April 22, 2012

Test Bowling: a peer analysis of spells

A statistical study analysing Test bowlers' performances relative to those of their peers
125

Muttiah Muralitharan had a career spell average twice as good as his peers
Muttiah Muralitharan had a career spell average twice as good as his peers © AFP

When I was perusing the scorecard of the South Africa - New Zealand match which finished in a draw, I was admiring Morne Morkel's bowling performance: 6 for 23. Mentally I compared that with de Lange's spell of 0 for 77 and computed in my mind that it was "19 times better". Then it struck me that this was when compared to a single bowler. What happens if we compared to all the other bowlers. The number came to around 44. I suddenly remembered that I had done this analysis for batsmen more than a year back, based on a Unnikrishnan suggestion but had not done it for the bowlers. And I was curious to know where Morne Morkel's performance stood, over 2000 Tests.

To view the Batting Innings Peer Index article please click here.

The greatness of this analysis is that it is the purest of peer analyses possible. All conditions remain the same. Against the same set of batsmen, in almost the same conditions, identical match situation, ball conditions (somewhat) similar, weather similar, same set of umpires and so on.

Once the spark comes, the system takes over. Soon I realized that this was totally different to the Batting analysis. The differences are outlined below.

1. There is no limit to the batsman runs nor the team runs. However the total wickets cannot exceed 10. Hence there is a cap on the combined number of wickets.

2. In completed innings, the highest share of a batsman is Bannerman's 67.3. Two bowlers have captured 100% of the team wickets, 14 bowlers 90% of the team wickets, 72 bowlers 80% of the team wickets and 246 bowlers 70% of the team wickets. There is a totally different dynamics in operation here.

3. It is certain that if a batsman scored x runs, the other-batsmen would have scored y runs, whatever be the situation, if basic precautions are taken. However there are many instances in which a bowler captures x wickets and the other bowlers capture no wicket. Morkel's is the perfect example. So this has to be taken into account.

4. There are two sub-analyses possible in the bowler analysis, unlike the batting analysis. I could do a peer comparison within the innings of the bowling accuracy and bowling strike rate. These are likely to produce totally different sets of performances.

How do I take care of all the above situations. First the terminology. The ratios are called Spell Peer Factor - 1/2/3 (SPF-1/2/3).

a. As far as I am concerned there is no 0 wkt situation. If the other-bowlers have not captured a single wicket, I take that notionally as 1 so that a division is possible. I am anyhow a very practical analyst. If a batsman started his career with an unbeaten 75, as far as I am concerned, his career figures should read 1-1-75-75.00 and not as "infinity" as some misguided purists would suggest. I have been irritated by the oft-repeated phrase "no average". This method would work very well in all situations, including the two 10-wicket performances and the trigger for this analysis, Morkel's spell.

b. For the Spell Peer Factor analysis of Bowling average and strike rate, I will only consider spells of 4-wickets or more. A 4-wicket capture is a very significant bowling spell and will add weight to the results. I had initially considered 3-wickets but decided to raise the bar. I have given a list of a few significant 3-wicket performances at the end.

c. For the Spell Peer Factor analysis of Bowling accuracy, I will only consider spells of 120-balls or more. This makes eminent sense. Otherwise a bowler with a single maiden over will throw everything out of gear. Let the accurate bowlers earn their spots over a decent 150-minute spell.

d. For the Career Peer Factor determination, I would exclude spells which are wicket-less and lower than 30 balls. This will ensure inclusion of bowling spells like these: Benaud 3.4-3-0-3, Kumble 2-1-2-2, Lawson 1-0-2-1 et al. To those who say 10 overs, I can only say, 30 balls present a fair chance of a wicket. Anyhow do not waste too much time. An example: Muralitharan has only two such short fruitless spells. So the impact is minimal.

e. Since this is a peer analysis of a bowler's performance against the combined performance of his team-mates, I have decided that the rest of the wickets will include all dismissals. What matters is that the rest of the team effected these dismissals. That is all. If I had excluded run outs etc., then the ratios would be higher across the board.

The formula for determining the SPF values is quite simple and outlined below.

Bowling average for other bowlers for innings
SPF-1  = ---------------------------------------------
Bowling average for bowler for spell

where

Bowling average Innings score - Extras - Runs conceded by bowler for other bowlers = ------------------------------------------------ for innings Innings wkts - Bowler wkts (If 0, taken as 1)

For SPF-2, use RpO (Runs per Over) instead of Bowling Average and For SPF-3, use BpW (Balls per wkt) instead of Bowling Average

Let me anticipate some comments. It really does not matter if the batsmen dismissed by the bowler are lower in the order and the other dismissals are top order. The bowler might have got the new ball and the other might not have. On the other hand the bowler might have got an old ball with reverse swing and the others might not have. The bowler might have bowled more to Richards. And so on. Let us forget all these factors. What we are looking at is a simple peer comparison within an innings, that is all. Instead of finding faults, let us draw some insights. I will be using this measure in my bowling spells rating work.

Readers have a knack of converting every analysis into a best bowler/batsman type of article. Please do not send comments such as Steyn is better than Morkel, Wasim is better than Sarfraz, Ambrose/Marshall are better than Lawson and so on. Possibly true, but not relevant here. Please read and understand the article which can be defined as "peer comparison within the same innings".

Now for the tables. There are four tables. Three cover the individual Spell Peer Factor values, one each for the Bowling average, Runs per over and Bowling strike rate. The last table highlights the Career Peer Factor values. As always, the top 25 or so entries are shown here and the full table, now standardized in the form of Excel sheets is available for downloading.

SPF-1: Bowling Average based

MtIdYearTeamVsT-WktsT-RunsT-AvgeBowlerWktsRunsAvgeSPF-1
 
20372012SafNzl0171.0171.0Morkel M6233.8344.61
481896EngSaf1116.0116.0Lohmann G.A9283.1137.29
16302002WinBng473.018.2Lawson J.J.C630.5036.50
17202004AusInd4192.048.0Clarke M.J691.5032.00
15832002SlkZim1161.0161.0Muralitharan M9515.6728.41
2901947AusInd554.010.8Toshack E.R.H520.4027.00
4281956EngAus0136.0136.0Laker J.C10535.3025.66
6831971WinInd1229.0229.0Noreiga J.M99510.5621.69
11101988AusWin1230.0230.0Hughes M.G88710.8821.15
9841984EngPak1170.0170.0Cowans N.G5428.4020.24
8491979PakAus1189.0189.0Sarfraz Nawaz9869.5619.78
10291985NzlAus1111.0111.0Hadlee R.J9525.7819.21
7991977WinPak2136.068.0Croft C.E.H8293.6218.76
12221993PakWin0106.0106.0Waqar Younis4235.7518.43
11651991SlkNzl1191.0191.0Labrooy G.F44210.5018.19
161884EngAus6515.085.8Hon.A Lyttelton4194.7518.07
15242000NzlZim1255.0255.0Martin C.S57114.2017.96
4831959IndAus1135.0135.0Patel J.M9697.6717.61
10201985AusEng1301.0301.0McDermott C.J814117.6217.08
12661994EngSaf1104.0104.0Malcolm D.E9576.3316.42
12171993AusNzl6194.032.3Warne S.K482.0016.17
10811987PakEng1100.0100.0Abdul Qadir9566.2216.07
10481986EngInd1122.0122.0Edmonds P.H4317.7515.74
8511979IndEng0427.0427.0Kapil Dev N514629.2014.62
3481952IndEng2200.0100.0Mankad M.H8556.8814.55
5271962WinInd2135.067.5Gibbs L.R8384.7514.21
15132000PakEng0290.0290.0Saqlain Mushtaq816420.5014.15
14431999IndPak0104.0104.0Kumble A10747.4014.05

Where does the trigger spell of Morne Morkel stand. Lo and behold! It is at the top. I had a sneaking suspicion that 44 was not a number which could be beaten. Imagine a bowler performing nearly 50 times better than the rest of his team-mates including Philander (avge ~ 14) and Steyn (avge ~ 23). Then, as expected, comes Lohmann, not surprising since he specialized in spells of 8/9 for nothing. Afterwards comes a spell of recent vintage. Lawson picking up the last six Bangladeshi wickets for just 3 runs and an SPF-1 value of 36.50. However the other bowlers also fared well, capturing 4 for 73. Then comes the very well-known Michael Clarke special, for an SPF-1 value of 32.0, 6 for 9 to move India from 181 for 4 to 205 all out. This match was incidentally played on a muddy lane around Wankhede Stadium. The top-5 are rounded off with a non-freakish spell by Muralitharan: the first nine Zimbabwe wickets for 51 against the 10th wicket by his colleagues for 161, an SPF-1 value of 28.41. Laker's 10-wicket spell finds a place in the top-10 with an SPF-1 value of 25.66. Hadlee's master class against Australia is there with an SPF-1 value of 19.21. Kumble's 10-wicket haul has a SPF-1 value of 14.05.

SPF-2: Bowling RpO based

MtIdYearTeamVsT-OversT-RunsT-RpOBowlerOversRunsRpOSPF-2
 
5511964IndEng1582991.9Nadkarni R.G3250.1612.06
10461986IndEng761612.1Maninder Singh2090.444.86
2851947SafEng931942.1Mann N.B.F20100.504.16
3061948EngSaf972022.1Compton D.C.S21110.524.00
5051961IndPak912482.7Nadkarni R.G34240.713.83
8181978NzlEng1703712.2Congdon B.E24140.583.74
4491958AusSaf1201791.5Burke J.W2080.403.73
10341986IndAus44942.1Yadav N.S33190.583.71
5911965EngNzl37922.5Titmus F.J26180.693.59
16752003SlkEng612123.5Muralitharan M40401.003.48
4381957SafEng561101.9Goddard T.L21120.563.46
5931965EngNzl581452.5Titmus F.J26190.733.42
5051961IndPak752002.7Nadkarni R.G52430.823.27
3221950AusSaf441313.0Noblet G22210.923.24
1931930EngWin912382.6Rhodes W20170.823.21
4331956IndAus471603.4Umrigar P.R20211.053.20
8281978EngNzl711331.9Edmonds P.H34200.593.20
971908AusEng1013363.3Armstrong W.W34361.053.17
2501936AusSaf421283.0O'Reilly W.J21200.953.15
7071973AusPak40872.2Walker M.H.N21150.703.09
4181955IndNzl38932.4Manjrekar V.L20160.803.06
18882008NzlBng1022182.1Oram J.D.P20140.703.05
6301968IndNzl1213122.6Nadkarni R.G36310.853.04
15032000WinEng471453.0Ambrose C.E.L22221.003.03
5631964EngAus30842.8Titmus F.J27250.933.02
4661959AusEng742142.9Rorke G.F24230.953.01

This table is based on the RpO value of the spell and the RpO value of the other bowlers. The spell has to be a minimum of 120 balls to be considered. At the top, by a few kilometres, is Nadkarni's famous spell. 32 overs of rather innocuous on-a-rupee-coin bowling, resulting in 5 singles. The other 187 balls were mostly padded away. Based on today's rules, Nadkarni would have got 80 wickets, and another 20 on DRS. Barrington and Bolus: I still get nightmares since I heard the whole innings on radio. The ratios here are smaller but Nadkarni leads with an SPF-2 value of 12.06. The second effort, also by an Indian left-arm spinner, cheese to chalk of Nadkarni, yielded an SPF-2 ratio of 4.86, just over a third of Nadkarni's figure. At least Maninder Singh was more attacking and picked up 3 wickets. It is possible that a two might have been scored off him. Nigel Mann and Compton (surprise !!!) take the next two places. Compton bowled 16 8-ball overs for 11 runs. Then Nadkarni appears again: this time for a spell of 34-24-24-1. It would be of interest to note that Nadkarni's spell in the other innings was an attacking 52.4-38-43-4. What does one say of a match performance of 86.4-62-67-5. One just gives up making sense. Muralitharan, Oram and Ambrose appear in this collection of ancients.

SPF-3: Bowling Strike rate based

MtIdYearTeamVsT-WktsT-BallsT-BpWBowlerWktsBallsBpWSPF-3
 
10201985AusEng1630630.0McDermott C.J821627.023.33
20372012SafNzl0384384.0Morkel M610016.723.04
10291985NzlAus1318318.0Hadlee R.J914215.820.15
481896EngSaf1160160.0Lohmann G.A9728.020.00
6831971WinInd1648648.0Noreiga J.M929833.119.57
4281956EngAus0594594.0Laker J.C1030830.819.29
12661994EngSaf1204204.0Malcolm D.E99911.018.55
19982011WinInd1474474.0Edwards F.H513827.617.17
11651991SlkNzl1480480.0Labrooy G.F411428.516.84
161884EngAus61196199.3Hon.A Lyttelton44812.016.61
15242000NzlZim1636636.0Martin C.S519739.416.14
9841984EngPak1267267.0Cowans N.G58416.815.89
17202004AusInd437293.0Clarke M.J6386.314.68
8491979PakAus1456456.0Sarfraz Nawaz928431.614.45
15262000SlkSaf1144144.0Muralitharan M66010.014.40
13951998ZimSlk1532532.0Streak H.H415037.514.19
13981998EngWin2342171.0Fraser A.R.C89712.114.10
12221993PakWin0228228.0Waqar Younis46616.513.82
13231996ZimNzl1618618.0Streak H.H418045.013.73
7631975AusEng61338223.0Walters K.D46516.213.72
1381921AusEng1420420.0Mailey A.A928231.313.40
3941955IndPak548296.4Gupte S.P5367.213.39
4611958IndWin1306306.0Gupte S.P920723.013.30
15132000PakEng0732732.0Saqlain Mushtaq844455.513.19
11101988AusWin1366366.0Hughes M.G822227.813.19
4281956EngAus1144144.0Laker J.C910011.112.96
14431999IndPak0204204.0Kumble A1015915.912.83
15832002SlkZim1342342.0Muralitharan M924026.712.83
11891992EngPak0624624.0DeFreitas P.A.J419849.512.61
14231998SlkEng1450450.0Muralitharan M932636.212.42

This is the peer comparison of the bowler strike rates. McDermott is the unlikely bowler at the top. He was in the news recently because of the way he has re-vitalized Australian bowling. His 8 for 141 was off 36 overs out of a huge score of 482 in 142 overs. McDermott's SPF-3 value is a huge 23.33. I like this since this is not one of those freak 5 for 2 type spells. Morkel's recent spell is quite close in the second place. His strike rate was 16.7 compared to the team strike rate of 384 leading to an SPF-3 value of 23.04. Two earlier acquaintances, Hadlee and Lohmann appear next. Then follows a a nine-wicket spell of Noreiga in the historic series-winning match, also happened to be Gavaskar's debut Test. Then comes Laker's ten-wicket spell. Kumble's 10-wicket spell comes at the end of the section with an SPF-3 value of 12.83.

Bowlers' Career Analysis: Based on SPF-1: Bowling average based

BowlerTeamDebutTestsWktsAvgeSpellsTotPtsC P F
 
Muralitharan MSlk199213380022.73228460.32.02
Barnes S.FEng19012718916.434998.62.01
Hadlee R.JNzl19738643122.30145271.81.87
Imran KhanPak19718836222.81139240.41.73
Laker J.CEng19484619321.2581139.51.72
Fraser A.R.CEng19894617727.3275126.11.68
Morkel MSaf20063913930.0473121.11.66
Streak H.HZim19936521628.12100165.31.65
Taylor B.RNzl19653011126.605283.71.61
Tayfield H.JSaf19493717025.916095.21.59
Croft C.E.HWin19772712523.305282.11.58
Gupte S.PInd19523614929.555992.91.57
Shoaib AkhtarPak19974617825.6982127.21.55
Pollock P.MSaf19612811624.195179.01.55
Davidson A.KAus19534418620.5380123.01.54
Collymore C.DWin1999309332.305178.51.54
McKenzie G.DAus19616024629.79109165.71.52
Grimmett C.VAus19253721624.226699.41.51
Underwood D.LEng19668629725.84145217.31.50
Mankad M.HInd19464416232.3267100.71.50
Donald A.ASaf19927233022.25126187.61.49
Steyn D.WSaf20045427223.19100148.01.48
Dilley G.REng19794113829.766391.01.45
Edmonds P.HEng19755112534.1875108.61.45
Ambrose C.E.LWin19889840520.99174250.11.44

This is a career summation and averaging of of all qualifying SPF-1 values. The cut-off for this table is 50 such spells. As an honorary invitee I have included Barnes who had 50 spells but one was excluded since that was 4-1-18-0. If anyone deserved it, Barnes is the bowler. Muralitharan and Barnes are the only bowlers who exceeded 2.0 in the Career Peer Factor value. Think of the significance of this. It has always been said that Muralitharan got his wickets because he played in a weak team. Fair enough. But, gentlemen, this is a peer comparison of performance measures. Not just did he get more wickets but got those at half the average of his peer bowlers. So let us give the great men, Muralitharan in particular, the credit. Just for a comparison, Bradman's figure for this value was 3.348.

Hadlee is next with an excellent career ratio of 1.87. A bow to one of the indisputable all-time greats. Then comes Imran Khan. He does not lose out because of the Tests he did not bowl in. A well-deserved fourth place at 1.73. Then comes Laker, not surprising because he out-performed his colleagues by wide margins. Finally a surprise sixth position for Fraser and a surprise seventh position for Morkel. I am happy to see Streak from Zimbabwe in the top-10. Then another surprise, B.R.Taylor of New Zealand. Tayfield, the incomparable South African spinner rounds off the top-10.

It is wonderful to see the classic leg-spinner, Subash Gupte topping the Indian bowlers, and nice also that Vinoo Mankad is the next Indian bowler. Colin Croft and Alan Davidson are the leading West Indian and Australian bowlers.

BCG Chart of Bowler Spell analysis: RpO vs BpW

BCG chart of career RpO and BpW
© Anantha Narayanan

Finally I am back to my favourite BCG charts for plotting the two contrasting measures, Career RpO and BpW Peer Factor values, which are the two components of the Batting average. The BCG chart will give a good idea of the way the accuracy and strike rate have interacted for the top bowlers. Remember these are not absolute values but peer values within the same innings. The top right quadrant houses bowlers who had above average figures for both accuracy and strike rate. The bottom left quadrant houses bowlers who had below average figures for both accuracy and strike rate. The top left quadrant houses bowlers who were strike rate centric. Generally the pace bowlers. The bottom right quadrant houses bowlers who were accuracy centric. Generally the spinners. But not written in stone. There are different types of bowlers mixed in all quadrants.

The top right group is the elite one and has six bowlers. Barnes has justified his special inclusion and could be termed the leader in this quadrant. Muralitharan is right up there, very close to Barnes. Imran Khan is also well placed. Hadlee has an excellent strike rate index value and just about manages to be in the right of the RpO middle line. Ambrose does it the other way. Pretty good on the RpO front and just about manages to be above the middle BpW line. But the real surprise is Kumble. He is also comfortably placed in the elite quadrant. Let us not forget that this is a peer index.

The bottom left group houses bowlers who have not been that great when compared to their compatriots. There are four bowlers in this quadrant. Lee has only a RpO peer factor of around 0.82, indicating that he has been a philanthropist when it comes to runs. He has a reasonable strike rate index. Hoggard, Benaud and Anderson are the other three bowlers here, possibly indicating that these bowlers played with other good bowlers through their career.

In the other two quadrants, the notable bowlers are Steyn, whose BpW peer factor value is simply amazing. Close to him in this regard are Donald and McDermott. Gibbs is there at the best RpO peer factor level. Bedi and Vettori are also there. Underwood, Garner and Walsh are the closest to getting into the elite quadrant. McKenzie, Kapil Dev and Lillee are the bowlers who are pushing to the elite quadrant from the BpW side.

The x-axis and y-axis lines are drawn at the median value positions. If the number of entries are counted it can be seen that both the lines divide the total population into 21-20 splits. The non-centric positioning of the two lines is because these are the top bowlers and their Peer Factor values run in the following non-centric manner. For lesser bowlers, the median values are likely to be closer to 1.00.

Career RpO Peer Factor: 1.38 to 0.83 (Median 1.10)
Career BpW Peer Factor: 1.77 to 0.93 (Median 1.23)

To download/view the Excel sheet containing the three Spell Peer Factor Tables please click/right-click here.

To download/view the Excel sheet containing the Career Peer factor tables please click/right-click here.

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

Comments have now been closed for this article

  • A. Khan on May 5, 2012, 7:57 GMT

    If someone is really interested in knowing the s/r of batsmen from that era, open statsguru and use period of Gavaskar’s playing days. You can use filters like minimum 1000 runs and avg 30-35 (to avoid lot many names) and sort by strike rate. There are very few batters (Crowe is one) who had their s/r not as well as Gavaskar’s. Frankly, I too was surprised at the s/r of Gavaskar and Mudassar Nazar, I thought Gavaskar’s to be 42 and Mudassar, around 38. All said and done, if someone did that BQI/pitch etc. analysis for strike rates, would be in for a huge disappointment, as gavaskar’s strike rate drops to 40 outside the subcontinent. It’s all there in cricinfo database. But unfortunately ball by ball data is not available for all matches. @Ananth: Please bear with me for being off topic, again.

  • A. Khan on May 5, 2012, 7:46 GMT

    When I said that Gavaskar was a slow player, to counter this, people resorted to different tactics. One doubted on my smoking habits and other “knowledgeable” person brought the names of 3 modern players, who are known for their slow batting and Boycott (once dropped for slow batting)/ and “celebrated” Vishy, without even caring to know why he was celebrated? definitely not for fast batting and also some players from pre-1970 era. For a PERSON who cites old players left and right, it was surprising to see such ignorance. By this count Sobers (s/r 49) was a toddler compared to these Pontings/Laras/Smiths? Was he? If someone has the time to READ about so many things of the past, he should also READ about when the pitches were started getting covered and what were the consequences. Coming back to slow batting, I knew that Mudassar was slower batsman than Gavaskar but cricinfo shows only a difference of 2 in their s/r! cont..

  • A. Khan on May 5, 2012, 7:32 GMT

    Initially I thought will stay out of this. This is not in any way to undermine sunil, I consider his 96 to be one of the greatest innings that I have seen and possibly the best player of spin bowling among the players I witnessed. But I have seen what I have seen; unfortunately I followed (worldwide) each and every match that was going on starting from that Mumbai test of 1980 against England. Although I don't get to see many now for last few years, I still follow them on cricinfo or sometimes watch 'em live on internet. I am not a computer, but after watching/listening the matches, an image is developed in the mind as to how fast or slow a player was. Cricinfo doesn’t need to tell me about a Tavare or about a Kapil Dev, it mostly confirms it. While playing with Aussies, Wessles was an aggressive batsman but when he played for SA, became a really dour batsman; needless to say his captaincy was similar. I don’t have to learn this from cricinfo. cont..

  • west indian follower on May 4, 2012, 19:44 GMT

    hi ananth, great piece. Was wondering, would you consider S.F Barnes to be the greatest bowler and potentially the greatest test cricketer over Bradman from a weighted cricket statistical point of view. [[ 1. Barnes' performances are not comparable to Bradman. Batting Avge: Bradman 100.0, Pollock ( Next best) 61.0 RpT: Bradman 100.0, Weekes (Next best) 68.9 Bowling Avge: Avge: Lohmann 100.0 Barnes (Next best) 65.7 WpT: Barnes 100.0 Lohmann(Next best) 88.8. Murali (Third) 86.1. These numbers indicate how far ahead was Bradman to the others. Others were quite close to Barnes. Ananth: ]]Also, after warne and mcgrath, who would you consider to be the third best spinner of all time. [[ Question not clear. Ananth: ]] Finally, when picking a greatest side ever, would you look at how the players measured up against others in their era, or an overall point of view? I know it is not relevant to the article, but just curious for your opinion [[ Picking a greatest side is generally one's personal choice. I will pick my side irresepective of how the numbers measure up. Ananth: ]]

  • Gerry_the_Merry on May 4, 2012, 4:25 GMT

    Ananth, there has been some talk about strike rates (whenever Gavaskar is dragged into a debate, people try to trash him by referring to slow strike rate) in this forum.

    I looked through your past articles, but could not find an in-depth study on this topic in Test matches. In particular 1) who are the top 30% batsman sorted by strike rate 2) ordered by BQI groups, how did their strike rate behave 3) sorted by pitch conditions RPI/RSI etc. how did they perform 4) has this been a virtue or a vice from a single team's perspective (we know what crowds want, but they also want T20) - e.g. Atherton 185* in Johannesburg and Greatbatch 146* in Perth contrast with collapses in the great Calcutta and Adelaide tests against India by Aussies in less than a day, similarly India against Pak in Bangalore, following 267 by Younis Khan and Afridi onslaught 5) top batsmen who played extremely slowly (e.g. >150 balls and <30 runs) frequently...

    Gilchrist, Gayle, Ponting, Lara and Richards top picks. [[ As you know, only recently have I almost doubled my Balls played information segment of the Database. I now have nearly two-thirds of the Tests covered. So now it is possible to do analysis like what you have asked for. In fact I now even have a decent handle on the batsmen career strike rates by doing a pro-rata allocation of team balls for the non-available one third segment. That too I intend to even strengthen by using the available strike rate as a guidance factor to do to the pro-rate allocation. It is a miracle. But I have the actual strike rate for 5235 of Bradman's runs. Ananth: ]]

  • Vikram on May 4, 2012, 3:13 GMT

    @Boll: Actually Imran Khan should be considered the most important PLAYER of a test team. This analysis would have included his last few years when he really didn't bowl much and hence would have suffered in peer comparison. However he made it up through his batting, leadership and talent-finiding skills. His batting average is nearly 50% higher than Hadlee. In terms of BOWLER, Barnes/Murlai/Hadlee had the highest contribution to their respective teams over their entire career.

  • shrikanthk on May 4, 2012, 2:43 GMT

    @shrikanthk: I have seen enough comments from people where SRT has been ranked outside top 3 in even today's generation of Indian batsmen. Even in such selections, SMG would have to be in top 5.

    I wonder who these people are making such outrageous rankings. Let's think for ourselves. Is there any reasonable argument for excluding either SMG or SRT from the top 3 Indian batsmen of all time? NO.

    It doesn't matter what your criteria are. Be it raw numbers, quality of opposition, degree of comfort exuded at the crease, completeness of defensive and offensive technique, or even their intangible contributions to Indian cricket in terms of their legacy. You simply cannot look beyond these two. [[ I am with you all the way. I will only add, as my choice, RD into the group of two and make it a three-some and say all what you have said. To me, keeping any of the three out of the top-three group is travesty, Ananth: ]]

  • Boll on May 3, 2012, 13:45 GMT

    Ah well, we`ve all been dragged a bit off topic at times..all I can say is (after 30 mins of looking at graph - actually pretty easy when you get into it) - look at Hadlee. Is he the most important player any test team has ever had? and then you add his batting exploits. Of course Murali is hard to go past for his bowling, statesmanship, and fame...and then Imran appears.

    Three men whose achievements have clearly been more than just leather-slinging.

  • Gerry_the_Merry on May 2, 2012, 18:03 GMT

    Ananth, fair enough. Though I was worried that you may put a filter like "ignore innings runs less than 15" which would not affect the top-100, or even 1000, but severely curtail the analytical significance of the output, if one has to extend calculations in different directions.

  • milpand on May 2, 2012, 8:10 GMT

    Here is an out of context quote that can be applied to decisions made by umpires. It appears in The Times Opinion column by Daniel Finklestein - 'Times political columnist, football columnist and father of three but not necessarily in that order' - whom I particularly like for his weekly "Fink Tank" football stats column. " .. should realise that the number of mistakes you make is not only determined by the proportion of decisions you get wrong. It’s also determined by the number of decisions you make." [[ Very well said. One can say that, in Cricket, the number of runs you make or wickets you capture depends not just on your quality and skills, but on the number of matches you play. Ananth: ]]

  • A. Khan on May 5, 2012, 7:57 GMT

    If someone is really interested in knowing the s/r of batsmen from that era, open statsguru and use period of Gavaskar’s playing days. You can use filters like minimum 1000 runs and avg 30-35 (to avoid lot many names) and sort by strike rate. There are very few batters (Crowe is one) who had their s/r not as well as Gavaskar’s. Frankly, I too was surprised at the s/r of Gavaskar and Mudassar Nazar, I thought Gavaskar’s to be 42 and Mudassar, around 38. All said and done, if someone did that BQI/pitch etc. analysis for strike rates, would be in for a huge disappointment, as gavaskar’s strike rate drops to 40 outside the subcontinent. It’s all there in cricinfo database. But unfortunately ball by ball data is not available for all matches. @Ananth: Please bear with me for being off topic, again.

  • A. Khan on May 5, 2012, 7:46 GMT

    When I said that Gavaskar was a slow player, to counter this, people resorted to different tactics. One doubted on my smoking habits and other “knowledgeable” person brought the names of 3 modern players, who are known for their slow batting and Boycott (once dropped for slow batting)/ and “celebrated” Vishy, without even caring to know why he was celebrated? definitely not for fast batting and also some players from pre-1970 era. For a PERSON who cites old players left and right, it was surprising to see such ignorance. By this count Sobers (s/r 49) was a toddler compared to these Pontings/Laras/Smiths? Was he? If someone has the time to READ about so many things of the past, he should also READ about when the pitches were started getting covered and what were the consequences. Coming back to slow batting, I knew that Mudassar was slower batsman than Gavaskar but cricinfo shows only a difference of 2 in their s/r! cont..

  • A. Khan on May 5, 2012, 7:32 GMT

    Initially I thought will stay out of this. This is not in any way to undermine sunil, I consider his 96 to be one of the greatest innings that I have seen and possibly the best player of spin bowling among the players I witnessed. But I have seen what I have seen; unfortunately I followed (worldwide) each and every match that was going on starting from that Mumbai test of 1980 against England. Although I don't get to see many now for last few years, I still follow them on cricinfo or sometimes watch 'em live on internet. I am not a computer, but after watching/listening the matches, an image is developed in the mind as to how fast or slow a player was. Cricinfo doesn’t need to tell me about a Tavare or about a Kapil Dev, it mostly confirms it. While playing with Aussies, Wessles was an aggressive batsman but when he played for SA, became a really dour batsman; needless to say his captaincy was similar. I don’t have to learn this from cricinfo. cont..

  • west indian follower on May 4, 2012, 19:44 GMT

    hi ananth, great piece. Was wondering, would you consider S.F Barnes to be the greatest bowler and potentially the greatest test cricketer over Bradman from a weighted cricket statistical point of view. [[ 1. Barnes' performances are not comparable to Bradman. Batting Avge: Bradman 100.0, Pollock ( Next best) 61.0 RpT: Bradman 100.0, Weekes (Next best) 68.9 Bowling Avge: Avge: Lohmann 100.0 Barnes (Next best) 65.7 WpT: Barnes 100.0 Lohmann(Next best) 88.8. Murali (Third) 86.1. These numbers indicate how far ahead was Bradman to the others. Others were quite close to Barnes. Ananth: ]]Also, after warne and mcgrath, who would you consider to be the third best spinner of all time. [[ Question not clear. Ananth: ]] Finally, when picking a greatest side ever, would you look at how the players measured up against others in their era, or an overall point of view? I know it is not relevant to the article, but just curious for your opinion [[ Picking a greatest side is generally one's personal choice. I will pick my side irresepective of how the numbers measure up. Ananth: ]]

  • Gerry_the_Merry on May 4, 2012, 4:25 GMT

    Ananth, there has been some talk about strike rates (whenever Gavaskar is dragged into a debate, people try to trash him by referring to slow strike rate) in this forum.

    I looked through your past articles, but could not find an in-depth study on this topic in Test matches. In particular 1) who are the top 30% batsman sorted by strike rate 2) ordered by BQI groups, how did their strike rate behave 3) sorted by pitch conditions RPI/RSI etc. how did they perform 4) has this been a virtue or a vice from a single team's perspective (we know what crowds want, but they also want T20) - e.g. Atherton 185* in Johannesburg and Greatbatch 146* in Perth contrast with collapses in the great Calcutta and Adelaide tests against India by Aussies in less than a day, similarly India against Pak in Bangalore, following 267 by Younis Khan and Afridi onslaught 5) top batsmen who played extremely slowly (e.g. >150 balls and <30 runs) frequently...

    Gilchrist, Gayle, Ponting, Lara and Richards top picks. [[ As you know, only recently have I almost doubled my Balls played information segment of the Database. I now have nearly two-thirds of the Tests covered. So now it is possible to do analysis like what you have asked for. In fact I now even have a decent handle on the batsmen career strike rates by doing a pro-rata allocation of team balls for the non-available one third segment. That too I intend to even strengthen by using the available strike rate as a guidance factor to do to the pro-rate allocation. It is a miracle. But I have the actual strike rate for 5235 of Bradman's runs. Ananth: ]]

  • Vikram on May 4, 2012, 3:13 GMT

    @Boll: Actually Imran Khan should be considered the most important PLAYER of a test team. This analysis would have included his last few years when he really didn't bowl much and hence would have suffered in peer comparison. However he made it up through his batting, leadership and talent-finiding skills. His batting average is nearly 50% higher than Hadlee. In terms of BOWLER, Barnes/Murlai/Hadlee had the highest contribution to their respective teams over their entire career.

  • shrikanthk on May 4, 2012, 2:43 GMT

    @shrikanthk: I have seen enough comments from people where SRT has been ranked outside top 3 in even today's generation of Indian batsmen. Even in such selections, SMG would have to be in top 5.

    I wonder who these people are making such outrageous rankings. Let's think for ourselves. Is there any reasonable argument for excluding either SMG or SRT from the top 3 Indian batsmen of all time? NO.

    It doesn't matter what your criteria are. Be it raw numbers, quality of opposition, degree of comfort exuded at the crease, completeness of defensive and offensive technique, or even their intangible contributions to Indian cricket in terms of their legacy. You simply cannot look beyond these two. [[ I am with you all the way. I will only add, as my choice, RD into the group of two and make it a three-some and say all what you have said. To me, keeping any of the three out of the top-three group is travesty, Ananth: ]]

  • Boll on May 3, 2012, 13:45 GMT

    Ah well, we`ve all been dragged a bit off topic at times..all I can say is (after 30 mins of looking at graph - actually pretty easy when you get into it) - look at Hadlee. Is he the most important player any test team has ever had? and then you add his batting exploits. Of course Murali is hard to go past for his bowling, statesmanship, and fame...and then Imran appears.

    Three men whose achievements have clearly been more than just leather-slinging.

  • Gerry_the_Merry on May 2, 2012, 18:03 GMT

    Ananth, fair enough. Though I was worried that you may put a filter like "ignore innings runs less than 15" which would not affect the top-100, or even 1000, but severely curtail the analytical significance of the output, if one has to extend calculations in different directions.

  • milpand on May 2, 2012, 8:10 GMT

    Here is an out of context quote that can be applied to decisions made by umpires. It appears in The Times Opinion column by Daniel Finklestein - 'Times political columnist, football columnist and father of three but not necessarily in that order' - whom I particularly like for his weekly "Fink Tank" football stats column. " .. should realise that the number of mistakes you make is not only determined by the proportion of decisions you get wrong. It’s also determined by the number of decisions you make." [[ Very well said. One can say that, in Cricket, the number of runs you make or wickets you capture depends not just on your quality and skills, but on the number of matches you play. Ananth: ]]

  • Gerry_the_Merry on May 2, 2012, 3:55 GMT

    Hi Ananth, will your batsman innings rating be done for all innings in history or only the top few 100? [[ Gerry, Without analyzing each one of the 71941 innings played so far, I cannot arrive at the top-100. So it has to be a complete analysis. However how I present this infomation, what I make available freely, to what level of detail I go to are things which will be looked at carefully nearer the publication, which as I have explained in a recent response, will not be before 3 months. Ananth: ]]

  • shrikanthk on May 2, 2012, 3:04 GMT

    If so, why are spinners forever looking at variety of deliveries: doosra, theesra, (probably) chaudah (in 2013) et al.

    Look I'm not saying all the bowler needs is command over length. What I'm saying is that command over length is what matters the MOST. Eight times out of ten a batsman gets out because of a misjudgment of length and not a misjudgment of line.

    And seldom do you find a first-class bowler who does absolutely nothing with the ball, be it a seamer or a spinner. I am sure even Nadkarni could spin it hard on a helpful wicket if he wanted to. What separates wheat from chaff is not "variation", but good old-fashioned persistence.

    There are similar examples even today if you care to look for them. Darren Sammy for instance. The guy is perhaps even less skilful in his trade than what Nadkarni was as a spinner. His greatest asset is accuracy. And he has done quite well. Sammy averages around 27-28 in both Tests and FC. Far better than someone like Fidel Edwards. [[ Yes, I agree with you, Shri. Length is paramount. And I am the guy who is testament to the fact that Nadkarni really had variations but played in a style suited to the team needs: a la Gavaskar. His 11 wickets at Madras in 1964, my first Test as a gangling teen, was a revelation. I like your exacpmle of Sammy. WIth seemingly no speed and variations and accuracy as his forte, he has 64 wickets in 24 Tests at an average under 31: acceptable figures when combined with a near-20 batting average. Ananth: ]]

  • shrikanthk on May 1, 2012, 16:39 GMT

    Nadkarni, in today's environment, would not have survived the Test scene

    I don't agree. Back in his day he averaged around 29-30 with the ball. No reason why he couldn't have enjoyed a similar level of success today.

    Ofcourse I haven't seen Nadkarni. But I am always wary of all this talk of a bowler being "too defensive". To paraphrase Sydney Barnes, bowling is about 3 things - "length, length and length". All other things come later. If you have command over those "three things", rest assured you'll have a decent average if not a great strike rate. [[ A dictum told about 100 years back holds good today !!! If so, why are spinners forever looking at variety of deliveries: doosra, theesra, (probably) chaudah (in 2013) et al. If it is only about length why do we see demise of spinners who do not have great variations. Ananth: ]]

  • Youvi on May 1, 2012, 16:12 GMT

    Anantha- This is an elegant piece of analysis which you have made it simple for all, a difficult thing to do. As a kid in the 70s I had chance to watch Nadkarni bowl in a domestic game. His top SPF-2 performance here is signifcantly higher than the rest which is amazing. Growing up the I had heard and then read so much about this feat. At the stadium during the game I allude to earlier when I watched Nadkarni bowl in the 70s, there was a mini- debate during a rain break among tthe people surrounding us on such bowling feats. Some questioned the utility of such a feat where the focus is on containment rather than getting wickets ! How times have changed as having a Nadkarni in the shorter format would be a boon to any team! Tx [[ Nadkarni, in today's environment, would not have survived the Test scene. He was too defensive a bowler. Although I have seen him flight the ball once in his life in 1964 and doing very well. However in the limited formats he would have been very difficult to get off with his phenomenal accuracy and would have secured quite a few wickets. Ananth: ]]

  • Alex on May 1, 2012, 15:21 GMT

    @Ananth: It is surprising that SMG still evokes passionate response on your blog. He is charismatic with very good PR & career management skills. That probably is one of the main reasons.

    1. This is a digression but I am puzzled that you put SMG & RD ahead of SRT in tests. Pl explain. Apart from one great series, SMG always failed in Eng & RD in SA & Oz. SMG's failure in NZ & Oz vs good attacks is noticeable. SRT's peaks have not been as great as those of SMG & RD but he has no such great weaknesses. E.g., if you take the two best series of SMG away, his average drops from 51 to 47. Dravid's drops from 52 to 49. In contrast, SRT's drops to 54 only. We can repeat the same exercise by taking out 3 or 4 best years ... SRT did well everywhere, year in year out, series in series out. [[ Alex, I have given you a lot of rope to dwell on your very strong pro-SRT stance. However now you have turned on me with your ludicrous comment n my being influenced by SMG's PR skills. I am the last person to be influenced by that. Frankly I do not like Gavaskar. As for explanations, forget it. I will not give one. Everything does not have to be backed up by numbers, often presented in a selective manner. Ananth: ]]

  • unni on May 1, 2012, 11:18 GMT

    Alex's link talks about Power mean which is a generic formula encompassing all other mean formulas. Still to use that you need to choose the value for p as -1, 0 or 1 in which case it becomes harmonic, geometric, arithmetic mean respectively. Generally geometric mean usage here has more logical meaning, but unfortunately 0 values make them non-usable. [[ The one solution would be to change the ratio itself. Instead of running from, say, 0 to 44, it runs from, say. 1 to 10, by a mapping exercise. And this range can be 1 to 10, irrespective of whether the actual range is 0 to 44 (Spf-1), 0 to 12.5 (Spf-2) or 0 to 23.3 (Spf-3). This would be a specialized mapping exercise with the specific intention of normalizing the numbers so that a better mean, Harmonic or Geometric, can be done. But I am deep in my next series of articles. So this has to wait for a while. Ananth: ]]

  • unni on May 1, 2012, 8:11 GMT

    @Ananth : I agree with Alex in that it needs some more thinking when combining the values from different innings. As far as I remember this was not settled for the batsmen case last year and you were on the opinion that we shouldn't even combine these indices(means the career based index shouldn't be calculated). When it comes to bowlers, it is even more complicated. But, I feel that we could still combine them. One way is that the way you have made the special case for SPF1. i.e, when there is a zero in the denominator assign 0 to the index. This is a very special case and it is tailor-made for AM. No issues, if AM is logically fitting. But, I have a gut feeling that we would miss a better method if we ignore this. And I see that if you assign zero to such cases, then GM is impossible. This is an intriguing little problem and I'm sure somebody would have already solved it. If you discuss this offline, pl. include me also. @Alex : Your key point is to use GM, right? If yes,I agree [[ Unni, please go through the link provided by Alex on Power Mean. Looks good. I think I will try that out at a later stage when I have some spare time. I am not hung up on anything. I only do not want to assign completely arbitrary values to replace 0, that is all. Ananth: ]]

  • Gerry_the_Merry on May 1, 2012, 7:59 GMT

    Would have to agree with Vikram on all his points. Additionally on Lillee, Sunny was challenged, during a tour in 1980-81 where we were being beaten black and blue, were down 0-1, on how he called himself a great batsman when he could not play on bouncy wickets in Australia (in reality we know the matches were in India friendly wickets Adelaide, Sydney and a crumbling MCG). Sunny replied that he did not rank himself in any case, and further that why did Lillee fail in Pakistan (3 wickets in 3 tests in 1979-80) since he should by the same yardstick succeed on flat tracks??

    Then we have his book recording that "Umpire Constant was constant in his support for England", and "The West Indian crowd's behaviour showed that they had not come down from the trees".

    Turning down MCC life membership and being criticised by Indians, especially Bedi...

    I am surprised that Sunny ever gets mentioned by any former player. I could see the same streak of defiance in Ganguly in tackling S Waugh.

  • Vikram on May 1, 2012, 6:31 GMT

    My last comment on SMG topic, atleast in this thread. I haven't seen enough of players like Hazare and Manjrekar Sr but there have been comments which rate them positively. Similarly Vishy has been considered better by some commentators and players. My point is that there are names in that pot but however people rank it, SMG will be there. @shrikanthk: I have seen enough comments from people where SRT has been ranked outside top 3 in even today's generation of Indian batsmen. Even in such selections, SMG would have to be in top 5. @ Alex: I didn't mean to say that Lillee is a nut case. As a bowler, he will be ranked pretty high in the all-time list. However, the point is that Lillee saying SMG is not his top 5 or whatever, is same as Warne saying SRT is the best. It counts for something but a very small something. My ranking in Indian batting: SRT/RD/SMG My ranking of batsmen in contribution to Ind test cricket: SMG/RD/SRT (Amongst all players, Kapil and Kumble rank above RD). [[ Second you in all these statements. Ananth: ]]

  • Alex on May 1, 2012, 4:33 GMT

    @Ananth: Fine, thanks. I am happy to take the GM/AM discussion offline and help you analyze that series.

    @shrikanthk:

    1. Why take things out of context? Hutton was possibly talking about only the emotions evoked in him. Also, "Indian cricket grew that day" is a complacent coffee table "sahib" category comment that arrived when Ind fell just short of the win in the 4th & final Oval test, thereby ensuring that Eng won the series 1-0. Incidentally, DBV & GRV hitting centuries in a dire situation in the 2nd test was the major turning point for India.

    2. There is a reasonable case to put RD ahead of SMG. IMO, all RD lacks is the media savvy and shrewd PR skills of MSG & RJS. RD, like Kumble & Ganguly, is actually quite cerebral & articulate. [[ When one looks at the overall contributions in Tests to the Indian cause, I am not certain it is a clear-cut SRT/SMG/RD sequence. Without even bringing in the shortened-format word once, SMG's and RD's contributions towards Indian Test cricket have been no less than SRT's. I would even say, clearly ahead. Between them, I would give SMG the marginal 0.1% nod, only because he never had the batting wealth around him and the tough away tours undertaken. Ananth: ]]

  • Gerry_the_Merry on May 1, 2012, 4:26 GMT

    I was keen to avoid further nit-picking on Sunny, but Alex's comment on when Hutton wrote, makes it seem I have got my facts wrong, so I must correct.

    Hutton's book was printed in 1986, but I stand corrected - it was first published in 1984. Some post-1980 cricketers are referred to in the book. The book itself is called "Fifty Years in Cricket" and Hutton made his debut in 1934. The book contains a para from Wisden 1983. Hutton specifically refers to Gavaskar's peak, which he would not need to do if written in 1979, when he was at his peak, and so was Richards.

    Some other comments are also interesting - he rates Lillee + Thompson better than West Indies quicks. He rates Holding as the best of the pace quartet. He naturally rates Lindwall - Miller higher, but his style of writing is typical of his stature - taking a panoramic view, rather than get messed up in detail.

    After this discussion batsmen will be afraid of ever retiring. "SRT retired in 2013, wary of Eng+Oz in 2015". Edited by Ananth.

  • Alex on April 30, 2012, 18:25 GMT

    @Ananth: You are free to choose whichever metric you like but do note that this series features infinite values (or, alternatively, zero values). In general, given a series featuring zero values, the generalized mean is a better choice than the arithmatic mean. For more info on generalized mean, see http://mathworld.wolfram.com/PowerMean.html . Arithmatic mean is a special case of it. [[ Will look into it. Ananth: ]]

  • shrikanthk on April 30, 2012, 17:47 GMT

    He ranks 221 at the same level as McCabe's 232, and says that Indian cricket grew up that day. He ends by saying that at his peak, Sunny was undeniably the world #1 bat

    These are the sort of remarks which make me not pay too much attention to the punditry of Hutton!

    Gavaskar's 221 and McCabe's 232 are two very different types of innings. I can't imagine how anyone can draw a parallel between these innings.

    McCabe's 232 was an explosive innings on a flat pitch against a spin-based England attack which lacked Bowes and in which Verity bowled too few overs. Gavaskar's 221 was a polished well-paced innings of great defiance against an excellent seam attack (Willis/Botham/Old).

    There is no parallel between the two. I personally rate DGB's 144 from the 2nd innings of that Trent Bridge test higher than McCabe's 232. DGB faced 62 overs of Verity, whereas Verity did not bowl much in the 1st innings!

    Also DGB's innings helped Aus save the game, regardless of McCabe's knock

  • shrikanthk on April 30, 2012, 17:11 GMT

    SMG has to rank amongst the top 5 Indian test batsmen

    What an understatement! It's like saying DGB is among the ten greatest batsmen of all time.

    SMG is among the two greatest Indian batsmen of all time. Period. Placing him outside the top two is not even an option for most people.

  • Alex on April 30, 2012, 17:10 GMT

    @Vikram: Since most readers on this forum are Indians, Ind players get debated the most here. On forums dominated by Eng, Oz, & WI supporters, their players get the most printed characters. They compare Pietersen & Cook favorably to SRT, and tout Anderson as the best bowler since Lillee. Prior to 1994, Pak media & fans found it hard to say one good word about even SMG, forget the rest of the Indian players. Things changed for better when India became an economic power. So, given the disdain in which India and Ind cricketers were held for so long, no harm if SRT and some other Indian players gets a trumpeted a bit too much on this forum.

    As for those 12-14 players, I said "arguably" and not "surely". Lillee is not a nut-case. He is a nice guy who is a straight shooter and a well respected coach. He was the best bowler of his (& SMG's) era and taught freely, even in India. Many great bowlers learnt from him (e.g., Imran & Hadlee) and Hadlee modeled himself after Lillee.

  • Alex on April 30, 2012, 13:43 GMT

    @Unni & @Ananth: The method I suggested will work for RPO. Say 3 bowlers bowled as follows:

    Bowler 1: 10-2-20-3 Bowler 2: 10-0-15-0 Bowler 3: 10-1-40-5

    Then, RPO's are, respectively, 2, 1.5, and 4. The average RPO is 2.5. Hence, the weighted RPO for each bowler becomes, respectively, 2/2.5, 1.5/2.5, 4/2.5. For every bowler, compute these weighted RPO in every match and then take the geometric mean of this series ... this is a better metric than the average.

    This will not work with SR and RpW because a bowler might not take a wkt in a match, thereby assigning infinite value to these metrics. So, approximations are needed. As a remedy, the infinite value for SR may be substituted by, say, (60 balls X 4=) 240 for a regular bowler (since he _will_ take a wicket in 4 matches) and 1000 for a non-regular bowler. Likewise, the infinite value for RpW may be substituted by 240 * (average RpO for the bowler). [[ The RpO Factor does not need any change at all. The highest is 12.xx and the nxt one is 4.xx. What is wrong with a simple AM. Why go out of the way and tweak when there is no need to do any. As I have already mentined I am ready to accept a way-out match for Morkel and his place in the top-20 bowlers for the sake of simplicity. Tomorrow if Deonarine scores a 401 and does not score another 100 in his Test career, surely we are not going to say, let us exclude the 401 while doing his batting average. No one has come out to my reference to Gillespie. The bottomline is simple. If there are going to be zero values, do not use anything other than AM. Ananth: ]]

  • Gerry_the_Merry on April 30, 2012, 7:29 GMT

    Ananth, to help us properly salivate over your next few articles, can you please give a quasi head's up on the topics? Waiting for the batsman innings rating and bowlers across batsman groups especially...and in tests, may I add... [[ It has already been done. Either in the last or the previous article. Ananth: ]]

  • Vikram on April 30, 2012, 5:56 GMT

    @ Ananth: the limit on words meant I had to delete the part where I wrote that I don't want to compare these two innings in quality or any other aspect. I was talking about the batsmen only because the question arose that SRT doesn't do well in ODI matches that matter. So if one innings of BCL makes him the a winner, shouldn't it apply to SRT as well, and specifically for ODIs that matter? It is not about comparing the two innings. I have watched the BCL innings 5 times, which is 5 times the SRT innings. So I respect that innings. It was about the brand value that that innings should provide. I don't believe that SRT is head and shoulders above other batsmen in ODIs, he is in the +/-2% range of atleast 5 other batsmen. However, to say, that he doesn't do well in ODIs that matter is as wrong as saying that that innings of BCL was not amazing. It was meant to be a provoking statement, but I think I didn't word it well enough. [[ No, the real problem is that we should not compare great efforts in different formats. It raises unnecessary arguments. Anyhow Lara's innings made him a winner in that match, And makes that innings a winner. That is all. Again, please do not bring in an unnecessary Lara/Tendulkar comparison, that too across formats. Ananth: ]]

  • Vikram on April 30, 2012, 4:35 GMT

    contd. 3. Of the names of batsmen listed, I am sure we can easily debate if Lloyd, Greenidge, Gooch, Crowe and Boycott were better than SMG. For other names, maybe true. 4. Since when did Lillee become the deciding factor for the worthiness of a batsman? He was a great bowler but definitely not the most sorted of people (aluminium bat anyone?). By that same token, the greatest bowler of this era, Warne, has said that SRT is the best batsman, so that debate should be over right? 5. I find it surprising that 1 BCL innings of 153 n.o. makes him the greatest winner but SRT's 143 and 134 in consecutive innings in a virtual semi-final and final don't make him the greatest winners in ODI. (please note the word winner). No reason to run down BCL, and I apologize for that, but my question is - why different yardsticks? [[ Normally, Vikram, you are a clear thinker and your comments are well-defined, I have even commented on this in this blogspace. Unfortunately you are getting your lines crossed here. How can you ever compare a Test innings played on the fifth day of a tight Test match with a pair of ODI innings played in a tri-series tournament, one of million, played in a day-night game in Sharjah. I am not going to even answer this. I would appreciate if this is the last time such a comparison is ever done. Tomorrow someone might say the 281 is superior to 189* or the 175* is better than the 154 and so on. Milind might not like it, but definitely, chalk and cheese. My comment on the 85 was only on the importance of the innings from the overall WC win point of view. If the 141 and 134 had not been played, India would not have won a Sharjah Tri-series, that is all. Ananth: ]] 6. SMG has to rank amongst the top 5 Indian test batsmen. He was a defensive batsman in a poor team and put "not losing" as his target and was fairly good at that.

  • Vikram on April 30, 2012, 4:19 GMT

    Unfortunate that such an interesting analysis of bowlers has become a discussion ground about SRT and SMG. Is it only Indian batsmen who have such vehement disputes surrounding them or are there similar forums for other batsmen/players as well. A couple of points from me as well 1. Any Indian player plays cricket for India. India plays Tests and ODI plus T20I (not IPL) so when a player should be analyzed it should be based on what he did as a player. We can analyze their test performances, ODI performances etc etc separately but as a player, they need to be analyzed across all formats because the team they represent plays all those formats (unless they choose not to participate). 2. You can say that I dont like a player because he is defensive but you can not (or rather should not say) that he is not a good player because he is defensive. Don's XI is his preferred team not necessarily the best team. There's a difference. contd.

  • Alex on April 30, 2012, 0:21 GMT

    @Ananth and @Gerry: Gerry's objection to SRT's failures in important matches has a sound basis. After 1998 through 2004, i.e., for fully 6 years, SRT was a terrible performer in semis & finals. In this period, his record:

    1. finals: 15 innings, 310 runs, two 50, ave=22. 2. semis: 3 innings, 63 runs, no 50, ave=21 ... I left out 83 vs Ken.

    Thereafter, he played final/semi in 2008 only. Thus, effectively, for fully 10 yrs since late 1998, he played in 18 finals/semis and posted ave=22 in finals and ave=21 in semis ... this is failure by any yardstick and more so for a batsman whose physical age was 25 through 34 in this patch.

    Gerry seems to focus on this phase only even though, amazingly, SRT bounced back on this count starting '08.

    @Ananth: Fully agree on SRT's 85. However, his brilliant 53 vs Oz was equally important, IMO. Ind-Oz was a better quality knock-out match and Oz was a tougher opponent & 3 times defending champion. Pak was a bundle of nerves. [[ Aesthetics of the innings, quality of the match, reputations of the opponents etc do not come into the picture at all. The Pakistan innings was twice as valuable as the Australian innings, borne out by the Rating points: 370 vs 187. Between the two, the Pakistan innings wins by a mile. And the two innings together. SRT ranks only behind Steve Waugh (1999) in the pre-Final knock-out contributions in a single WC. I know 1999 vs SA was not really a knock-out, but in reality it was. Ananth: ]]

  • Pranav Joshi on April 29, 2012, 17:42 GMT

    @ Gerry the Merry

    I would say that your statistical knowledge of cricket is woefully incomplete, and you continue to display the unwillingness to find out.

    How many knockouts did India play in WCs during Sachin's time? Would you even blame him for scoring 65 in a SF in which his team faced an abysmal collapse after his dismissal? Were SL not a tough team then? Was the pitch not difficult.

    Lemme give you another stat. From the '96 WC onwards, Sachin played 10 knockout games in WCs (counting the 2nd and 3rd super six match in the 99 WC). He scored 4 fifties in those 9 knockout games.

    And a statement like "any tournament that ends with Zimbabwe in the final says a lot" is ridiculous. The other team were SL.

    Here's a list of SRT's significant innings against top opposition in WCs:

    54, 84, 70, 90, 137, 65, 98, 97, 120, 111, 53, 85. [[ I agree with your selection. I maintain the importance of 85 only because of the end result in the tournament, that is all. Like no one should forget Steve Waugh's 120 and 56 in the two matches against SA in 1999, but for both innings, there was no Final for Australia. Ananth: ]] I have not begun on his 10 fifties in non-WC finals yet. Not enough space here.

    But if you are determined to close your eyes, no one can help you.

  • Alex on April 29, 2012, 16:32 GMT

    @shrikanthk: SMG was not ultra-defensive but certainly content to wear the attack out. The Don had a major objection against this mindset. In any case, we don't know if the Don's All-Time XI is actually his. Also, in reality, SMG's ODI mediocrity existed well until 1985. India winning the B&H under him in 1985 seemed to remove a mental block for him and he suddenly became a pretty good ODI opener at the ripe age of 36. He had rarely smiled or joked around on field until 1985 but that became a common sight after the B & H until his retirement.

    @Gerry: As per your Olonga argument, we should trash the '96 WC since SL actually won it: what was their bowling attack and didn't they get thrashed routinely outside the subcontinent & Sharjah? Also, your collection of SRT's 4 innings is nice but misses out on 95 vs Pak & 74 vs SL. Also, IMO, SRT's greatest ever ODI innings is 138 vs SL, played in an away final but you don't refer to it. Why such an obsession with skin color?

  • Alex on April 29, 2012, 15:57 GMT

    @Gerry:

    1. To judge a player in his era, I look at it with a cover of plus-min 5 yrs. E.g., for SMG, look at 1965-1991 over which the following 12 are arguably better (disregarding aggregates, place in history, importance to his own team, etc.): Sobers, Kanhai, Viv, Lloyd, Greenidge, Boycott, Gooch, G Chappell, AB, Miandad, & Crowe; one may also consider Pollock & B Richards.

    2. Roberts' rating used to be: 1. Ian Chappell; 2. SMG; 3. Boycott. I don't know if Sobers ever wrote an article rating batsmen he has seen. His recent book has an entire chapter on Lara and not a single word on SRT.

    3. I distrust many of the utterances of the Britishers since they don't shoot straight. Anyway, Hutton's comment was @1979. SMG was phenomenal in the Packer depleted era (1977-79) in which he did very well vs Imran, Botham, Willis, & Thommo. So, in '79, many believed he will do as well vs Lillee & other top bowlers. But that did not happen.

  • A. Khan on April 29, 2012, 9:09 GMT

    @Gerry_the_Merry My sincere apologies if my language came out poorly. As you said - everyone has his own opinion, I have mine and for sure Len/Sobers(who in 1979, rated Viv/Barry Richards as the best batsmen in the world)/Hutton can have theirs. Surely these guys are/were not senile, and for the records, I smoke plain cigarettes (That was quite eloquent from your part).

  • shrikanthk on April 29, 2012, 7:25 GMT

    One last comment on why I think Hobbs is a truly remarkable figure in the history of sport -

    In the summer of 1925, he played 30 first class games! Yes. 30. And scored 3024 runs at an average of 70.32 with 16 hundreds. All in a single summer. So over a four month period, he was playing cricket for 90 days! All this at the age of 43!!!

    Lots and lots and lots of people will make huge sacrifices to hit 16 first-class hundreds over a career, leave alone a season! And this guy did this when he was in his forties! Truly amazing.

  • Gerry_the_Merry on April 29, 2012, 6:43 GMT

    ...and not given to irrelevant comments. It is in 1986, when KApil Dev had already scored 175*, Richards had already made 189*, India had shaken up the world by winning world cup and WCC '85, and Gooch had scored 129.

    The chapter itself is called "Gavaskar and the other greats". He says "Gavaskar is not as good as Bradman, but very close". He ranks 221 at the same level as McCabe's 232, and says that Indian cricket grew up that day. He ends by saying that at his peak, Sunny was undeniably the world #1 bat. He rates Richards as brilliant but somewhat unreliable.

    All this is fromSir Len Hutton. If you guys rank him also as senile, you ahve to disclose what you ahve been smoking.

    Does this make Sunny #1 in my books? Not at all. He is #2, after Richards. Everyone has his own opinoin. But I would challenge anyone to bring me 12 players in the 1971-87 era about whom three greats have written in such unambiguous terms.

    AQ Khan's language has killed my interest in this.. Am signing off.

  • Gerry_the_Merry on April 29, 2012, 6:35 GMT

    ...I would never include any of them in career stats as there is no true tension in such matches, and I have seen even Tendulkar give McGrath a royal pasting in such festival matches. I only meant to say that Sunny was in good touch, since Vengsarkar and Haynes were easily bounced out by Marshall.

    5. Pranav, I anyway think that Tendulkar is one of the top 5 batsman of his time. It is just that you guys rate him even higher, which I dont. But that is merely my opinoin. Just as rating Sunny only in the top 12 is Alex's opinoin.

    6. But where can we say that esteemed players rate Sunny? We have already seen the "Capricious" Sobers rate him as the best he has ever seen. A few months back I posted a similar link (Simon Briggs interviewing Andy Roberts in 2009), where Roberts made it clear that he thought Sunny was the best.

    7. Here is a tribute that will make Alex reach for his smelling salts, written by a highly esteemed player, an automatic choice for any World XI (first or second)...

  • Gerry_the_Merry on April 29, 2012, 6:30 GMT

    1. I dont know whether to laugh or to cry at the obsessive urge to mix up ODI and tests. Dravid, the best batsman in the recent disaster in England, was not picked as a one day batsman for several years. The two games are completely different.

    2. Pranavm you have answered your own question, by mentioning Olonga etc. A tournament that ends up with Zimbabwe in the final says it all. In reality, Tendulkar's great innings are 143, 134, 117* and 91. All other tournaments are likely to have featured teams which did not include their best players.

    3. AQ Khan, Alex and Pranav need to make up their minds on whether dropped catches, plumb LBWs are counted or not. If yes, then Sunny's 188 is useless, but so is Tendulakr's 85 in WC SF, and hence we are left with no meaningful knockout contributions against top teams in 4 world cups but a bunch of failures. [[ I would never ever downplay an innings because of what happened within the innings. Lbws come and go, catches I would have taken, in my sleep, are dropped, catches no one should ever have taken get taken, umpires raise their fingers when the ball is defleced off the bat into the pad which is 12" away or umpires keep their fingers steadfast to themselves when the batsman pads to a ball 6 inches in front of middle stump, umpires lift their fingers with alacrity when the ball and bat are 6" apart and the wk appeals, umpires keep their fingers in their pockets when there is a deflection off the bat to second slip and so on. Tendulkar's 85 might have been an awful innings by any standard. BUT IT IS THE MOST IMPORTANT INNINGS PLAYED BY TENDULKAR IN ODIS. But for that innings, Pakistan and Sri Lanka would have contested the Final at Mumbai. So it was an extremely invaluable innings. Period. I do not know enough about the 188 to comment. It is not official. Ananth: ]] 4. Alex, I anyway think that 188 by Sunny is useless. All these matches, Kerry Packer, etc. are irrelevant as per me. I would never...

  • shrikanthk on April 29, 2012, 6:30 GMT

    However it is presumptuous to assume that playing Cricket at 45 is tougher than playing (and winning) in Golf at 45

    Ofcourse both are tough propositions. I am just saying that I would regard cricket as slightly tougher. And many would agree. Several golfers have enjoyed success in their late thirties/forties - Ben Hogan and Gary Player for instance. It is not uncommon.

    It is not uncommon in cricket either. We have cases like Sydney Barnes who was a phenomenon in league cricket into his sixties. The greatness of Hobbs is that his whole legacy owes a lot to his achievements after the age of 40, unlike say Nicklaus who still built his legend mainly before the age of 40.

    More than half of Hobbs' FC centuries were scored after the age of 40. That's just too astounding for words. And during those years in the twenties, he averaged 50+ practically every season often faring better than much younger players like Woolley, Hammond, Sutcliffe.

  • shrikanthk on April 29, 2012, 5:59 GMT

    There have been great players who have not won a single major. Jack Nicklaus, after he turned 40, won 3 majors (equivalent to and no less difficult than Tennis Grand Slams) and finished runner-up on 2 occasions. In addition he won 2 other Tour events, equivalent to the Tennis Master's events.

    Not dismissing Golf. Nevertheless, I guess everyone agrees that playing top class cricket at 45 is atleast slightly harder than playing top class golf at the same age. [[ I do not want to start an argument. However it is presumptuous to assume that playing Cricket at 45 is tougher than playing (and winning) in Golf at 45. 72 holes over 4 days, 4 to 5 hour rounds, upwards of 5 kilometres of walking, not a minute to relax et al. In fact, I would go to the extent that playing a 4 day major at 45 is more strenuous than a playing a Test match. In a Test, when you are out, you can come back and relax for the rest of the innings. Not in Golf. Anyhow, Shri, these are two totally different sports. One is individual and the other, team. Do not compare. Leave them in their own sphere. Ananth: ]] I'd rate a 100 first-class centuries after the age of 40 a bit higher than Nicklaus' achievement. And during that period, he was also one of the best batsmen in the world even in Test cricket. He played widely acknowledged masterpieces on gluepots even as late as 1928 when he was 46.

  • A. Khan on April 29, 2012, 4:17 GMT

    Ananth Interesting analysis. I never thought Morkel's figure would be so unique! Although I dont agree with adding one wicket to others when they have not taken a wicket, but then again, there is no better alternative to this. Besides, this puts Morkel on top. The most interesting table (to me) is the SPF-2 table. Compton, Umriger and Manjrekar were real surprise. [[ First, do not think of it as adding 1 to the wickets column. I have decscribed it in the article itself. One not out innings of 50 should lead to an average of 50 and not infinity or "no average". This problem does not exist in bowling. However the ideas are similar. 0 for 150 should lead to an average of 150 and not infinity. And Congdon and Burke are also really batsmen. Manjrekar's was in that awful match which was on one of the most batsmen-friendly pitches ever. 1100 runs for 10 wickets. And Manjrekar, what type did he bowl, bowled 20 of the 58 overs. "Occasional off-break bowler", C M-J says. Ananth: ]]

  • shrikanthk on April 29, 2012, 3:19 GMT

    which is why the Don did not include him (or Boycott/Hutton/Hobbs)

    Boycott/Hutton I agree. But Hobbs is a different matter altogether. Don probably didn't include him as he saw Hobbs only towards the end of the latter's career when Hobbs was nearing 50! I guess he was loath to include him based on that evidence.

    If one looks at Hobbs' career as a whole, you'll realize that he was a very consistent opener famed for his controlled aggression. Hobbs pre-War was in fact a very aggressive opener scoring often at 4 runs per over.

    Even his batting in the 20-21 Ashes series against an excellent Aus attack (McDonald and Gregory) suggests that he was an extremely free-scoring batsman (you can check the SRs on cricinfo). And let's not forget that he was approaching 40 in '20-21 already.

    Hobbs' achievements never cease to amaze me. I don't think there has been any outdoor sportsman anywhere in the history of the world who has achieved so much after the age of 40. [[ Let me bring out a comparable achievement. First do not dismiss Golf as a simple sport which can be played at any age. The competition level at the top is fierce. There have been great players who have not won a single major. Jack Nicklaus, after he turned 40, won 3 majors (equivalent to and no less difficult than Tennis Grand Slams) and finished runner-up on 2 occasions. In addition he won 2 other Tour events, equivalent to the Tennis Master's events. And he played competitive golf at the highest level for nearly 10 years. Ananth: ]]

  • shrikanthk on April 29, 2012, 3:12 GMT

    However, he was a poor improviser and primarily defensive, which is why the Don did not include him in his all-time XI and why his ODI record is average.

    Fair enough. But these are the dominant perceptions in people's minds reinforced by "legendary" innings like the 36 not out at Lord's, in an era when several people didn't take ODIs too seriously.

    If you focus on facts, one observes something slightly different. Gavaskar's SR in tests was around 44 as per Ananth's research. That SR is better than that of Dravid, Chanderpaul, Kallis, Ian Chappell, Boycott, Hutton, and well even the "flamboyant" Mr.Hammond!!! Why, it's even better than that of his celebrated brother-in-law GR Viswanath!

    So obviously the man was not as defensive as people imagine him to be! An SR of 44-45 is not slow by any means, given the era and the team he played in.

    I am not saying he was a swashbuckler. But nor was he a blocker! He was no less aggressive than the average test batsman of his era.

  • A. Khan on April 29, 2012, 3:02 GMT

    about what to do, at what rate he should be scoring? What is the target to be set. In almost all the matches where India won and he scored runs, the real defining innings were played by some other player in the side. Putting him in the same bracket as some of the really greats of the game (such as Sachin) is no surprise to me. I stopped being baffled when I got into a cricketing discussion in a forum some 8 years ago and I was enlightened that not only Wasim was a better bowler but also a better allrounder than Imran, believed by, well, almost everybody! Before that I thought the only area where he was better than Imran was in fielding, he was a poor fielder and Imran, poorer.

  • A. Khan on April 29, 2012, 2:36 GMT

    @Alex Indeed. I don't think people do really understand, at most they would click cricinfo summary, 13 centuries, 65 avg, against most fearsome attack, awesome! Boring as he could be many a times for not making the scoreboard move, he was a purist's delight. I still consider him the best player of spinner that I have seen (not just coz 96). I remember one snap in sportstar during his playing days where he and a windies player were together and there was a caption that Sunny seems to be saying him that I know your team has got great fast bowlers and have beaten teams all around the world, but I have scored 13 centuries and avg 65! The media deliberately act ignorant many a times. And you are right, just as Tendulkar has a weakness against incutters, Sunny had weakness against legcutters/Out swingers. Although Sunny's record doesn't look as bad, with avg of 35, in ODIs, he was a really poor player in the shorter format, especially in the first innings where seldom had any clue cont..

  • Alex on April 28, 2012, 15:14 GMT

    @A. Khan: Today, SMG bandwagon overlooks his true defining merits but lauds him to skys. I include Boycott in this bandwagon: as a commentator, it is in his best interests to do so!! To this day, Lillee doesn't consider SMG a great batsman. I idolized SMG as a kid. He batted as if batting was his salvation and taught at least the subcontinent the virtues of disciplined batting, big scores, longevity, & consistency. His technique was superb barring a weakness vs medium-pace/fast bowlers for deliveries moving away outside the off. He had enormous powers of concentration and organized his innings very well. Also, stylish (even in defense) & dignified with an array of textbook strokes.

    However, he was a poor improviser and primarily defensive (121 & 90 etc. are exceptions), which is why the Don did not include him (or Boycott/Hutton/Hobbs) in his all-time XI and why his ODI record is average. Today, SMG says he wanted to attack but couldn't, perhaps due to a mental block.

  • A. Khan on April 27, 2012, 5:12 GMT

    @Alex Alas, I never thought that LBW let off at very early in SMG's first innings 188 will be remembered by any one! Commentators opined that probably the unofficial nature of the match led this to be given notout, it was so plumb. I dont know how he would fared in the forthcoming windies series, but his averages were slowly but surely falling. Being a conscious man that he was, he timed his retirement pretty well. He averaged around 58 till 1980, but after that he averaged 45 (from 1981). And for the delectation of some people he averaged 28 in the 23 matches (7 wins avg 29), which produced results rubbing shoulders with Kapils and Kirmanis. Mohinder and Vengsarkar fared much better. But yeah, when ever he got the CHANCE he took it with both hands. That's the advantage with the tukku tukku batsmen, they more than make up when come across dead tracks or poor attack. Some 20 years down the line we label them as match savers based on a handful of really match saving innings.

  • Pranav Joshi on April 26, 2012, 8:26 GMT

    @ Gerry

    Your 2nd definition is wrong, purely on factual basis.

    Just as a snapshot - 40 finals. 10 scores between 50-99. 6 scores from 100 to 138. Average of 54.

    India won all six finals in which SRT scored over a hundred. One century came against Kenya. One against Zimbabwe - in 71 balls. (this was when he took revenge on Olonga). Two centuries came against Australia, one in Australia in 2008, the other following the epic 143. Both were memorable in their own right. Two centuries came against SL. 3 of the 6 centuries came while chasing.

    All 6 centuries came after 1996.

    The above is among the best records in ODI finals, EVER. SRT's two failures in WC finals have clouded everyone's mind. But I suggest you look at his SF knocks as well, both in World Cups and in the ICC Knockout tournaments. All significant!

    At least, SRT has been India's best ever batsmen in ODI knockouts.

    Also, look up his performance in away test series in the defining match, with the series 1-1.

  • unni on April 26, 2012, 4:21 GMT

    @Ananth : I didn't get how can there be zero values for SPF1. To get SPF1, you have already eliminated zero possibility from bowler average(otherwise you will have to divide by zero !). And from the above tables I see that you eliminate zero from rest-of-team average. So, no chance of zero for SPF1. What did I miss? @Alex : I went through the article, but unfortunately couldn't deduce your point myself. elaborate? [[ Unni, Zero values are present in a big way. Any spell in which no wicket is captured is awarded zero for Spf-1 and Spf-3. In fact Nadkarni's 32-27-5-0, which leads the Spf-2 table, by a few miles, gets zero for Spf-1 and Spf-3. That is the way it should be. Out of qualifying 34500+ spells, over 9000 have 0 values for SPf-1. This is done right at the top so that division by zero and other nonsense does not come in. And replacement of 0 for 1 is done only for the special cases where the rest of the team does not get a wicket. And it would be silly to assign the value of 1 for 0 and find that 5-0-25-0 gets a better Spf-1 score than 20-10-30-1. So zero values are there and prevent both HM and GM to be used. Anhow I have given enough justification for AM to be used. Just because there is a way-out value does not mean that the AM is wrong. Pl look at Gilliespie's case which I had brought out in response to Arjun's comment. And I do not want to even think of replacing 0.00 with some number. 0.1 will completely distort the result since it is a product. Only 1 is acceptable since it is a product. And I am not going to give 5-0-25-0 a value of 1.00 which is the same given to 30-5-100-5 out of a total of 200 a.o. Ananth: ]] On further thinking I see that what I actually had in mind is not even geometric mean. It is simply the multiplication of the SPF ratios. This sounds wild in the first look and one could be very apprehensive about getting larger values. And also, one could wonder if it is justifiable to use the 'accumulated' value from different number of innings for different players. But, my hypothesis is that this will come out well. One is 2 times better in one innings and .25 times better in the next one, isn't it logical to conclude that he is (2*0.25 = 0.5) times better overall? [[ The zero values make this too a non-starter. Ananth: ]]

  • Yash Rungta on April 26, 2012, 4:16 GMT

    @Boll: I agree there have been many batsman who were better than Dravid/Ponting at No.3 and their averages reflect that. But very few of them have scored more than 5000 runs at that position, if at all. If you give some points for longevity, I'm sure Dravid/Ponting will come on top(except Bradman) since they've scored about 10000 runs each.

    Anyways, sorry to drift from the topic!!

  • Anshu N Jain on April 26, 2012, 3:22 GMT

    Isnt Harmonic Mean the reciprocal averaging method itself?

  • shrikanthk on April 26, 2012, 2:50 GMT

    beware the wrath of @shrikanthk for straying from the topic

    Boll: I am all for straying. I enjoy straying from the topic myself as you'd have doubtless noticed.

    On the previous post it was one of my pet peeves which made me object to some of the digressions - as a lot of people seemed too keen to bring in ODI stats to bolster the credentials of a particular player in a test-based article. That's all.

    Let's stray by all means.

  • Ananth on April 26, 2012, 1:39 GMT

    To all readers: We are going to be away from now (Thursday 7 AM) to Saturday evening. And for once I am not taking my laptop but only the iPad using which I can only browse or check mail but cannot do any work nor enter the blog publishing platform. So no more responses until Saturday evening. Unni: The GM is also a non-starter since the distribution will contain zero values. And one website suggests substitution of zero with some other considered value. But I do not have a substitution value. Zero suggests a very poor performance by the bowler. 1 suggests a very good performance. Substituting zero with 1 seems safest because this is a product but against the cricketing interpretation of the ratios. Ananth

  • Alex on April 25, 2012, 21:52 GMT

    @Gerry: When you cast off laziness, I request you to pl look up which 4 players these stats belong to:

    A. finals: ave=54, semis: ave=40, quarters: ave=66. B. finals: ave=38, semis: ave=37, quarters: ave=56. C: finals: ave=56, semis: ave=51, quarters: ave=43. D: finals: ave=41, semis: ave=29, quarters: ave=31.

    @Ananth & @Unni: For the geometric mean to work well, Ananth needs to re-define the factors as:

    SPF-1 = (bowler's ave in an innings)/(ave of _all_ bowlers in that innings);

    and likewise for SPF-2 & SPF-3. A simple illustration of why this should be so is available in the wiki article on "geometric mean". [[ Anyhow GM is also a non-starter. Ananth: ]] @Ananth: I find it commendable that you respond to a lot of comments and also code several suggestions & post those results. That is a lot of work even for a seasoned programmer! [[ Most are straight-forward. Some are tough and very few have been virtually impossible to do. For instance if you ask me what is the most favourite birthdate of Test players, I will give up. You could "Ask Steven" instead. Ananth: ]]

  • dinesh on April 25, 2012, 18:16 GMT

    @ananth: sorry for barging in and commenting about batsman here. I can understand the amount of work you have undertaken to bring out these stats for us and we guys indulge in our favorite pasttime SRT VS BCL VS RTP irrespective of the kind of articles done. I have been a following your blogs for a long time and i really appreciate the work you do and the kind of arguments and inputs readers like Alex,Gerry,Boll and others give you regarding improvements are too good. [[ No problems, Dinesh, as long as it does not go out of hand. And I get most of my ideas on new analyses from you readers. Ananth: ]]

  • dinesh on April 25, 2012, 17:59 GMT

    @Boll: I never compared his and dravid's records. I would say in My opinion Ponting was/is/will always be a far better batsman than Dravid. But when i looked up in Statsguru found some stats which you might want to look at.In England Dravid scored the same number of runs as Ponting but played 7fewer matches.But still ponting's better.In SA Dravid struggled.No doubts about that and ponting was far better. And coming to Australia its ponting's home.He is expected to score runs there,if not he then who will.So i would bring in India as well where on those very batting paradises ponting was made to look worse than Chris Martin(No one ca succeed him) for an entire series.And if you say those were batting paradises then probably in every match scorecard would have read 600/2. P.S: i never intended to compare Dravid and Ponting. My comment about Test batsman's analysis was about BCL and SRT. Ananth: Sorry for barging in and commenting about batsman in your awesome bowlers analysis.

  • unni on April 25, 2012, 14:19 GMT

    One comment from me on the methodology of combining the individual match figures. I think this is the place to use the geometric mean; why? Here the base value is 1 (identity of division/multiplication). i.e, in this analysis, if the player gets a perfect 1, he is equal to his team mates. If he gets more than 1 overall the bowler has done better than his team mates. If less than 1 then he performed worse than his team mates. So, situations where the value 1 should kind of 'penalize' the bowler. i.e, less than 1 value should pull down the overall figure. Arithmetic mean doesn't pull things down, it 'only' reduces the contribution. In other words, when the elements are ratios or products, it is better to use the geom. mean. When the elements are sums or differences use the arithmetic mean. I don't see any logical argument for using the harmonic mean in this situation. [[ Ah! nice to have you back with us, after a month of bliss, eating mangoes, reading and having never-anywhere fun. Today as I was writing my last comment on Harmonic Mean, I thought to myself, "what about the GM". However since we are leaving tomorrow morning for a 2/3 day break, I did not want to start it now. Will do it on Saturday. In fact this will be the perfect time to get a good handle on the different means and when to use what. Learnt that zero values preclude use of HM. Ananth: ]]

  • Boll on April 25, 2012, 13:44 GMT

    Argh, I just can`t help myself. At No.3, Dravid scored about 10,500 runs at 52.9 (SR: 42.8), Ponting has scored 9,900 runs at 56.3 (SR: 61.9) - I know who I`d be picking. And this is an article about test bowling spells! - I am a bad person.

  • Boll on April 25, 2012, 13:30 GMT

    @dinesh. Thanks for illustrating quite clearly (and promptly) why a Dravid vs Ponting comparison article would swiftly head down the road better not travelled. First someone brings up Ponting`s poor test record in India, then someone mentions that against the 2 best bowling teams of his Time (SA and Aus) Dravid could only score 4 centuries in more than 100 innings at an average of about 35. Then someone brings up Sydney 2008, and before we know it Ananth is struggling to maintain the integrity of this wonderful site. There are plenty of other places on cricinfo to sledge and demean and point out faults (and sure, I sometimes grimace when I look back over things I`ve written in the heat of the moment here) but we can at least try to rise about that stuff can`t we? [[ Thanks, Boll, for the kind words. As long as malice is absent I can take anything. And people should have confidence in me if I say it would be done. Remind me if I forget, that is all. Ananth: ]]

  • Boll on April 25, 2012, 13:10 GMT

    @Yash Rungta - beware the wrath of @shrikanthk for straying from the topic! I tend to agree with Ananth here though, that a Dravid vs Ponting analysis may quickly deteriorate into parochialism and put downs, rather than reveal too much we don`t know about either of these wonderful players.

    I have just looked back through the archives-, and had a look at Ananth`s Apr 2010 article entitled `Batsmen with highest averages at each position in Tests`. I would make just a couple of comments re. the top 10 players by average at that time at No.3 (Ponting at 7, Dravid at 9) - Dravid is quite unusual in that he appears at first glance to be clearly the least attacking of the top 10 (I would imagine his SR reflects this), and also that Ponting or Dravid as clearly 2nd/3rd best after Bradman is a very big call. Headley, Barrington and Hammond all average over 70 at No.3, Richards and Lara both over 60. Sanga doesn`t look too bad either.

  • Gerry_the_Merry on April 25, 2012, 12:47 GMT

    Ranga, Dinesh, will clarify as accurately as possible one last time:

    1) tough matches are defined as in and v/s E/A/S/W, but, in the figures I had quoted, I had mentioned no batsmen from these countries. I had compared Tendulkar only to his own career avg and to other batsmen from subcontinent.

    2) tough matches as in significant matches / knock-outs etc.: This was the basis of picking Ponting/Richards as being several notches higher.

    So there are two definitions I was using, and perhaps my descriptions ended up blurred.

    Tendulkar was a brilliant batsman, and I was a big admirer of his, till 1996. Thereafter, if you look at his record carefully, he started slipping up on both the above metrics.

    The one exception was Aus in 1999-2000, where against McGrath+Warne, he performed well, and apart from his skills, I would put it to his promoting team goals far above his own batting, as he was the captain. Abdicating captaincy cramped him mentally as a batsman under pressure.

  • dinesh on April 25, 2012, 12:13 GMT

    Ananth Since you are doing an an analysis on ODI batsmen's performance vs various countries and in different locations , i hope you do an analysis on TEST batsmen's performance vs various countries and in different locations . I think this will Open many eyes as well. [[ Yes, bowlers and batsmen in tests and odis. Ananth: ]] and @Gerry: if you say that Ponting was better than Sachin by comparing ODI's and runs scored in India,i would say how many runs did Ponting score in Tests in INDIA. It took him 5tours to score a ton and i think you would remember his famous runless streak in 2001 in india. [[ I think most of us are repeating the same words again and again. This applies not just to this comment but to all. Ananth: ]]

  • Ranga on April 25, 2012, 11:45 GMT

    @ Gerry - I Think you might have to add "Tough" for whom. If Eng was tough, then Eng would be world beaters 9/10 times. They arent. In 2007 Dinesh Karthik averaged the highest for an Indian in Eng series. Much ahead of SRT, RD, et al. How tough was it? Aus has always been fair and even. Aus produces good wkts, good contests. Aus has better cricketers.Most teams play well in AUs. But eventually Aus wins.Aussies win everywhere and hence, winning in Aus was always easier for them. And if India was the easiest condition to bat on, I dont think any result would have been possible for everyone would have scored 1000 for 6 or so. SL has always been tough for everyone, including India though it is a part of Subcontinent.

    I fully agree on your proposition on counting tough runs, I think it neednt be a general country based criteria. Facing McGRath/80s WIndies/Hadlee is tough anywhere, tougher in helpful conditions. So toughness is a relative term and not absolute. [[ Hopefully my next article, which will be published on 9 March, will shed a lot of light on this aspect. Ananth: ]]

  • Ananth on April 25, 2012, 11:21 GMT

    Arjun It slipped my mind that this could be the Harmonic Mean. I have used Arithmetic Mean and Geometric Men extensively. But rarely HM. However this has made it quite clear. The base definition as well as Wolfram Maths site clearly say that the numbers have to be greater than 0. There is no exception. In one reputed site, it is mentioned that HM is zero if one value is zero. In another site, they have identified a specific water-flow problem, requiring HM computation with a possibility of zero flow. There is a complex adjustment procedure. "" Because a zero flow cannot be used in the calculation of harmonic mean flow, the second term in the harmonic mean equation is an adjustment factor used to lower the harmonic mean to compensate for days on which the flow was zero. "" The bottomline is to give the HM a miss. Ananth

  • Gerry_the_Merry on April 25, 2012, 9:48 GMT

    Pranav, apologies for offending you. To clarify by my definition (you can disagree) 1) a meaningless ODI is one where the team's fortunes are not really at stake. In general tournaments, including World CupS, this is usually the semi final and the final, but occasionally includes other knockout-like matches (e.g. S.Waugh 120*) 2) In the World Cup '92, every match mattered to qualification, and all these should be counted, and indeed in any such tournament 3) I personally have stopped watching any bilateral ODI in India or abroad and consider all these meaningless.

    To popularize cricket thru WC, team list has been expanded resulting in many easy groups, diluting matches vs top teams.

    Take Richards (a) 138 in 157 balls in WC 1979 final. (b) take 149 in 99 balls in a meaningless ODI in India. Which is the better innings? Also, 100s in meaningless ODI is fine, AS LONG AS "tough matches", which i will ALWAYS define "in and v/s" Eng/Aus/SA/WI yield good avgs. You are free to disagree.

  • Pranav Joshi on April 25, 2012, 8:29 GMT

    Sorry Ananth, but Gerry's nauseating anti-Tendulkar rhetoric finally got to me.

    I will address just one point. He said "in meaningless ODIs Tendulkar will score 150 on any day against any opposition". I suggest Gerry look up Tendulkar's performance in ODI finals over his 23 years.

    When any other batsman scores a fluent century, it is a fascinating performance. When Sachin does the same, it is a "meaningless ODI". And by Gerry's metric most ODIs are meaningless. And few good batsmen should ever play ODIs at home, or on small grounds, because they are "meaningless". Even if this is taken as true, I would ask Gerry to look up Tendulkar in World Cups (as these aren't meaningless ODIs). And Indian grounds weren't nearly as small, or as flat, till about 7-8 years ago.

    I always wonder why turning tracks are flat pitches, and seaming tracks are tough pitches.

    And as someone said before, taking one element of a batsman's career, to conclude that he wasn't nearly as good, is very myopic. [[ Yes, I agree. You will see that my comments on SRT are only on the post-2/4/11 ODI career. I have no problems at all before that. The so called cheap 100s are there in every player's career. The fact that I publish a comment does not mean I endorse the views. It is also correct that another reader contests such comments, rather than myself. Anyhow my next article which is an analysis on ODI batsmen's performance vs various countries and in different locations will open many eyes. Ananth: ]]

  • Arjun on April 25, 2012, 8:29 GMT

    Ananth,

    A bowler with 9 performances of '2.00' and one perf. of '40.00' will average 5.80. With reciprocal method it will average '2.21'; which i think is more closer to '2.00'. Now big question is 'what to do with zeros' If we replace zero with 1 then 2+2+2+2+2+2+2+2+2+1 = '1.82' which is closer to 1.80(avg. of 9 2s and one zero). [[ I am not questioning the benefits, Arjun. What I have a problem is only the fact anything other than 0 is wrong. Why 1. SInce is a decimal number, why not 0.1 or 0.01. Morkel's 44 is in no way more outrageous than Gillespies's 201. Ananth: ]] Please have look at 'Harmonic mean' in wikipedia.

  • Vikram on April 25, 2012, 7:12 GMT

    While this overall analysis is awesome, it leads to a few more queries. Anderson has had two clear phases in his career, so where is Anderson 1.0 and Anderson 2.0. How was Kumble before and after his re-invention (adapting to foreign conditions)? Botham with and without Willis? Mcgrath with and without Warne and vice versa? Ananth, it might be fascinating if you took one particular era from each country where the team had a very powerful bowling unit, and then drilled down to see how each one contributed (in a peer comparison analysis) to a team's performance. I might be asking for too much :) [[ I think the Test bowling pairs is the first step towards such an analysis. That is "with"; "without" is quite dicey. Ananth: ]]

  • Vikram on April 25, 2012, 6:47 GMT

    @Alex: sorry mate, I was only talking about SMG post World Cup. Pre/World Cup is a horror story, apparently of his own creation to mar Kapil Dev. SMG was a little "different" in the '83-'87 era. This was the second phase of SMG where he was a little more attacking (a flip of SRT where SRT has curbed his natural strokeplay), with a little less technical control. Bat flying off vs Marshall, golden ducks vs WI in India, a century in ODI against NZ in 80 odd balls, dropping down the order, all are symptoms to his fight against Father Time. If the '83 trip to WI is anything to go by, '88 series would have been a disaster. However, I believe that he was better in '88 than in '83, mentally and in better form. So he would have done ok. While I am a huge SMG fan (how many cricketers do I like :)), I agree with you Alex in your assessment of his ranking in world cricket. The only rejoinder is that his contribution to Indian cricket goes way beyond his batting skills. Anyways, enough on SMG.

  • Gerry_the_Merry on April 25, 2012, 5:39 GMT

    Alex, that we disagree is a happy state of affairs for me. On Gavaskar, you are on shaky ground, as you are speculating that he would have failed. On Tendulkar, I am on firm ground as he has contributed nothing to prevent his team being beaten 8-0 except a fine 72 in MCG. Bringing One-days into the equation can be done within an era, and even there, as I exposed his stats a few days back, with dodgy numbers abroad, Tendulkar is not the most threatening batsman by any stretch of imagination. That has to be Ponting. On small Indian grounds, in meaningless ODIs, I would pitch for him to score 150 any day against any opposition. On whether Sunny was in top 3 / top 10/50 that is your opinion, though there also you have tried to distort the picture by mentioning that "till 1990, no one really thought Sunny to be even the best opener in his own era, let alone the best batsman". I have not bothered to counter such distortions, out of laziness.

  • Alex on April 25, 2012, 5:04 GMT

    @shrikanthk: Now that Ananth has thrown his hands up on Gupte, we have only capricious Sobers to rely on! As a child, I ought to have heard more from Mumbai press on Gupte since it always trumpets its own. I remember them extolling Merchant, Hazare, & Manjarekar but not Gupte. As a bowler, Shivalkar was their favorite. Maybe the circumstances that led to the ouster of Gupte (read his wiki page) did not go well with the prim & proper Mumbai establishment.

    @Gerry: We don't see eye to eye on many issues. Slander or not, I don't believe SMG would have done well vs WI in '88 & '89 ... look up his year-wise & opponent-wise stats '83-'87. And he was plumb LBW to Marshall on zero during the 188* on a flat Lords; Marshall clean bowled him for 0 (I think) in the 2nd. In the 1970-1987 era, SMG was arguably one of the 2-3 best test openers, one of the top 12 test batsmen, and a mediocre ODI batsman. Best not to compare him with SRT who is among top 2-3 in all 3 categories of his era.

  • Ananth on April 25, 2012, 4:55 GMT

    Arjun, I must have been really off to even consider the reciprocal method, with its attendant problems in handling zero values. What is a 0.0. That indicates that the bowler failed. Changing the zero to anything is wrong because it assigns an arbitrary undeserved value to the bowler performance. The problem is also that assigning 0.01 gives one set of player values, assigning 0.1 gives another set and assigning 1.0 gives a third set. This is totally wrong. What do we do for batsmen averages. How do we handle that. A 0 is a proper score and has to be taken as such. The 0 should lower the average and the 400 should increase the average. That is the way averages are determined. As far as Morkel is concerned, that was his day in the sun. He worked for it and performed brilliantly for 5 hours. Surely he deserves that credit. Simple example. Gillespie's top-5 innings are 47, 48*, 50*, 54* and 201*. Only a numbers-guy with no knowledge of cricket will say that the 201 should be excluded. Gillespie, with his 8-hour marathon, deserves his average to move up from 15.69 to 18.78. No one should grudge him that. So the bottom line is that the Spf-4 is an excellent measure. But we should stop at that. Neither should we use that for Player career analysis not should we use the insufficient Reciprocal method to modify a perfect calculation. By the by is there a Wikipedia reference on the Reciprocal calculations. Ananth

  • Yash Rungta on April 25, 2012, 4:22 GMT

    @Ananth(regarding my Dravid vs Ponting suggestion_: Then maybe you could do a special article for no.3 batsman. Of course, Bradman will steal the show but Dravid and Ponting should be the other 2 in the top 3. Both have about 10000 runs at that position.

  • shrikanthk on April 25, 2012, 3:18 GMT

    @shrikanthk: Not sure about Grimmett on googly. He is reputed to have invented the flipper

    Yes. He did invent the flipper. But from what I've read about him, his leg-break did not turn a LOT. He was not as quick as O'Reilly but still a fairly brisk spinner with a round-arm action (Even more round arm than Warne).

    Nevertheless all this still points to a bowler with a conventional line of attack (i.e attack the off-stump or just outside, from where the ball can either turn away, come back or go straight on). Among modern leggies, the likes of Amit Mishra and Piyush Chawla have similar lines of attack (they seldom bowl a leg-stump line).

    Maybe Ananth can put more light on Gupte and Benaud. I think Benaud was a conventional leggie with a middle-and-off stump line. Not sure about Gupte. [[ Poor Ananth had the privilege of watching two Test matches befiore he was 25 and saved to buy a second hand radio at Rs.25, while at college, to listen to Vizzy, Berry Sarbadihikari, Merchant, Chakrapani, DeMello et al. So he can only tell from his newspaper/Sports&Pastime memories that Gupte had a high arm action and used googly as a surprise weapon, as hopefully Benaud did. Certainly I remember the radio commentators exclaiming, "It was a googly, not read by the batsman", but no more than once an over. Ananth: ]] There were spinners before Warne who turned the ball a mile. But none had his extraordinary drift which enabled him to target the batsman's blind spot despite bowling "over the wicket". Genius!

  • Gerry_the_Merry on April 25, 2012, 3:15 GMT

    Shrikanthk, Warne gave the ball a tremendous spin. With that fizz, he got significant in-swerve when the ball entered the second half of the loop, which again it was in Warne's case that I saw first. So repeatedly the ball would start on the stumps and fall in the batsman's blind spot, which also coincided with the rough marks, causing Warne to be a terror to particularly to Englishmen and Safs. Lara did not have this blind spot at all being a leftie, which probably explains to some extent Warne's OK record against him. Warne was also unorthodox in the sense of inventing so many new deliveries, which no other spinner in modern cricket did.

  • Gerry_the_Merry on April 25, 2012, 3:07 GMT

    Alex, your comment that "SMG would have surely failed vs West Indies" is a facetious remark. I can perfectly understand if he failed outright, based on which you predict a failure for the future. But if at the age of 34 he scored 3 centuries against WI and retired at 38, the first has no relevance for the second. At 38, if you mean that he was failing, that is not what his farewell series and World XI 188* indicated. If you mean that at 38, batsmen become senile, that too is not true. Greenidge scored 226 at 40yrs, Gooch scored heavily until 41, Boycott made a century against West Indies at 41, Lloyd played brilliantly until 41. All these batsman Lloyd, Gooch, Gavaskar were champion fighters, unlikely to be passive bystanders if their team was being whitewashed. You are probably referring to Tendulkar's RPI being 28 at this age. It has nothing to do with age. It is the pressure mounted by the opposition. In that context, your comment on SMG amounts to just a slander.

  • Ananth on April 25, 2012, 1:33 GMT

    Arjun While I am happy that the Spf-4 values have come out very nicely, I am not happy when I apply these values to the players. Strange things happen. Using the similar criteria for spells, Steve Waugh and Jayasuriya are ahead of Murali. They both do nt have a single Spf-4 value greater than 2. I think it is the peculiar nature of comprison and the fact that average splls end with scores around 1. I think I will do the reciprocal method for the standard Spf-1 value. I will take 0 as 0.1, not as 1. That way the impact will be minimal. Ananth

  • Alex on April 24, 2012, 18:26 GMT

    @shrikanthk: Not sure about Grimmett on googly. He is reputed to have invented the flipper and Wiki says that he bowled the flipper as often as the leg-break. Warne might not have bowled googly a lot but he still had tremendous variety: gargantuan leg-break, flipper (his signature until the shoulder injury), & top-spinner. Because of his pace, Chandra was really unorthodox. Kumble too was unorthodox in the sense that here was a spinner who hardly spun the ball!

  • Alex on April 24, 2012, 18:10 GMT

    @Vikram: It was not 4 out of 7 innings ... SMG played 11 tests vs WI, 1982-83, and, more specifically, 20 innings. He was excellent in only 4 of those (147*, 121, 90, 236*) and an outright failure in the remaining 16. Of course, he did well in 4 out of 11 tests ... and _that_ is not that bad.

    @shrikanthk: "Classical" was my euphemism for clean action.

    @Ranga & @Vikram: The complication with Bhishma was due to (i) he was in the war not to win it but to wear it out to save his kins, and (ii) he wanted the honor of dying at the hands of only a truly great warrior whereas no such warrior wanted to pull the plug on him. On the 9th night, he met with Yudhishthira & Krishna and accepted the futility of (i). They conceded (ii) to him by placing Arjun behind Shikhandi & creating a smoke-screen that Arjun shot arrows and not Shikhandi. So, I think SRT needs to accept the futility of him playing on and, maybe, some good team has to let him score a 300 or 400 so that he retires in peace.

  • shrikanthk on April 24, 2012, 16:42 GMT

    All leggies are inherently unorthodox

    Good point. The art of leg-break bowling really took off around 1900-1910 when Bosanquet pioneered the googly and the SA bowlers mastered it. At that point, people used to seldom refer to someone as a "leg-spinner". The common term was "googly bowler", which suggests that the wrist spinner of that era was someone who attacked the off-stump and made the ball turn both ways.

    Even in the 30s the Aussie leg-spinners used the "googly" as the attacking weapon. The leg-stump line was used mainly to tie down batsmen (Eg : O'Reilly tying down Hammond or Grimmett tying down Headley)

    Those pioneers epitomised "classic" leg-break bowling which was basically googly bowling mixed with the odd innocuous leg-break.

    Warne changed this fundamentally. At last here was a leg spinner who hardly bowled a googly! He showed the world that a leg-spinner can actually get wickets by bowling HUGE legbreaks pitching regularly outside leg-stump! That was his genius.

  • tashfeen on April 24, 2012, 16:32 GMT

    Ananth, this is the third time I am reading your article. "Brilliant" would be a severe understatement - hats off to you sir!

  • Arjun on April 24, 2012, 15:02 GMT

    Ananth,

    I hadn't checked mails so wasn't aware of this. But still it can be done. eg. Your original method 0+2+4+35+2 = 43/5 = '8.60'. Reciprocal method can't be done because of zero values. However, if we replace 0 with 1 then result is '2.19' fair enough. A lot more acceptable than 8.60 [[ I think the need for the reciprocal method disappears now when I use the new index which runs from 0.00 to 8.06. This is a normal and acceptable distribution and the impact with high numbers is not there. Morkel has become 5.29 so he is not going to get the huge push. I don't want to arbitrarily replace the 0 with 1 when I do not know of the impact. 0 represents a really awful performance. 1 represents a quality performance indicating that the bowler included figure is the same as bowler excluded figure. That sort of change cannot and should not be made. Beer's 2 for 56 has a Spf-4 value of 0.96. There are a few 0 for 100s which have values nearing 0. And using the reciprocal method must have overriding benefits. Ananth: ]]

  • Tom on April 24, 2012, 13:33 GMT

    Fantastic stuff as usual Ananth. I can't quite make it out due to the location but am I right in seeing Warne's 02 next to Harbajhan's 10 on the graph? If so it's interesting to see that Warne, McGrath, Lee and Gillespie all have very similar BpW peer scores with RpO making the difference. Nice to see some sensational one-off performances by batting specialists at the top of tables 1 and 2, too! [[ Sorry for the cluster there. Let me give you the numbers. Warne: 1.12 & 1.12 Harbhajan:1.12 & 1.13 Ananth: ]] Hoggard and Anderson are also interesting cases as their career figures are very similar (although on your consistency measures Hoggard has a much higher mid-3rd slice %). I suspect their position on the graph here is partly because they have been used in similar ways by captains for much of their careers - as effective second/third "reliable" seamers (although they have both often opened the bowling) behind the more intimidating pacemen like Flintoff, Harmison, Finn, Broad, etc.

  • Arjun on April 24, 2012, 12:44 GMT

    Ananth,

    thanks for the table. You are right, great spells have come on top.

    Have you changed the career table with revised reciprocal method ? [[ Arjun, I have already mentioned twice, including a direct mail to you, that this cannot be done because of zero values. Ananth: ]] Also what about 2nd career table with this new peer ratio ? [[ I have mentioned in my comment that this will be done soon, probably by morning. Ananth: ]]

  • Ananth on April 24, 2012, 12:27 GMT

    Arjun: I have incorporated the fourth Spf value: The ratio between the Team Average without the concerned bowler and the Team Average with the concerned bowler. I had some reservations because this is not a true peer comparison. But the results have proved otherwise. It is a fantastic comparison. First the highest value is 8.06, very well under control. Lohmann for his 9 for 28. Next is Muralitharan (7.59) for his 9 for 51. Third is Laker (7.20) for his 10 for 53. Second is that only top spells are in the top. Freak spells have little chance of getting in. Morkel, who had 44+ in Spf-1, has 5.29 here. Great, but not outrageous. I have incorporated the worksheet in the Spf tables and uploaded the same. I will now incorporate this in the Player data and do the career data again. Should be very good. The link is given below. http://dl.dropbox.com/u/39210851/Bowler_Spell_Peer_Index.xls Ananth

  • Yash Rungta on April 24, 2012, 11:45 GMT

    Nice article Ananth! :)

    Although I've got a suggestion for a new write-up. How about a 'It Figures' blog on 'Dravid vs Ponting' in Test Matches. On the face, they have very identical batting stats and both batsmen played majority of their career at no.3. But if you go deeper, you realise Ponting has played 89 matches at home out his 165 and Dravid has played only 70 matches at home out of his 164 which suggests that Dravid has been a slightly better batsman than Ponting since they have a similar average. Ponting was probably a better fielder but catching wise both were equal or probably Dravid was a little better. Anyways, is there any chance of a write-up on their Test Batting Analysis? [[ I don't like these types of analysis. They have a fixed agenda. Instead, a comparison of the top 10 players in all aspects will be worthwhile. Ananth: ]]

  • Ranga on April 24, 2012, 10:38 GMT

    I think it is extremely difficult to get a highly objective analysis out of the numbers as all of us come here because of subjectivity, or our qualitative preference for the game. To an extent, these blogs try to bring out various dimensions and substantiate in numbers. May be the advent of TV and the involuntary diet of the partisan commentary (here I dont see any nationality - Greig/Hussain as partisan as Taylor/Chappell as partisan as Ravi/Sunny - May be there are a few exceptions). The point is many times we get clouded by this w/o our knowledge and look for OUR preferences in numbers. Typical example is Ishant's "excellent" launch in Oz (6 wkts in 3 tests).Initially I too had trouble accepting the numbers here but I remembered the days in which to watch 3am matches from NZ, I used to watch on Mute. Without influence of someone else's opinion.I started reading these blogs from that angle. Just my own interpretation of what is in front of me.Like a match on mute.My own judgement

  • Gerry_the_Merry on April 24, 2012, 9:49 GMT

    Shrikanthk - who exactly is an example of a classical leggie according to you? especially in post 1970 period? Dont tell me Chandra, Qadir and McGill. All leggies are inherently unorthodox, and you just have to move the normal to the right. Off spinners are generally called orthodox (not like Ajmal, however), which must mean that the terminology evolved to accommodate the perception that leggies are unorthodox anyway.[[ To my knowledge Subash Gupte and Benaud come close. Ananth: ]]

  • Vikram on April 24, 2012, 8:12 GMT

    @Ananth, don't get me wrong. Arjun's request has a lot of value. I am just trying to see if the 2.39 there means Clarke as compared to others or team with Clarke vs. Team without Clarke and hence what should the interpretation be. If it's the latter, then is it peer comparison or team effectiveness comparison. That is what my question was. And I would definitely be looking out for it, so hopefully you have added it to your to-do list. :)

  • Vikram on April 24, 2012, 7:34 GMT

    @Ranga @Ananth: yes this is the cleanest of comparisons. In teh case of batting analysis by excluding pitch and bowler, it's a fake (but valid) situation being created. This is the best sort of analysis that can be done. And the relative analysis is fascinating. Murali vs. Vaas, how Waqar and Akram complimented each other, what does the fact that Mcgrath, Warne and Gillespie are all clustered mean for the team and for the opposition - just fascinating. And as a player, your biggest objective is to play for your team, so your value to the team will (should) mean the most to the players. @ Arjun: taking your example, Clarke was 35+ times better than his peers, however the team was 2.39 times more effective with Clarke in the team. Isn't that what the interpretation should be? [[ I see value in Arjun's suggestion. But is not a pure peer comparison as defined in this artcle. But that has never stopped us from exploring. Ananth: ]]

  • Arjun on April 24, 2012, 7:00 GMT

    Ananth,

    By mistake i mix-up 2 different methods. What actually new table should include is Ratio of with and without. eg. Test # 1720, M Clarke took 6 wkts for 9 runs. Avg. including clarke is (20.10) 201/10 Avg. excluding clarke is (48.00) 192/4 so the ratio is 48/20.1 = '2.39' This ratio indicates he bowled 2.39 times better than his peers. [[ Arjun This is a play of the numbers. The real peer comparison is 9/6 (1.5) with 47.00 (188/4). That comes to 35+. Your suggestion is an alternate measure but, as Som pointed out it compares two scenarios, both including Clarke. I will still do it since it provieds an alternative with manageable ratios. Ansd I had mentioned about the problems of reciprocal method, with zero values. Ananth: ]]

  • Vikram on April 24, 2012, 6:47 GMT

    @Alex: exactly my point mate. In India, where memories are short-lived, SMG ensured that his last series was a good one. If he had persisted, like Kapil did, in going for a couple of more tours, it would have been a disaster for him. Here he went with one of his best innings as his abiding memory. As for those 4 innings against WI, that can not be called a bad performance. He played well in 4 out of 7 innings, those 4 innings came over 3 test matches. WI smashed India in the other 3. Also, (something which shrikanthk may again not like), he did lead India to the B&H World Championship, and retired from capataincy after that, very shrewdly but also at the same time built himself a favorable image. So overall, I think he timed his retirement perfectly. As for SRT vs Bhishma example, who will be Shikhandi? Maybe Amarnath, if he becomes the chief of selectors? [[ And Srikkanth then must be the vidushak. Ananth: ]]

  • Ranga on April 24, 2012, 6:30 GMT

    (cotd): To me, I read the entire series of articles (whether they may be connected or not), as "What did one do with the oppportunities presented" rather than "who had more opportunities"

    It doesnt matter that BCL had more 2nd inns than someone else because Windies folded quickly or Sir Hadlee had more chances of getting a wkt than others coz of weak bowling team. But what they did with the opportunities presented is what needs to be enjoyed and I feel that aspect comes perfectly with this article. [[ I love these peer level analysis articles. As you have mentioned, this is the purest of peer analysis. Everything is the same, including fileding/catching support which I forgot to mention in the article. I remember when I did the peer analysis across teams, there were questions raised on the catching support McGrath/Warne got. I cannot get over the match between Saf and Nzl. Two of the most attacking bowlers in the world did nothing, while bowling 33 overs, other than bowling economically while the third bowler gets a wicket every 16 balls. In fact the lack of success of Philander (for once) and Steyn also cost the South Africans the match and the no.1 position. And in the first innings, Steyn and Philander capture 8 wickets at a wicket every 34 balls and Morkel bowls 120 balls without a wicket. That is Test cricket. 6x6s and 7x4s in 45 balls, as happened yesterday, is no compensation.The myriad nuances of Test cricket !!! Ananth: ]]

  • Ranga on April 24, 2012, 6:25 GMT

    I read this blog piece again and again at least for 2-3 times to digest the facts. Actually this is one article where a very clear cut peer is defined. Contribution to one's team is more prominent than performance vis-a-vis others playing for other teams. A revelation to the generation that watched live cricket from 80's is the presence of Chandra as a strike bowler! In the famed quartet (Venkat & Pras may not qualify due to cut off on wkts),Chandra is seen as most attacking. He is in the league of fast bowlers, showing how much he attacked. Going by all this, I get a feeling that "Playing for a weak team", is not a great advantage at all!! It is a disadvantage rather!!! I get surprised when Lara or Hadlee are credited for playing for weak batting and bowling teams respectively! (Cotd)

  • Ranga on April 24, 2012, 6:10 GMT

    @ Alex: Thanks for the wonderful reference to Bheeshma! Nice to have someone recollect a legend when thinking of these two legends. Also the most intriguing part of Bhishma's death was he himself, for the benefit of those innocents who were killed because of his invincibility, had to tell Yudhishtra the secret of his death!!

    That leaves us with SRT himself!! The grand old man of Indian cricket, fighting a war against himself and coming up with Pyrrhic victories losing all the credibility he had earned over 22 years of career characterised by his cricket alone which defined his personality. Like Bhishma, he has to put an end to the miseries by taking a call on where he wants to go and decide without miusing the trust the billion(s) have on him. This could be anything - be it playing or retiring - it is upto him. He still has good cricket left in him, but then, can he play better roles in Indian cricket than what he is doing currently? And is he stopping himself from doing so?

  • shrikanthk on April 24, 2012, 5:31 GMT

    I do prefer a classical spinner like Warne/O'Reilly over him though

    Not sure what you meant by classical. Bill O'Reilly was anything but classical. He was a medium-paced bowler (essentially Kumble like but with more turn). Truly a unique bowler in cricket history.

    Even Warne was not strictly classical. By classical I mean a leg spinner with a high-arm action and good flight who attacks the off-stump and moves the ball away. Warne changed leg-spin fashion fundamentally by having a slightly round-arm action (atleast not a perpendicular arm!) and an aggressive leg-stump line which used to be regarded as a negative line in the past. He was very much a revolutionary and more of a heretic than a classicist.

  • Alex on April 24, 2012, 2:53 GMT

    @Vikram: SMG's well-timed retirement is a myth, IMO. He really struggled throughout 1983-1987 in tests --- during these 4 years, he could score big vs minnows only (Oz were minnows that time), the only bright spots being a good farewell series vs Pak, 4 good innings vs WI, and two 50's vs Eng. That's all. The next series up was vs WI, and no doubt he would have failed in it. (However, he became a competent opener during 1986-87 in ODIs).

    @Ananth: Sadly, SRT has been averaging even less than 25 RpI for a while now. His new bizarre hair-cut recalls Kapil's own weird hairstyle circa '93-'94 ... Kapil sported a pony-tail back then and the crowd used to heckle rowdily at it. [[ Yes, the RpI for the last 11 matches is 28. Ananth: ]] On retirement, Kapil or SRT, esp. SRT, make me think of Bhishma of Mahabharata. Everyone knew that Bhishma had to be killed but no one who mattered wanted to pull the plug on him. A tame war went on aimlessly for 9 days and then on the 10th day somehow Bhishma fell and life resumed.

  • Raghav Bihani on April 24, 2012, 2:24 GMT

    Since this is a peer analysis Murali, Hadlee, Kumble etc are expected to dominate. What is interesting is to see how teammates are placed:

    1. Wasim and Waqar: they complement each other brilliantly as can be seen by their opposite stats. each excellent on one measure. This is despite other very good bowlers in the team.

    2. AUS team of 90-2000s: Interesting to see McGrath, Warne and Gillespie all in the same quadrant. None of them could pull away on any measure beyond 1.2. How Gillespie stays afloat is tribute to his value?

    3. WI of 80s: Marshall and Holding are excellent in BPW., while Garner was a mean machine. Each of them prevents the other from entering the top quadrant. They complement each other well. Ambrose in top quadrant despite playing alongside Walsh, Marshall etc - Top Notch or is he there only because of his latter years.

    Plotting of other players like McGill, Roberts, Bishop, Saqlain, Pollock would make for interesting team analysis. [[ I am already plotting 41 bowlers. I could just about add a few. Let me see. Shri has asked for O'Reilly. Ananth: ]]

  • Vikram on April 24, 2012, 2:19 GMT

    @Alex: If you look at the stats for the 3rd last year, 7 tests and 25 wickets, it shows you that his last two years would have pulled him down in this analysis (peer based analysis, where in the last 2 years other bowlers would have picked up more wickets). Also, his RPO has gone up in the last two years, so the 3rd last year again would have a much better RPO. It is about timing. SMG did that. Kapil had a chance after the AU trip where he bowled AB through the gate twice, if I recall correctly. The last two years marred his legacy. He retired statistically the highest-wicket taking bowler, but with a reduced image/memory and maybe a lower ranking here, same as SRT (just after world cup would have made him an all time legend with the abiding memory of him being caried around, now he is being questioned).

  • Vikram on April 24, 2012, 2:06 GMT

    Ananth: a request - while the graphical representation can be restricted to the top 10-12 pairs, the analysis should be for atleast a few pairs. I would say that we should take the top 50 pairs, and then off those the top 10-12 pairs are represented graphically. Or am I being greedy? [[ Vikram, that is how it has always been. As you might have noticed I follow a three-tier structure almost always. Graphs: Selected group of 10-12. Tables in the article: around 25-35. Downloadable Excel sheets: No limit. However when we come to bowler groups we ourselves must set the bar high not to dilute the analysis. I would say, only the first two should apply. If McGrath/Warne captured 1000 wkts together (just a number), the last pair should not be below 250. Incidentally I had done an analysis of the bowling pairs four years back, link given below. Now I would do a different type of analysis, with more peer comparisons. http://blogs.espncricinfo.com/itfigures/archives/2008/09/good_in_isolation_great_as_a_p.php Ananth: ]]

  • Vikram on April 24, 2012, 1:34 GMT

    A bowling partnership based analysis would be really useful. Even within these 41 bowlers, a high % bowled in partnership. Warne-Mcgrath, Ntini-Kallis, Kumble-Harbhajan, Harbhajan-Zaheer, Akram-Younis, Ambrose-Walsh, Botham-Willis, Lee-Warne, Marshall-Holding, Holding-Garner, Vaas-Murali, Donald-Pollock, Pollock-Ntini, Mcgrath-Gillespie, Gillespie-Warne, Gillespie-Lee, Bedi-Chandrashekar. Looking back at this graph, if we ignore a few outliers (Lee, Kallis, Hadlee, Murali, Barnes), all the other bowlers seem to lie clustered around a straight line (x,y = 1.4,1.7). Finally, it would be rather comforting for most batsmen to see that very few active bowlers appear in any of the SPF based tables, a further prove, if any was needed, that bowling levels have declined. [[ We must set rather high standards for bowler pairs. No more than a dozen pairs since graphical depiction then becomes difficult. I am now struggling with this problem in a big way in the ODI Batsman-vs-countries tables/graphs. Ananth: ]]

  • Proud Indian on April 23, 2012, 20:47 GMT

    It is hard to believe that by any analysis bowlers like Wasim Akram, Warne, Macgrath, micheal holding, marshall will miss out in list.... [[ I give up when I see a comment like this. That too after requesting in the article not to post comments like this. Ananth: ]] Not trying to deny anything here...it's like you do batsman analysis and don't see Sachin!... [[ An analysis of 250+ scores in Test cricket will certainly exclude Tendulkar. Ananth: ]]

  • Alex on April 23, 2012, 18:50 GMT

    @Vikram and @Ananth: It is a myth that Kapil the bowler was disastrous in his final 2 or 3 yrs (or final 19 tests). His SR had diminished but he remained very economical and an effective bowler. Indeed, his stats during final 3 yrs:

    19 tests, 49 wkts, ave= 28, SR=83, RPO=2.1.

    His stats during the final 2 yrs:

    12 tests, 24 wkts, ave=28, SR=75, RPO=2.2.

    Statistically, Kapil's real horror phase was 1986-1990, which was salvaged by 2 good years of 1988 & 1989. His stats in that phase:

    37 tests, 96 wkts, ave=34, SR=80, RPO=2.6.

    A reason for this sub-par performance was that India played on batsman-friendly wkts in '86, '87, and '90. Not surprisingly, the '86-'87 phase was SMG's bright period (culminating in a well-timed retirement) and DBV's purple patch. [[ At the end of 115th Test Kapil had 401 wkts at 29.66. At the end of 131st Test Kapil had 434 wkts at 29.65. So there is absolutely no average change. He captured 33 wkts in 15 Tests at 29.60. I think this average of 2.2 WpT is a huge drop from the 3.48 when he reached 400. You could think differently. Currently Tendulkar's ODI RpI is a rather high 40.76 (4th best ever). It is like accepting a RpI of 25 over the next 25 matches. Ananth: ]]

  • Som on April 23, 2012, 17:37 GMT

    @Arjun - Without/With cannot give 'times better', I think. Ananth's approach of Without/Only, IMO would give 'times better'. [[ Without/With is not a true peer comparison. It is also a different type of analysis. Also the BowQ ciomparison will give the same ratio whether Morkel had a spell of 16.4-7-23-6 or 16.4-0-100-0. Ananth: ]]

  • Arjun on April 23, 2012, 13:16 GMT

    Ananth,

    There are 3 different SPF tables...average-based, RPO-based and SR-based. There is need for New 4th table with raw unadjusted figure. The formula you use for BowQ, ER-SR method (bowling quality for an innings). This table should provide the ratio with a particular spell and without that particular spell. Say in the example of morkel, his team's BowQ inculding him was 32.47 and without him it was 54.29. The ratio is 54.29/32.47 = 1.67. It basically means he was 1.67 time better than his teammates in that particular innings. Another benefit of this table is you do not have to use artificial wkts taken for other bowlers as 1 if they have taken 0 wkts. I think this peer table will be very interesting. [[ A new peer idea. I will do and add to the same Excel table as a fourth worksheet. And the numbers will have a finite range. Continuing Arjun: why should be done for these qualifying spells only. What you are asking is a much more general analysis which can be done for ALL spells. By restricting it to these spells we will miss some important ones since I am also working wirth some cut-offs. Ananth: ]]

  • sumanth on April 23, 2012, 12:08 GMT

    First of all, thank you very much for giving one more dimension in analyzing the Test bowling.

    This statistics may not be liked by Warne&Mcgrath, Wasim&Waqar [as they got good records but unfortunately they had to play together mostly in their career] and for some extent to Ambrose [as he played with walsh]. Kumble gained a lot because there were no world class bowlers in his tenure [Srinath may be an exception for some extent].

    If we derive the same statistics with pair against their peers, I am sure Waqar&Wasim along with Warne&Mcgrath and Ambrose & Walsh will top the list. What do you say? [[ We will look at, say, 10 top bowler-pairs and analyse everything including the peer analysis compared to the rest of the teams. Let us wait for that analysis. Ananth: ]]

  • Vikram on April 23, 2012, 10:35 GMT

    Very very interesting analysis. There are a couple of bowlers who bowled in the same team and it is really interesting to compare them. Also, the charge against Harbhajan that he was never the team leader seems to be correct. Apart from that 2001 series, he has never really led the team. Two of his contemporaries, Zaheer and Kumble, have had to take the lion share of that load. Also, Kallis seems to be an outlier but for a batting allrounder to be even part of this chart, shows his value to the team. As an overall contribution to the team, he might outrank everyone. An interesting comparison between Imran, Botham, Kapil and Hadlee in their relative importance to the team. Alex, Donald was better than Mcgrath :) He could be aggressive (Atherton), could be outlandish (SRT bowled through the gate, what a sight) and could be the channel bowler. A package. Finall, Kapil Dev - just outside the elite quadrant. If not for the last 2 years, he might have left a better image on cricket followers. [[ Yes, I am surprised that, despite those disastrous last 16 Tests when he added 32 wickets, he is still knocking on the first quadrant sandwiched between McKenzie and Lillee. Ananth: ]]

  • Anshu N Jain on April 23, 2012, 9:28 GMT

    [What justification do I have...]

    The justification you have is that the bowler's peers couldnt take even a single wicket in that innings! Anyway, the outperformance factor that you derive is not a cricketing measure as such either, so why not the tweak?

    I understand the problem of having a legit divisor. I also understand that using 0.1 or 0.5 or any such number between 0 and 1 will result in an SPF of magnitude several times the base case of treating 0 as 1. However, adding 1 to the denominator in both cases does not have this magnified impact, while still giving due weightage to a bowler whose peers took no wickets at all, vs those whose peers managed some wickets, albeit just one.

    As it is, I'm guessing there wouldnt be too many such instances all put together, leave alone in a single bowler's career.

    I'll be happy with whatever you decide to do finally and wont come back with another rejoinder on your response to this :-)

    Will go back to the detailed tables now!

  • Gerry_the_Merry on April 23, 2012, 9:13 GMT

    Ananth, why not do the career PF on averages like this? Say the matches are numbered 1, 2, 3 etc. A runs / B wickets is the average of the other bowlers in match 1. Over the bowlers career it becomes sigma (A1, A2, A3...) / sigma (B1, B2, B3...). Let us call this OBCA (other bowlers' career average). Then the CPF is career bowling average / OBCA. [[ Then it will not be an Innings peer analysis. What would it signify. Ananth: ]]

  • chris on April 23, 2012, 7:55 GMT

    Really enjoyed this one, interesting to see another correlation between Hadlee and Murali isn't it?

  • Pankaj Joshi on April 23, 2012, 7:51 GMT

    Beautiful premise and, on first cut, logic seems quite sound - these are two different things. Congrats Ananth. Some meaningful perspective would help - for example Hadlee's not heavily present in the first two tables. Does that mean the Kiwi attack was not such a one-trick pony or that Hadlee simply bowled more overs than his mates to get more wickets? It may be difficult for packs of bowlers to figure here, but surely Hadlee or Kapil would (9-93, 8-85, 8-106, 7-56) figure a bit more, like Murali? Second, another freak thought now that we're in the zone. What about an exercise within the top spells where a bowler runs through the top half and makes it easier for his mates to shine? That may deserve a few more credits, if there's a base to link wickets taken and their order in the batting. [[ A different exercise altogether. Ananth: ]] That muddy lane near the Wankhede - getting better and better ain't we? Rgds [[ I couln't resist that since there was a possibility that we could have defended 75, leave alone 105 Ananth: ]]

  • Arjun on April 23, 2012, 7:36 GMT

    Ananth,

    In the career analysis table there are few surprises like morne morkel etc. not because they were that good than peers but because of wrong method of calculation. what you have done is summed all the ratios across career and divide it by no. of spells. Morkel benefits because of his single spell ratio of '44.61'. The correct method is reciprocal method. I am sure steyn(1.48) would be and should be above morkel(1.66) [[ Arjun, to be honest I clean forgot that the recprocal method can be used for all averaging. I was under the impression that it was useful only in the weighted averages scene. Let us say 1, 2, 3, 4, 50 would lead to an average of 12.0. Is my reciprocal method given below correct. 5/(1.0+0.5+0.333+0.25+0.02) This works out to 2.37. However see the problem in front. Every table changes. Every description para changes. The graph changes. Let me see what can be done. On second thought not as big a problem since only the career table will be changed and the graph. Ananth: ]]

  • Mystic on April 23, 2012, 6:38 GMT

    ummmm... where did you get your doctorate from?

    brilliant!!! [[ From the readers of this blogspace. Ananth: ]]

  • Alex on April 23, 2012, 5:59 GMT

    @Ananth: Thanks. The xls sheet is most absorbing. Note the similarities in Steyn & Donald, Imran & Fazal, Gillespie & Miller, etc. A real surprise - Gupte & Hadlee!! Sobers rates Gupte the best leg spinner he ever saw. Gupte's career figures are superb but who did expect him to measure up to Hadlee in this analysis? [[ A very under-rated performer. His spell at Kanpur against West indies was one of the best ever on the first day of the match. Ananth: ]] Murali was extraordinarily skilled but, like Shoaib, had a slightly suspect action at times, IMO. This is not a place to discuss that and, now that this great icon has retired, it is best forgotten. I do prefer a classical spinner like Warne/O'Reilly over him though.

  • Balumekka on April 23, 2012, 5:42 GMT

    Ananth - Brilliant work! Love your opinion on the 'no average'/ average infinity) issue! Performance analysis compared to peers is very sensible. However, I was wandering whether it would be more meaningful if you can do a correction for the final figure considering the quality of the wickets captured(based on the batting average of the batsman dismissed, may be). [[ Balu, if this door is opened a bit it will get opened completely. Let us leave the concept as it is. A peer analysis with no other factors coming in. Ananth: ]]

  • Som on April 23, 2012, 5:30 GMT

    Absence of Mcgrath and Warne in any of these tables is quite a revelation (and actually shouldn't be). They perhaps with some effort from Gillespie, Lee foiled each others chances of making out-of-peer group contributions. But OZ definitely did not mind. An interesting way to look at this data would be to group Mcgrath and Warne together and run these same numbers against others. That way, when done for all teams across era's would give us the best hunters in 'pairs'. The same theory applies to Macko and to a lesser degree to Wasim.

    I had done some analysis on my end when selecting India's all time best test XI, and had found Gupte to be making a compelling argument (only to be losing out marginally to the comprehensive figures of Kumble).

    Nadkarni was a champ! Never seen another like him. [[ Som, I have a reference in my Pending Box, "Two bowlers together". I will do an analysis on that and as part of that I can consider the two as Siamese twins and do the peer analysis also. At a later stage I will ask for possible pairs. Ananth: ]]

  • Ananth on April 23, 2012, 4:12 GMT

    The Career Excel table has now been uploaded with the Career Spf-2 and Spf-3 information. You could get the information for all bowlers, for Spf-1, Spf-2 and Spf-3, here. the graph will not be updated. The link is again provided below. http://dl.dropbox.com/u/39210851/Bowler_Spell_Peer_Index_CareerSummary.xls Ananth

  • Alex on April 23, 2012, 2:42 GMT

    @Ananth:

    1. What would be the numbers for Roberts and Fazal Mahmood? I think Thommo's numbers would be similar to Lee's and Shoaib's would be just a bit better.

    2. Sobers missed your cut-off of 240 wkts by just 5. What are his numbers It'd be nice to see where he is placed w.r.t. Kallis. [[ Instead of answering individual queries, I will upload another table which contains all three Spf values so that you can see for yourself. Will be done by today. Ananth: ]] 3. Gibbs had an extraordinary RPO=1.97. I always feel he belongs in the all-time WI XI, if for nothing then just for the sake of variety and to ensure adequate over-rate. Pl give an honorary mention in the excel sheet to Nadkarni (RPO=1.67!!).

  • milpand on April 22, 2012, 23:37 GMT

    I liked 'muddy lane' in "This match was incidentally played on a muddy lane around Wankhede Stadium." very much.

    But not a big fan of the repeated use of 'cheese & chalk' both by you and others. It does not irk me too much in this post but their are instances in recent past where a better alternative could have been used.

    I think we should use 'chalk and cheese' to indicate that apart from a superficial similarity, nothing else is same. I wonder whether Nadkarni and Maninder are that dissimilar? [[ I think the one major difference was that Nadkarni's obsession was to bowl run-less balls and overs. Maninder was more attacking. But two varieties of cheese, rather than chalk and cheese. I remember that Nadkarni bowled differently just once in life. He flighted the ball and tried to capture wickets. Against Australian during 1964 in Madras. He succeeded so well (5 for 31 and 6 for 91) that it was clear that he had played differently, maybe in the team's defensive cause. Like Gavaskar. This was my first Test and was where Hanumant Singh played, in my opinion, one of the great innings in a losing cause, a la Laxman's 167. 18 exquisite fours out of 94, mostly in the 'V', off McKenzie and Hawke. Ananth: ]] Two round fruits of similar size, widely available, generally affordable, eaten raw or in slices, by themselves or part of a medley in both solid and liquid forms can be compared (http://improbable.com/airchives/paperair/volume1/v1i3/air-1-3-apples.html). Despite these similarities, apples and oranges do not taste the same. So if we compare the salary of a clerk with his officer we are comparing apples with oranges.

  • azhar on April 22, 2012, 21:57 GMT

    big bang theory is piece of cake compared to understanding this analysis. based on this all mediocre bowlers are there with the exception of only handful of greats. [[ Lack of understanding should not lead people to make generally worthless statements. This is an analysis of single innings. Why should one player not have one such day in his career. All bowling analysis need not come around to the same set of 10 bowlers. Then what is the fun in it. On that December day at Adelaide, Agarkar had his single great day in the sun. He performed 3.67 times better than his more illustrious team-mates. What prevents us from accepting this single show of briliance and give him his due. He won the match for India, with his bowling. I will give up if anyone brings in Dravid and Laxman. They have been discussed ad infinitum. Ananth: ]]

  • Alex on April 22, 2012, 21:51 GMT

    @Ananth: Nice article.

    1. I am a bit surprised that Morkel, a superb bowler, is way ahead of Steyn on career analysis based on SPF-1. [[ This happens when a bowler who has bowled around 50 spells, has one mind-blowing (and more) performance. Morkel has one such. In two year's time his figure will come down. Clarke has one such spell but he is a non-regular bowler. Ananth: ]] I think you took CPF as the average of SPF-1. If you compute CPF as CPF = (avg of all other bowlers across all innings in which Morkel bowled)/( Morkel's career avg) then you might get a different, and perhaps a more valid, result. [[ But that will be a different peer comparison and has been done in the past. This is different. The base here is what happened in a single innings. Ananth: ]] 2. It must be pointed that Barnes' phenomenal stats are thanks to the then minnows SA. He averaged 22 vs Oz, an excellent figure but nothing out of the world. I leave Murali out of all discussions. [[ I hope not because you have a complex against Murali. Ananth: ]] 3. Happy to see McDermott and Donald do well in the RpO vs BpW chart. McDermott was genuinely great outside the subcontinent and Donald, a largely forgotten figure now, was arguably a top 3 fast bowler 1992 through 1999. Till 1999, many placed Donald ahead of McGrath.

  • Anshu N Jain on April 22, 2012, 18:55 GMT

    Good one Ananth..

    Havent gone through the tables in detail yet, but a quick suggestion on treating 0 wickets captured by the team as a notional 1 for SPF-1 and SPF-3.

    Extend the largesse to the bowler in question as well. Not only is it fair to the bowler, it also does not throw the SPFs out of gear.

    So for Morkel, Laker, Waqar, Kapil, Saqlain and Kumble in the table for SPF-1 (and all other bowlers whose peers were completely inept in taking wickets in that innings for SPF-1 and SPF-3), divide the runs conceded and balls bowled respectively by adding 1 to the wickets captured (even if for Laker and Kumble, this means dividing by 11). [[ Anshu I have not increased the number by 1. I have considered that a '0' cannot be the divisor. That is the only reason I have taken it as 1. And if the others captured 1+ wickets I have kept it as such. What justification do I have to divide 74 by 11. How do I explain that to the nice cricket enthusiasts who come to read this article. Ananth: ]]

  • shrikanthk on April 22, 2012, 17:25 GMT

    Nice Chart. As soon as I saw the chart, I checked to see if Waqar and Trueman are close to each other. They were! Made me feel good.

    Trueman's Economy rate is very good by modern standards. But by the standards of the fifties he was a somewhat expensive bowler, with a brilliant strike rate. So I naturally expected him to be close to Waqar (despite the absolute difference in their eco rates).

    Also where does O'Reilly figure in the BCG table. His BpW is often scoffed at. But in a peer analysis, it may not seem all that bad. [[ Problem is that O'Reilly had only 48 spells and just missed the qualification. I probably should have had him included as an honorary entry. His Spf-1 is 1.58, Spf-2 is 1.23 and Spf-3 is 1.34. So he would be comfortably in the top-right qudrant. But I am not going to include him in the graph. Difficult thing to replace. I will include him in the Excel sheet and table. They are under my control. Ananth: ]]

  • Som on April 22, 2012, 16:56 GMT

    Ananth - This analysis is freakishly brilliant. More on the content later, just wanted to comment on the context. You rock !

  • No featured comments at the moment.

  • Som on April 22, 2012, 16:56 GMT

    Ananth - This analysis is freakishly brilliant. More on the content later, just wanted to comment on the context. You rock !

  • shrikanthk on April 22, 2012, 17:25 GMT

    Nice Chart. As soon as I saw the chart, I checked to see if Waqar and Trueman are close to each other. They were! Made me feel good.

    Trueman's Economy rate is very good by modern standards. But by the standards of the fifties he was a somewhat expensive bowler, with a brilliant strike rate. So I naturally expected him to be close to Waqar (despite the absolute difference in their eco rates).

    Also where does O'Reilly figure in the BCG table. His BpW is often scoffed at. But in a peer analysis, it may not seem all that bad. [[ Problem is that O'Reilly had only 48 spells and just missed the qualification. I probably should have had him included as an honorary entry. His Spf-1 is 1.58, Spf-2 is 1.23 and Spf-3 is 1.34. So he would be comfortably in the top-right qudrant. But I am not going to include him in the graph. Difficult thing to replace. I will include him in the Excel sheet and table. They are under my control. Ananth: ]]

  • Anshu N Jain on April 22, 2012, 18:55 GMT

    Good one Ananth..

    Havent gone through the tables in detail yet, but a quick suggestion on treating 0 wickets captured by the team as a notional 1 for SPF-1 and SPF-3.

    Extend the largesse to the bowler in question as well. Not only is it fair to the bowler, it also does not throw the SPFs out of gear.

    So for Morkel, Laker, Waqar, Kapil, Saqlain and Kumble in the table for SPF-1 (and all other bowlers whose peers were completely inept in taking wickets in that innings for SPF-1 and SPF-3), divide the runs conceded and balls bowled respectively by adding 1 to the wickets captured (even if for Laker and Kumble, this means dividing by 11). [[ Anshu I have not increased the number by 1. I have considered that a '0' cannot be the divisor. That is the only reason I have taken it as 1. And if the others captured 1+ wickets I have kept it as such. What justification do I have to divide 74 by 11. How do I explain that to the nice cricket enthusiasts who come to read this article. Ananth: ]]

  • Alex on April 22, 2012, 21:51 GMT

    @Ananth: Nice article.

    1. I am a bit surprised that Morkel, a superb bowler, is way ahead of Steyn on career analysis based on SPF-1. [[ This happens when a bowler who has bowled around 50 spells, has one mind-blowing (and more) performance. Morkel has one such. In two year's time his figure will come down. Clarke has one such spell but he is a non-regular bowler. Ananth: ]] I think you took CPF as the average of SPF-1. If you compute CPF as CPF = (avg of all other bowlers across all innings in which Morkel bowled)/( Morkel's career avg) then you might get a different, and perhaps a more valid, result. [[ But that will be a different peer comparison and has been done in the past. This is different. The base here is what happened in a single innings. Ananth: ]] 2. It must be pointed that Barnes' phenomenal stats are thanks to the then minnows SA. He averaged 22 vs Oz, an excellent figure but nothing out of the world. I leave Murali out of all discussions. [[ I hope not because you have a complex against Murali. Ananth: ]] 3. Happy to see McDermott and Donald do well in the RpO vs BpW chart. McDermott was genuinely great outside the subcontinent and Donald, a largely forgotten figure now, was arguably a top 3 fast bowler 1992 through 1999. Till 1999, many placed Donald ahead of McGrath.

  • azhar on April 22, 2012, 21:57 GMT

    big bang theory is piece of cake compared to understanding this analysis. based on this all mediocre bowlers are there with the exception of only handful of greats. [[ Lack of understanding should not lead people to make generally worthless statements. This is an analysis of single innings. Why should one player not have one such day in his career. All bowling analysis need not come around to the same set of 10 bowlers. Then what is the fun in it. On that December day at Adelaide, Agarkar had his single great day in the sun. He performed 3.67 times better than his more illustrious team-mates. What prevents us from accepting this single show of briliance and give him his due. He won the match for India, with his bowling. I will give up if anyone brings in Dravid and Laxman. They have been discussed ad infinitum. Ananth: ]]

  • milpand on April 22, 2012, 23:37 GMT

    I liked 'muddy lane' in "This match was incidentally played on a muddy lane around Wankhede Stadium." very much.

    But not a big fan of the repeated use of 'cheese & chalk' both by you and others. It does not irk me too much in this post but their are instances in recent past where a better alternative could have been used.

    I think we should use 'chalk and cheese' to indicate that apart from a superficial similarity, nothing else is same. I wonder whether Nadkarni and Maninder are that dissimilar? [[ I think the one major difference was that Nadkarni's obsession was to bowl run-less balls and overs. Maninder was more attacking. But two varieties of cheese, rather than chalk and cheese. I remember that Nadkarni bowled differently just once in life. He flighted the ball and tried to capture wickets. Against Australian during 1964 in Madras. He succeeded so well (5 for 31 and 6 for 91) that it was clear that he had played differently, maybe in the team's defensive cause. Like Gavaskar. This was my first Test and was where Hanumant Singh played, in my opinion, one of the great innings in a losing cause, a la Laxman's 167. 18 exquisite fours out of 94, mostly in the 'V', off McKenzie and Hawke. Ananth: ]] Two round fruits of similar size, widely available, generally affordable, eaten raw or in slices, by themselves or part of a medley in both solid and liquid forms can be compared (http://improbable.com/airchives/paperair/volume1/v1i3/air-1-3-apples.html). Despite these similarities, apples and oranges do not taste the same. So if we compare the salary of a clerk with his officer we are comparing apples with oranges.

  • Alex on April 23, 2012, 2:42 GMT

    @Ananth:

    1. What would be the numbers for Roberts and Fazal Mahmood? I think Thommo's numbers would be similar to Lee's and Shoaib's would be just a bit better.

    2. Sobers missed your cut-off of 240 wkts by just 5. What are his numbers It'd be nice to see where he is placed w.r.t. Kallis. [[ Instead of answering individual queries, I will upload another table which contains all three Spf values so that you can see for yourself. Will be done by today. Ananth: ]] 3. Gibbs had an extraordinary RPO=1.97. I always feel he belongs in the all-time WI XI, if for nothing then just for the sake of variety and to ensure adequate over-rate. Pl give an honorary mention in the excel sheet to Nadkarni (RPO=1.67!!).

  • Ananth on April 23, 2012, 4:12 GMT

    The Career Excel table has now been uploaded with the Career Spf-2 and Spf-3 information. You could get the information for all bowlers, for Spf-1, Spf-2 and Spf-3, here. the graph will not be updated. The link is again provided below. http://dl.dropbox.com/u/39210851/Bowler_Spell_Peer_Index_CareerSummary.xls Ananth

  • Som on April 23, 2012, 5:30 GMT

    Absence of Mcgrath and Warne in any of these tables is quite a revelation (and actually shouldn't be). They perhaps with some effort from Gillespie, Lee foiled each others chances of making out-of-peer group contributions. But OZ definitely did not mind. An interesting way to look at this data would be to group Mcgrath and Warne together and run these same numbers against others. That way, when done for all teams across era's would give us the best hunters in 'pairs'. The same theory applies to Macko and to a lesser degree to Wasim.

    I had done some analysis on my end when selecting India's all time best test XI, and had found Gupte to be making a compelling argument (only to be losing out marginally to the comprehensive figures of Kumble).

    Nadkarni was a champ! Never seen another like him. [[ Som, I have a reference in my Pending Box, "Two bowlers together". I will do an analysis on that and as part of that I can consider the two as Siamese twins and do the peer analysis also. At a later stage I will ask for possible pairs. Ananth: ]]

  • Balumekka on April 23, 2012, 5:42 GMT

    Ananth - Brilliant work! Love your opinion on the 'no average'/ average infinity) issue! Performance analysis compared to peers is very sensible. However, I was wandering whether it would be more meaningful if you can do a correction for the final figure considering the quality of the wickets captured(based on the batting average of the batsman dismissed, may be). [[ Balu, if this door is opened a bit it will get opened completely. Let us leave the concept as it is. A peer analysis with no other factors coming in. Ananth: ]]