards, Ponting, Lara, Bevan, (Dhoni) will do it for me - Sobers and Bradman show potential but have yet to impress...
[[ Nightmare team. The bowlers would go off to play IPL/BBL/BPL/CPL and whatever 'L's are there. Ananth: ]]

Posted by Charith99 on (March 17, 2013, 12:46 GMT)

nice work Ananth can't wait to see the next part. Regarding sanaths worst streak i think it should be noted that unlike many batsmen who you have chosen sanath played more as a bowler who could bat a bit during 1990. Infact he held the record for the best bowling figures for SL a long time. (The sad thing is he is loosing all the respect and admiration he gained during his playing days by getting himself tangled in politics)

i agree with sam paul that batsmen like ganguly and aravinda deserves to be in the first set of selections ahead of someone like KP ( i'm not saying KP is bad but has he played more great one day knocks and won the match for his country than de silva or ganguly have done)
[[ Team balance !!! Ananth: ]]

Posted by liamhsiemllac on (March 17, 2013, 12:43 GMT)

Thanks for the response... Ha, I hope so - I'm a massive Bell fan, I do have a soft spot for the stylists, despite the frequent frustrations. As for the left arm spin issue with KP- I never really bought into it when it was first suggested he had a problem - but of course, every team started bowling an awful lot of left arm spin to him - which in turn inflated his percentage of that particular mode of dismissal!

That said, I'm quite excited by the English ODI side. Not exactly the guns blazing model - but wily bowling and a really solid top order.

Look forward to part two, as always.

Posted by   on (March 17, 2013, 12:30 GMT)

Welcome back in your full glory Ananth san. Great article to start with..."accessible and free of trash". As usual there`s so much to think about - so many different aspects to the game, and so many changes over the past 40 years; bowling restrictions, fielding restrictions, smaller grounds, very different pitches - and all of them have favoured the batsmen. Very tough to compare them. I appreciate why some people may feel that Ganguly and Hussey have missed out a bit, and my 15 may have been slightly different, but I think there`s a pretty good balance of the best batsmen over the years here. re. Calvin`s featured comment, Ananth has written an article on ODI SR/Ave across the decades. That Richards has the 3rd highest ave. and 3rd highest SR of these 15 is, as you suggest, simply phenomenal. He`s quite clearly the best one-day batsman I`ve seen, and a comparison with some brilliant contemporaries shows his clear supremity. Richards at 3, Gilchrist and Sachin opening.
[[ Dave I have missed all you guys. Today I will sleep well. Ananth: ]]

Posted by   on (March 17, 2013, 11:56 GMT)

It will be interesting if you could include an analysis on how the shortlisted batsmen fared against pace & spin. I expect the opening batsmen to have better figures against pace vis-a-vis the middle order batsmen in the list & vice versa in the case of spin.
[[ Ball-by-ball data not available for most matches. Ananth: ]]

Posted by   on (March 17, 2013, 11:53 GMT)

At the risk of looking / sounding biased, how do the below players score / where would they stand in your analysis (Since your analysis & ranking is based on statistics, I am sure they are not on this list for good reason, but would like to know anyway):

1) Yuvaraj Singh, India. 2) Arjuna Ranatunga, Sri Lanka. 3) Allan Lamb, England. 4) Herschelle Gibbs, SA

Another request on the side - I have seen a lot of insights into the impact of batsmen / bowlers on cricket matches - what about fielders ?

I know Jonty Rhodes used to make it almost impossible for batsmen to score past point / Solkar past Square-Leg, resulting in batsmen trying to manufacture strokes & lose wickets ..bowlers got the MOM's in most cases, but theirs was the contribution (Donald / Pollock / Bedi / Prassana / Chandra will vouch for this). ...Can Cricinfo come up with something for such stalwarts whose efforts might have won matches for their teams, MOM's for bowlers ...but never got them their due recognition ?

Posted by liamhsiemllac on (March 17, 2013, 11:41 GMT)

As an Englander, the reintroduction of Mr Pieterson to our side should prove a little problematic since Bell has performed so well since replacing him at the top of the order. I'm not sure if I'd want to break the Cook/Bell partnership, but Mr Pieterson's numbers are mighty impressive, albeit from a small sample. That said, his lower order numbers are also pretty special. I just don't know if I would be tempted to use him later in the innings, have him opening again, or just settle in at number four? Thoughts?
[[ Apologize for late introduction of response. The comment was published by someone else despite instructions. Bell's numbers are quite good and I would say he is a little bit behind Pietersen now. Another couple of good years could see him becoming the best Englaish ODI batsman. I think Pietersen would be vary valuable at no.4 since he plays the spinners quite well. In ODI's he should not really bother about the average left-arm tweakers. Attack them and they would disappear. Ananth: ]]
Oh if I had your skills I'd crunch the respective numbers myself - super article btw.

Posted by   on (March 17, 2013, 11:33 GMT)

Since you have included World Cup and Champions Trophy, would it not be fair to the earlier generations to also add events like the world championship where all/most of the test playing nations participated?
[[ I am very particular that I would only include only events in which all 8 Test-playing countries are included, that too an official World event. Nehru Cup, Hero Cup, B&H event all fall just short. Ananth: ]]
Another additional check for the non-openers could be the situation in which they started their innings, and when it ended. For example -Scenario 1 Coming in at 50/2 after 12 overs - Rpi - 25, RR - 4.16; Multiply Rpi and RR = 104. Ending at 200/5 after 40 - Rpi - 40 at RR of 5; Multiply RR and Rpi - 200. Percentage Increase = 200/104 = 100% (approx). It is better than Coming in at 150/2 after 30 overs ; Multiply RR and Rpi = 375, and finishing at 350/5 after 50; Multiply RR and Rpi = 490. Percentage increase 490/350 = 40%
[[ Unfortunately, Sitanshu, when I say i do not have the batsman coming-in information, I also do not have the ball information, when the wicket fell. Ananth: ]]

Posted by Mundrik on (March 17, 2013, 11:31 GMT)

Great Job... I am sure you have worked your ____ off to get this on board.
[[ Worth it. Ananth: ]]

Posted by   on (March 17, 2013, 11:07 GMT)

Thanks a lot Anantha for your reply..

Just one more point - The stats of Viv (oh yes, I'm a huge fan!!), the balls faced & strike rate, are slightly different than what is there in Cricinfo site ( 7451 balls in cricinfo, 7581 in your analysis).. Was there some World Series matches or just a printing mistake??
[[ It is true that my data has been created using the wonderful scorecards of Cricinfo. However this has been done over the past 15 years and Cricinfo constantly corrects its scorecards. It is impossible for me to keep track of these changes. There are likely to be minor variations. Having said that I would specifically look at this and correct, if possible. Ananth: ]]
And, thanks for such a wonderful article.. I first doubted the inclusion of Flower & Inzi, but then realized that Flower having relatively lower stats than few others also is impacted by the team he played in, & his peer index would seal his place in the above list!! Thanks..
[[ In any such analysis balance is essential. Between teams, between eras and between batting responsibilities. Ananth: ]]

Posted by rawr94 on (March 17, 2013, 10:22 GMT)

So who would be your all time ODI team, based off your findings?

Posted by   on (March 17, 2013, 10:09 GMT)

Hello Sir, Have you given any weightage or deducted any factor for performance against weak teams?? No offence, but Zimbabwe and Bangladesh or the Canada, Namibia etc? That will have some influence on the numbers , definitely. Also I thought of home/away numbers, but that would have a larger influence in Tests than here .WHen you do for tests, please include that(since Sanga,Jaya,Sehwag etc would be filtered out!!!)

Posted by   on (March 17, 2013, 10:05 GMT)

Hi Ananth, I would like to ask why you ignored AB de Villiers and chose Pietersen who has played fewer matches and has aggregated less impressive stats than the former. Even a Hashim Amla for that matter, despite his relatively shorter career would have been a worthy inclusion. Pietersen may be a good player but in my opinion, the two South Africans are better.
[[ 2 from South Africa and none from England. Does not seem right. Ananth: ]]

Posted by JohnnyRook on (March 17, 2013, 9:50 GMT)

Hi Ananth, Here are my two cents

1) I feel longevity should be given importance in Batting index. It is extremely difficult to maintain the high level of performance for a long period of time. Within this list, Tendulkar has played 3.8 times that much of Kevin Peterson and scored 4.2 times runs. Even when Kev retires, this factor will be at least 2. This shouldn't be ignored in calculations. If we assume 25 to be generally accepted good BPI and 200 to be a decent number of matches a good player plays, think how much India has benefitted because of Sachin's longevity.

2) Everything in the world is relative. Richards averaged 47 with SR of 88 when 35 and 70 respectively were considered great figures. So this should really be considered. I remember you had written an article earlier in similar lines. Can some calculations from that be used. Team share of runs/balls in next article will help to some extent but I feel the batting position influences it just too much.
[[ The Index is a performance-based measure. There is no point in bringimg in longevity. The fact that a batsman sustained a leval of performance twice as long as another does not make him better. It only underlines his terrific value to the team over twice the time.
If I were doing a piece to determine the best ODI batsman of all time, I would include a weight for longevity. But this is not such an exercise. Ananth: ]]

Posted by   on (March 17, 2013, 9:13 GMT)

Great analysis Ananth. Some of the figures make me kind of rue the fall of 50-over format due to the rise of t20. The various stages in an ODI took measure of skills and temparement of players and only then such figures deserve respect. Although, your 'Index' measure will help in analysing t20 stats as well.
[[ DB, the Part 2 analysis will throw lot more insights. I wanted to set the tone with some easy-to-understand numbers. Ananth: ]]
@ true that Bonehead_maz- those players could have been good finishers as well. But tendulkar didnt have such good figures coming at no.4 or 5. Still, its flabbergasting to see tendulkar averaging 48 as an opener for such a long period.

Posted by   on (March 17, 2013, 8:52 GMT)

I know this will be the first of many...but Ganguly-the ODI batsman- would feel hard done by.. Andy Flower is a top notch test batsman but definitely Ganguly has better one day records...just to make it even countrywise we shouldnt lose sight of the bigger picture...this coming from a not-really-a-big-ganguly fan
[[ I was as much upset to keep out Ganguly as I was, keeping Hussey out. But I could not help but be influenced by Dhoni's late order batting exploits. And he has won many key matches, batting at no.6/7. Ananth: ]]

Posted by SankerV on (March 17, 2013, 8:46 GMT)

Excellent post and nice to see the stats also! But I found Ganguly's name missing - the fastest to score 7000, 8000 & 9000 runs, also the 2nd fastest to reach 6000 & 10000 runs in ODI. At the same time I happy to see Dhoni's name too in the list:

sankartypo3.wordpress.com/2013/02/24/mahendra-singh-dhoni-the-captain-of-indian-cricket-team/
[[ Pl see response to earlier comment. Ananth: ]]

Posted by Bonehead_maz on (March 17, 2013, 8:20 GMT)

Very nice work Ananth. I'm not a ODI fan, but glad to see a few on that list of unlucky to miss out. (will never forget Aravinda in that WC semi final .... wow).

Am amazed to see so many having had a few bats at No.6 or below. Sure looks like Tendulkar, Lara and Miandad would have made outstanding "finishers" if they'd had to that more often.

Murray
[[ Murray I am also surprised to see so few crossing the minimum mark. Maybe all teams come around to the view that the best batsmen should occupy the top 4/5 slots. Ananth: ]]

Comments have now been closed for this article

ABOUT THE AUTHOR

Anantha Narayanan
Anantha spent the first half of his four-decade working career with corporates like IBM, Shaw Wallace, NCR, Sime Darby and the Spinneys group in IT-related positions. In the second half, he has worked on cricket simulation, ratings, data mining, analysis and writing, amongst other things. He was the creator of the Wisden 100 lists, released in 2001. He has written for ESPNcricinfo and CastrolCricket, and worked extensively with Maruti Motors, Idea Cellular and Castrol on their performance ratings-related systems. He is an armchair connoisseur of most sports. His other passion is tennis, and he thinks Roger Federer is the greatest sportsman to have walked on earth.

All articles by this writer