|Photos||Video & Audio||Blogs||Statistics||Archive||Games||Mobile|
I started this. So I have to finish it...
It is amusing. A few days back whole lot of people were lambasting me for not having Tendulkar on top. Now another set of people are screaming that Tendulkar is on top. Hey guys, this is only an analysis. I am one insignificant analyst who works with a computer and a Cricket database. The greats remain greats, whatever I (or for that matter you all) say.
Just one more thing. Unlike what some have suggested, I have not gone out of the way to put Tendulkar on top. He is one of the greatest but NOT my favourite batsman.
As done before I have incorporated a summary response to readers' comments at the end.
In my previous article I had taken two important ODI batting measures and attempted to analyse batsmen skills using those. It elicited the usual comments on the additional parameters for consideration. Hence instead of doing a straightforward follow-up to that analysis, I have gone the whole hog and after considering all relevant parameters, come out with what I feel should be a very fair ODI batsmen ranking based on what they have achieved over their careers.
The following 8 facors are considered.
1. Total runs scored (TRS) 2. Batting Average (AVGE) 3. Runs per Innings (RPI) 4. Strike Rate (STRT) 5. Quality of bowlers faced (BOWQTY) 6. % of Team runs (TRPER) 7. Wins achieved - Absolute number of wins (WINS) - Win % of matches played (WINSPER) 8. MOM awards received/frequency (MOM).
A brief description of each factor and the weights given to each parameter is outlined below. The total points add up to a nice round sum of 100.
1. Total runs scored (20 points)
This is a recognition of the longevity of the player. There is no doubt that the runs scored has to be given decent weightage. At the same time care has been taken to see that the olden era players such as Richards, Greenidge et al do not suffer unduly. My belief is that it is very unlikely for any batsman, including Tendulkar, to exceed 20000 runs. Hence the limit seems correct. The formula used is
2. Batting Average (15 points)
This is a straightforward calculation. We need not worry about not-outs since there is a separate factor for that. Since the batting average is unlikely ever to exceed 60.0, we are within the maximum level. The formula used is
Note: David Barry is doing some simulation work with a view to establish a correlation between Average and Strike Rates. It is too early to incorporate these first level findings. Hence at this stage I have taken the simple, easily understandable method of separating the Average and Strike Rate measures with individual weightages. Similarly Jeff Grimshaw's ideas about treating balls played as a resource and giving credit for the same is quite good. However I do not want too many overlapping parameters. Already I have Average and RPI.
3. Runs per Innings (5 points)
This is to mitigate the factor of a high number of not-outs, especially for middle-order batsmen. Again a straightforward calculation. Since the Batting average is unlikely ever to exceed 50.0, we are within the maximum level. The formula used is
Note: I briefly toyed with Abhihjeet Dongre's excellent suggestion of excluding from the total number of innings the innings in which the batsman has finished not out at a score below his batting average. This redresses the balance towards middle order batsmen slightly. However I have finally rejected this tweak since I feel that they have already got the full benefit of not outs while calculating the Batting Average. The purpose of separation of these two factors will be lost if I do not use the full complement of innings played.
4. Strike Rate (25 points)
I consider this factor as the most important measure and that is reflected in the weightage. However much we talk about the importance of scoring runs, it is essential that these are scored at a reasonable pace. It does not mean that every century should be a run-a-ball one. However, it is true that many a match has been lost because the batsmen have not moved up the scoring rate at the right time.
However a major tweak has been done. The actual strike rates have been adjusted up or down based on the decade scoring rates pro-rata. In other words, if Viv Richards played between 1975 and 1991, his actual scoring rate has been adjusted pro-rata for the three decades, viz., 1970s, 1980s and 1990s. In general this will mean that the older players will get a slight benefit since the scoring rates were lower, as indicated in the table below.
AllMats 1970s 1980s 1990s 2000sThe actual and adjusted strike rates for a few top players is given below. All these adjustments seem very reasonable. The only clear cases are for batsmen such as Pietersen and Dhoni who have played all their matches in the current decade and hence have the same adjustment of -4.4%. The others are pro-rata. For instance, Tendulkar's and Lara's strike rates have been adjusted much less since they have played during 1980s, 1990s and 2000s. Zaheer Abbas gains the maximum since his career spanned 1975-1985, the low-scoring years.
Matches played 2759 82 516 933 1228 Batsmen innings 47947 1418 8838 16266 21425 Runs scored 1142018 30292 202884 386508 522334 Balls bowled 1473233 46208 277516 505727 643782 Runs per ball 0.775 0.656 0.731 0.764 0.811 % of all-matches avge 100.0% 84.6% 94.3% 98.6% 104.7%
Batsman Prev SR Adj SR % chgSince the only Strike Rate to exceed 1.00 is that of Shahid Afridi, I have accepted the fact that only he will exceed the maximum level. The formula used is
Richards I.V.A 90.2 98.4 +9.1% Haynes D.L 63.1 66.9 +6.0% Jones D.M 72.6 75.4 +3.9% Greenidge C.G 64.9 70.8 +9.1% Zaheer Abbas 80.0 89.7 +12.1%
Tendulkar S.R 85.5 84.6 -1.1% Jayasuriya S.T 91.0 90.1 -1.1% Gilchrist A.C 96.9 94.4 -2.7% Lara B.C 79.5 78.6 -1.3% Sehwag V 99.1 95.3 -3.8% Shahid Afridi 111.2 108.2 -2.7% Klusener L 89.9 88.2 -1.9% Dhoni M.S 91.3 87.2 -4.4% Pietersen K.P 87.5 83.6 -4.4%
5. Quality of bowling faced (15 points)
This is a double weighted adjustment. The objective is to make sure that the runs acored against stronger teams such as Australia are given much higher weighting than the runs scored against weaker countries such as Zimbabwe. Care also has to be taken that the weaker Australian teams such as those during the mid-1980s are treated accordingly. The complex process is explained below.
First a bowling quality index is found for each innings. This is done by the following formula (somewhat similar to the one used by David Barry). I had thought of this earlier, but dismissed it as too complex. Now I think it is necessary.
Sum of (Balls bowled by each bowler x Bowler's bowling avge) Innings BQI = ----------------------------------------------------- Sum of (Balls bowled by each bowler)In one of my earlier articles on Team Strength analysis I used a simple average of the top 5 Bowling averages. That was when I was trying to find the strength of team as it walked on to the field. However here I am trying to find how valuable the batsman's innings was. Hence the actual deployment of the bowling resources is necessary. Wasim Akram will make the Pakistani team that much strong, on paper, however, if he did not bowl a single ball, to that extent the bowling lacks sting.
Now comes the second weighting. For this the actual scores of batsman and the Innings BQI are used. The formula is explained below.
Sum of (Batsman innings score x Innings BQI) Batsman career BQI = -------------------------------------------- Sum of (Batsman innings score)There is some convergence of values as batsmen score many runs. Note the BOWQTY value for the top 5 batsmen. Hence special care has to be taken to assign points. Amongst batsmen who have scored greater than 2000 runs, Craig McMillan is the best with a BQI of 34.48 and Habibul Bashar the worst with a BQI of 43.47. If we lower the limit to 1000 runs, Nicky Boje is the best with a BQI of 31.3 and Glenn Turner the worst with a BQI of 47.63. No batsman has a career BQI below 30.00 and no batsman has a career BQI above 50.0. The Batsman career BQI is used to derive the index value based on the following formula.
6. % of Team runs (5 points).
The value of a batsman to the team is also determined by the share of the batting load he takes. In other words the % of team runs he scores. This is a secondary parameters and has a weighting only of 5 points. With a criteria of 2500 runs and above, the highest share of team runs scored is by Zaheer Abbas with 21.6%, followed by Greenidge with 19.2%, then by Richards with 19.2% and finally by Tendulkar with 18.1%. The formula used is
Upto this point, the full weight will be given only if the batsman has scored above 2000 runs. Else the points secured will be proportionately downsized.
7. Wins achieved (5 points)
Winning is something special, if not everything (as the Americans profess). No one wants to lose. Hence we should give value to this important aspect of the game without going overboard. This is done in two parts. The first is to derive an index value solely based on the number of wins achieved. This will benefit players who have played more games and have been part of successful teams. The highest number of wins achieved is 220 by Jayasuriya, followed by Ponting with 216, Inzamam with 214, Gilchrist with 214 and Tendulkar with 206. The formula used is
8. Win % achieved (5 points)
What about Richards who achieved 132 wins in 187 matches (a 70.6 win %), which is much higher than that of Tendulkar, 206 wins in 417 matches (49.4%) or Steve Waugh, 196 in 325 (60.3%). His win % suffers only in comparison to the current Australian team, some of whom having over 75%.
This factor addresses this problem. Credit is given to the % of wins achieved, subject to minimum number of matches being reached. The formula used is
9. MOM awards received (5 points).
The last parameter is on the MOM awards achieved. This is the only subjective measure, as pointed to by Shankar Krishnan of Riyadh. However since this is the only individual evaluation measure available I have to consider it. Whatever be the idiosyncracies of the adjudicators there is no doubt that the MOM awards are a pointer to the contribution to the wins achieved by the team.
I have tried to remove the subjective factor, to a certain extent, by considering the frequency of awards also in addition to the absolute number of awards. This is also fair to the older players. Consider this. Richards has got 31 awards in 187 matches. He lags far behind Jayasuriya who has got 45 awards in 415 matches. However when we consider the frequency, Richards has a frequency of one in 6 matches, while Jayasuriya, one in 9.2 matches. Incidentally Tendulkar leads the absolute number of awards with 55. The frequency ranges from 6.0 to 20.0 (limiting value). The formula used is
For the last two points, the full weight will be given only if the batsman has played above 50 matches. Else the points secured will be proportionately downsized.
Now the table of top 30 ODI batsmen of all time. The table is current upto match 2759, the facile English win over the hapless South Africans, giving them a 4-0 lead.
The top ODI batsmen of all time - as on 28 August 2008.Tendulkar is on top, and deservedly so. He has not only scored lots of runs but scored these at a good pace, scored these against good bowlers and contributed more than his share to the Indian cause.
No.Cty Batsman Total Runs Avge R/I S/R BwQty Wins Win% % TS MOMs
100.0 20.0 15.0 5.0 25.0 15.0 5.0 5.0 5.0 5.0
1.Ind Tendulkar S.R 79.27 16.36 11.08 4.02 21.16 12.13 4.12 2.47 3.61 4.32 2.Win Richards I.V.A 73.14 6.72 11.75 4.02 24.60 12.20 2.64 3.53 3.85 3.83 3.Slk Jayasuriya S.T 72.24 12.80 8.18 3.13 22.52 12.01 4.44 2.64 2.89 3.63 4.Aus Ponting R.T 71.87 11.11 10.81 3.81 19.63 12.63 4.32 3.59 3.19 2.78 5.Aus Gilchrist A.C 71.12 9.62 8.97 3.45 23.59 12.12 4.04 3.52 2.93 2.88 6.Win Lara B.C 67.70 10.40 10.12 3.60 19.64 12.38 2.78 2.33 3.43 3.01 7.Saf Kallis J.H 67.70 9.61 11.17 3.64 17.33 12.98 3.50 3.15 3.26 3.05 8.Pak Inzamam-ul-Haq 66.72 11.74 9.88 3.35 18.31 11.81 4.28 2.83 3.04 1.48 9.Aus Bevan M.G 66.00 6.91 13.40 3.53 18.31 13.88 3.10 3.34 3.00 0.53 10.Ind Ganguly S.C 65.87 11.36 10.26 3.79 18.14 10.64 2.98 2.40 3.27 3.03 11.Pak Saeed Anwar 65.55 8.82 9.80 3.62 20.20 10.95 2.82 2.85 3.42 3.07 12.Aus Waugh M.E 65.50 8.50 9.84 3.60 19.39 12.29 3.04 3.11 3.22 2.52 13.Aus Symonds A 65.43 5.01 10.09 3.19 22.41 12.58 2.98 3.86 2.58 2.74 14.Slk de Silva P.A 65.02 9.28 8.73 3.14 20.61 12.52 2.56 2.08 3.06 3.05 15.Win Haynes D.L 64.62 8.65 10.34 3.65 16.74 12.16 3.18 3.34 3.63 2.93 16.Saf Gibbs H.H 63.66 7.59 9.12 3.39 20.24 11.95 2.88 3.12 3.03 2.36 17.Ind Dravid R 63.47 10.59 9.87 3.44 17.36 13.25 3.14 2.36 3.00 0.47 18.Pak Mohammad Yousuf 63.22 9.24 10.80 3.64 18.19 10.81 3.12 2.90 3.12 1.40 19.Saf Kirsten G 63.06 6.80 10.24 3.67 17.88 13.45 2.40 3.24 3.35 2.03 20.Saf Klusener L 62.90 3.58 10.28 2.61 22.05 13.92 2.18 3.19 2.27 2.83 21.Aus Jones D.M 62.82 6.07 11.15 3.77 18.86 11.99 1.96 2.99 3.55 2.48 22.Aus Hayden M.L 62.80 6.13 10.95 3.96 19.36 12.26 2.38 3.70 3.34 0.72 23.Pak Javed Miandad 62.74 7.38 10.43 3.39 17.98 13.04 2.38 2.55 3.40 2.18 24.Saf Rhodes J.N 62.73 5.93 8.78 2.70 20.11 15.00 3.10 3.16 2.46 1.47 25.Eng Pietersen K.P 62.46 2.82 11.96 3.87 20.90 14.46 0.68 2.10 3.44 2.24 26.Ind Sehwag V 62.31 5.81 8.11 3.12 23.82 12.41 1.90 2.49 2.70 1.95 27.Ind Dhoni M.S 62.26 3.79 11.85 3.54 21.80 12.73 1.30 2.71 2.90 1.63 28.Ind Azharuddin M 62.10 9.38 9.23 3.04 18.97 12.09 3.20 2.40 2.90 0.89 29.Aus Waugh S.R 61.95 7.57 8.23 2.63 19.26 13.75 3.92 3.02 2.38 1.21 30.Saf Cronje W.J 61.91 5.57 9.66 3.18 19.26 13.51 2.34 3.11 2.92 2.35
Richards is in second place, again deservedly so. He has scored only 6721 runs, but made up for the huge shortfall in index points with his outstanding average, strike rate, win % and MOM frequency. He may very well move a little bit down in the list in the years to come. But will not lose any of the aura.
Jayasuriya is next, having made up for his low Average and RPI with a mountain of runs scored at a scorching pace. The Lankan readers will be happy that the contributions of the entertainer non-pareil have been recognized. He has managed to retain the third position depsite a poor run of ODI matches against India.
Ponting and Gilchrist, two great Australian batsman, follow in the next two positions, through different combination of high points. Ponting with high average and good strike rate while Gilchrist with lower average and excellent strike rate. Both have great win related numbers.
Lara, Kallis Inzamam, Bevan and Ganguly complete the top 10. This elite placing of these quality batsmen cannot be debated. In fact Lara and Kallis exchanged places after the last match.
There is no doubt that players such as Pietersen (25th currently), Sehwag (26th), Dhoni (27th) and Hussey (39th) will move up the list as they score more runs. However this may be partly compensated by the possible decrease in their averages. Dhoni is surely on the way to becoming an excellent finisher in the Bevan/Hussey mode and as such is unlikely to drop his average. Pietersen's average could drop a little bit. Hussey's could drop significantly unless otherwise he does what Bevan did over a long career.
It should be noted that if we change the weightings, the batsmen will move up or down the list. For instance, Strike Rate could be reduced to 20 points. In that case, Jayasuriya and Ponting will exchange places. But these are minor movements only. It is my firm belief that the top 2, Tendulkar and Richards will remain where they are, whatever be the weightings.
Batsmen such as Kluesener, Dhoni and Pietersen, who have not even scored 4000 ODI runs have managed to reach the top 30 positions in the all-time best batsmen table. This indicates that the weightings for non-longevity measures have been given due importance.
Finally, one important point to be noted. No analyst starts with an idea to prove that one batsman is superior to another or push their favourite batsmen on top. Such shallow analyses will be found out in no time at all. The idea is to come out with a vehicle for healthy discussion and exchange of views. Hence please avoid rude and vicious comments. They have no chance of being read by any one. Pl make your point in a courteous and acceptable manner. The readers have their right to be heard but also their responsibilities to be constructive and courteous.
To view the complete list, click here
This list consists of batsmen who have scored a minimum of 1000 ODI runs. Please remember that many of the calculated points are downsized for batsmen in the 1000-2000 range. They are included only to show where some of the batsmen from the lesser countries stand.
Summary response to readers' comments (Possible tweaks)
1. Avoidance of double weighting for "Wins".
2. Possible cap on Runs scored weighting.
3. Adjust for the paucity of matches played during the early 10 years.
4. Giving weight to key tournament wins such as World Cup and Champion's Trophy.
5. The subjective nature of MOMs, already mentioned by me in the main post does not go well with readers.
6. Quite a few readers have, while accepting Tendulkar's position at no.1, have questioned the wide gap between Tendulkar and Richards. It worries some readers that this gap will keep on widening.
Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systemsFeeds: Anantha Narayanan
© ESPN EMEA Ltd.
|Comments have now been closed for this article
Anantha spent the first half of his four-decade working career with corporates like IBM, Shaw Wallace, NCR, Sime Darby and the Spinneys group in IT-related positions. In the second half, he has worked on cricket simulation, ratings, data mining, analysis and writing, amongst other things. He was the creator of the Wisden 100 lists, released in 2001. He has written for ESPNcricinfo and CastrolCricket, and worked extensively with Maruti Motors, Idea Cellular and Castrol on their performance ratings-related systems. He is an armchair connoisseur of most sports. His other passion is tennis, and he thinks Roger Federer is the greatest sportsman to have walked on earth.