Trivia - bowling November 19, 2008

Analysing the 'who' and 'when' of ODI spells

After a couple of rather heavy Test articles, I have reverted to ODIs, that too bowling, an area which tends to be neglected
32

After a couple of rather heavy Test articles, I have reverted to ODIs, that too bowling, an area which tends to be neglected. Also a simpler post which would make for easier reading.

Let us compare two bowling innspells. (We again renew our acquaintance with the term "innspell" which indicates the complete bowling effort in a single innings by a bowler.)

G.A.Mcgrath   7.0-4-15-7
A.D.Mullally  8.0-1-18-4 

McGrath's first innspell contained the "priceless" scalps of Jan-Berrie Burger, Morne Karg, Danie Keulder, Gavin Murgatroyd, Deon Kotze, Louis Burger and Bjorn Kotze, Namibian batsmen who would find it difficult to get into the Tamil Nadu, Victoria or Surrey state/county teams.

The batsmen who Mullally dismissed in the second innspell are Mark Waugh, Adam Gilchrist, Ricky Ponting and Damien Martyn, all for low scores. There is no doubt which is, by far, the more outstanding innspell.

This article analyses ODI bowling innspells from the point of view of "who" was dismissed and more relevantly, "when" was he dismissed. Let me say at the outset that this article is not to determine the best ODI innspells ever, in which case I have to take into account the bowling accuracy, result, match status, match importance and quality of opposition (in a summary form).

We only look at the wickets captured by the bowlers. Who was dismissed credits the bowler with the appropriate batting measure of the batsmen dismissed. (Viv Richards is better than Ponting who incidentally is better than Herschelle Gibbs who in turn is better than Shiv Chanderpaul and so on.) The point is that, at any stage in the innings, it is important to capture Sachin Tendulkar's wicket. I am sure a reader will point to a dismissal of Tendulkar in the 50th over and argue that the wicket is not valuable. I accept that. However I am talking of most matches.

When was the batsman dismissed is the other equally (or probably more) important factor. It is important to dismiss Tendulkar; it is more important to dismiss him sooner than later. Any dismissal below the batting average is good. The lower the batsman score at the time of dismissal the better it would be for the bowling team.

The measures used are equally simple. For the first, who was dismissed, I give credit to the bowler to the extent of the runs per innings value, since that is the fairest of measures. I seriously thought of ODI Batting Index, already presented and discussed in an earlier article. But the big problem of lower strike-rates during the early years is a deterring factor and decade level adjustments make the work quite complex. Batting average tends to favour those with high number of not-outs. Hence runs per innings seem to be the appropriate measure.

The when situation is addressed in an equally simple manner. Credit is given only if the batsman is dismissed for below the batting average (yes, this time we use the more relevant batting average measure). The extent of credit is the difference between the batsman score and batting average. The higher this factor is, the lower the batsman has been dismissed for. The added advantage of using the batting average is that a higher credit can be given for a very low score (0-5) dismissal of a top batsman.

The other important decision is only to consider the wickets 1-8. Lower-order wickets will only distort the overall picture and place more importance on the number of wickets than the quality of wickets. Of course, there may be situations in close matches where lower-order wickets are crucial. Again, my analysis is a limited one covering only the quality of wickets taken.

Readers might ask, why 1-8 and not 1-7 or 1-9. Just to strengthen this point I did a quixotic analysis of the No. 8 batting position of qualifying ODI innings (in all 3907). The results are fascinating. The average batting average of the batsmen who batted at No. 8 is a fairly high 17.80 and the highest is Bevan (53.58). Also 547 (14%) of these no.8 innings have been played by batsmen with batting average exceeding 25.00. This is mainly because many fine allrounders such as Shaun Pollock, Lance Klusener, Wasim Akram, Shahid Afridi, Chris Cairns et al have often batted at No. 8. Also one pinch-hitter at the top forces a top batsman into the position. Hence my decision to consider batting positions 1-8.

I have considered bowlers who have taken four wickets and above. I have lowered the cut-off to four wickets since there are quite a few cases of a bowler taking four top-order wickets. Even with this cut-off, there are 1075 instances. When required, batting averages are tweaked to take care of the low matches played (below 25).

An aside. We have already seen that Bevan (53.58) is the best batsmen to have batted at No. 8. Richards (yes, the great Viv with an average of 47.00) once batted at no.9. Klusener (41.10) once adorned the No. 10 position and the outstanding Salim Malik (32.89) once came in last. These are only aberrations. Richards was run-out and Saleem Malik remained not out.

After this fairly long introduction, let us look at the top 20 performances. This analysis covers matches upto the second ODI between India and England. Two four-wicket innspells were performed in this match, one each by Yuvraj Singh and Broad.

No Bowler          Year MtNo For-Opp Top Dis TopPts DisPts TotPts
InnsSpell

1.Waqar Younis 2001 1724 Pak-Eng 7 7 185.17 172.96 358.13 10.0-0-36-7 1. Trescothick M.E 0 [Rpi:35.53] (Avg:37.37 - +37.37) 2. Knight N.V 9 [Rpi:36.36] (Avg:40.40 - +31.40) 3. Stewart A.J 18 [Rpi:28.87] (Avg:31.60 - +13.60) 4. Vaughan M.P 2 [Rpi:23.88] (Avg:27.15 - +25.15) 5. Shah O.A 3 [Rpi:24.67] (Avg:27.75 - +24.75) 6. Collingwood P.D 0 [Rpi:27.29] (Avg:34.69 - +34.69) 8. Cork D.G 4 [Rpi: 8.57] (Avg:10.00 - + 6.00)

2.Vaas WPUJC 2001 1776 Slk-Zim 7 7 169.77 173.24 343.01 8.0-3-19-8 1. Ebrahim D.D 0 [Rpi:18.99] (Avg:20.61 - +20.61) 2. Flower G.W 1 [Rpi:30.83] (Avg:33.69 - +32.69) 3. Carlisle S.V 16 [Rpi:25.61] (Avg:27.68 - +11.68) 4. Flower A 0 [Rpi:32.63] (Avg:35.34 - +35.34) 5. Wishart C.B 6 [Rpi:20.96] (Avg:23.23 - +17.23) 7. Taibu T 0 [Rpi:22.24] (Avg:27.39 - +27.39) 8. Streak H.H 0 [Rpi:18.51] (Avg:28.30 - +28.30)

3.Aaqib Javed 1991 0685 Pak-Ind 7 6 196.25 144.32 340.57 10.0-1-37-7 1. Shastri R.J 15 [Rpi:24.28] (Avg:29.05 - +14.05) 2. Sidhu N.S 21 [Rpi:34.76] (Avg:37.09 - +16.09) 3. Manjrekar S.V 52 [Rpi:28.49] (Avg:33.23) 4. Azharuddin M 0 [Rpi:30.45] (Avg:36.92 - +36.92) 5. Tendulkar S.R 0 [Rpi:40.20] (Avg:44.34 - +44.34) 7. Kapil Dev N 8 [Rpi:19.11] (Avg:23.79 - +15.79) 8. Prabhakar M 7 [Rpi:18.96] (Avg:24.13 - +17.13)

4.Collins P.T 2005 2212 Win-Aus 5 4 172.51 140.48 312.99 10.0-1-43-5 1. Clarke M.J 21 [Rpi:32.88] (Avg:42.47 - +21.47) 2. Hayden M.L 3 [Rpi:39.57] (Avg:43.81 - +40.81) 3. Ponting R.T 0 [Rpi:38.06] (Avg:43.24 - +43.24) 5. Lehmann D.S 4 [Rpi:30.48] (Avg:38.96 - +34.96) 6. Katich S.M 76 [Rpi:31.52] (Avg:35.78)

5.Bond S.E 2003 1986 Nzl-Aus 6 6 167.65 143.88 311.53 10.0-2-23-6 1. Gilchrist A.C 18 [Rpi:34.48] (Avg:35.89 - +17.89) 2. Hayden M.L 1 [Rpi:39.57] (Avg:43.81 - +42.81) 3. Ponting R.T 6 [Rpi:38.06] (Avg:43.24 - +37.24) 4. Martyn D.R 31 [Rpi:29.37] (Avg:40.81 - + 9.81) 7. Hogg G.B 0 [Rpi:12.15] (Avg:20.26 - +20.26) 8. Harvey I.J 2 [Rpi:14.02] (Avg:17.88 - +15.88)

6.Waqar Younis 2001 1725 Pak-Aus 5 4 169.03 137.11 306.14 8.0-0-59-6 2. Waugh M.E 0 [Rpi:36.02] (Avg:39.35 - +39.35) 3. Hayden M.L 0 [Rpi:39.57] (Avg:43.81 - +43.81) 4. Bevan M.G 5 [Rpi:35.27] (Avg:53.58 - +48.58) 5. Waugh S.R 56 [Rpi:26.28] (Avg:32.91) 7. Symonds A 35 [Rpi:31.89] (Avg:40.37 - + 5.37)

7.Zoysa D.N.T 2004 2158 Slk-Saf 5 5 158.49 145.61 304.10 8.0-0-26-5 1. Smith G.C 14 [Rpi:38.14] (Avg:40.89 - +26.89) 2. Gibbs H.H 7 [Rpi:33.63] (Avg:36.18 - +29.18) 3. Boje N 14 [Rpi:19.92] (Avg:26.68 - +12.68) 4. Kallis J.H 0 [Rpi:36.70] (Avg:45.28 - +45.28) 5. Rudolph J.A 4 [Rpi:30.10] (Avg:35.58 - +31.58)

8.Styris S.B 2002 1843 Nzl-Win 6 5 174.05 124.77 298.82 7.0-0-25-6 1. Gayle C.H 60 [Rpi:37.23] (Avg:40.06) 3. Lara B.C 0 [Rpi:36.00] (Avg:40.49 - +40.49) 4. Hooper C.L 24 [Rpi:27.97] (Avg:35.34 - +11.34) 5. Sarwan R.R 2 [Rpi:35.13] (Avg:44.18 - +42.18) 7. Hinds W.W 4 [Rpi:26.50] (Avg:28.93 - +24.93) 8. Hinds R.O 11 [Rpi:11.22] (Avg:16.83 - + 5.83)

9.Lee B 2005 2284 Aus-Icc 4 4 144.30 147.33 291.63 9.0-2-30-4 1. Gayle C.H 0 [Rpi:37.23] (Avg:40.06 - +40.06) 4. Kallis J.H 2 [Rpi:36.70] (Avg:45.28 - +43.28) 5. Lara B.C 0 [Rpi:36.00] (Avg:40.49 - +40.49) 7. Dravid R 16 [Rpi:34.37] (Avg:39.50 - +23.50)

10.Broad S.C.J 2008 2754 Eng-Saf 5 5 146.71 142.15 288.86 10.0-3-23-5 1. Smith G.C 9 [Rpi:38.14] (Avg:40.89 - +31.89) 2. Gibbs H.H 10 [Rpi:33.63] (Avg:36.18 - +26.18) 3. Kallis J.H 6 [Rpi:36.70] (Avg:45.28 - +39.28) 5. Duminy J.P 8 [Rpi:26.39] (Avg:35.57 - +27.57) 7. Botha J 1 [Rpi:11.85] (Avg:18.23 - +17.23)

11.Joshi S.B 1999 1504 Ind-Saf 5 5 145.59 140.64 286.23 10.0-6- 6-5 1. Dippenaar H.H 17 [Rpi:36.01] (Avg:42.23 - +25.23) 2. Gibbs H.H 18 [Rpi:33.63] (Avg:36.18 - +18.18) 5. Cronje W.J 2 [Rpi:31.80] (Avg:38.65 - +36.65) 7. Rhodes J.N 1 [Rpi:26.98] (Avg:35.12 - +34.12) 8. Pollock S.M 0 [Rpi:17.17] (Avg:26.46 - +26.46)

12.Harmison S.J 2005 2251 Eng-Aus 5 4 176.09 106.75 282.84 10.0-0-33-5 1. Gilchrist A.C 26 [Rpi:34.48] (Avg:35.89 - + 9.89) 2. Hayden M.L 31 [Rpi:39.57] (Avg:43.81 - +12.81) 3. Ponting R.T 0 [Rpi:38.06] (Avg:43.24 - +43.24) 4. Martyn D.R 0 [Rpi:29.37] (Avg:40.81 - +40.81) 6. Hussey M.E.K 84 [Rpi:34.61] (Avg:57.14)

13.Imran Khan 1985 0325 Pak-Ind 5 5 137.69 144.48 282.17 10.0-2-14-6 1. Shastri R.J 0 [Rpi:24.28] (Avg:29.05 - +29.05) 2. Srikkanth K 6 [Rpi:28.22] (Avg:29.02 - +23.02) 4. Vengsarkar D.B 1 [Rpi:29.23] (Avg:34.73 - +33.73) 5. Gavaskar S.M 2 [Rpi:30.31] (Avg:35.14 - +33.14) 6. Amarnath M 5 [Rpi:25.65] (Avg:30.54 - +25.54)

The tables are self-explanatory. For sheer brilliance, intensity and top-drawer quality, the best performance in this regard is Waqar Younis' 7 for 36 against England. Seven wickets taken, all seven top-order wickets, all seven batsmen dismissed below their respective averages. This is sheer magic.

Chaminda Vaas gets into second place with figures somewhat similar to Waqar Younis. Seven top-order wickets, all well below their batting average. Even though this was against Zimbabwe, readers should remember that this Zimbabwe team was not bad, with decent averages and having Heath Streak bat at no.8.

I would personally place Aaqib Javed's performance at par with Vaas. Seven top-order wickets and only Sanjay Manjrekar managing to go past his bowling average. The hat-trick to boot.

Pedro Collins is the highest placed five-wicket bowler in this table. Five top-order Australian wickets, all below their averages. Shane Bond's is somewhat similar. It is only the fact that two of Bond's scalps are of lesser batsmen (Brad Hogg and Ian Harvey) which has pushed him below Collins.

The best-placed Indian bowler is the unheralded Sunil Joshi. His collection of five top South African batsmen, all below their averages is truly outstanding. The best placed Australian bowler is Brett Lee whose collection of four (yes, only four) ICC XI scalps has pushed him into the Top 10. But then the batsmen are Chris Gayle, Jacques Kallis, Brian Lara and Rahul Dravid.

Just for interest, let me go back to the two innspells at the top. McGrath's 7 for 15 is placed in 399th position while Mullally's 4 for 18 is placed in 33rd position. I am sure most people would agree with these placings. Let me add that the incomparable McGrath has seven innspells placed in this table ahead of this particular one (and eight below).

To view the complete list please 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

  • Ross on January 17, 2009, 9:55 GMT

    The BCs had 2 for them and none against, we must be ok... :P

  • Allen Johnston on January 2, 2009, 4:16 GMT

    In terms of the great ODI bowling spells, it must surely be hard to beat the ridiculous one by Phil Simmons vs Pakistan in 1992. Simmons took 4 wickets, for 3 runs in ten overs. 3 Runs in 10 overs! Simmons scalps consisted of Javed Miandad, Aamer Sohail, Saleem Malik and Asif Mujtaba - all in the top 5 of the batting order.

    The West Indies managed to score 214 runs on the pitch.

    (3 runs!!) [[ Yes. Symond's performance is great. It is in the Top-50 with very respectable tally of 235 points. Ananth: ]]

  • Allan Pinchen on December 4, 2008, 9:18 GMT

    Great stuff!! Waqar Younis is about the best I`v`e ever seen,I used to love those inswinging yorkers.By the way I`m Aussie`[58yrs] and loved D K Lillee,he was the best!! Waqar at his peak was unplayable and a great bloke..loved to watch him!

  • Uday Trivedi on December 1, 2008, 11:32 GMT

    One interesting observation on full list data. for India, Agarkar is the top most bowler with 12 entries in the list. In bracket I have put total wks the bowler has taken till now.

    Agarkar - 12 (288) Kumble - 10 (334) Srinath - 10 (315) Zaheer - 8 (205) Prabhakar - 6 (157) Sachin - 6 (152) Pathan - 5 (148) Kapil - 4 (253) Harbhajan - 4 (200) Prasad - 4 (196) Nehra - 4 (90)

    Agarkar has 12 outstanding spells for India,he is no. 3 is all time ODI bowlers list for India, still he didn't get the credit he deserves. I remember during 1997 to early 2000s, Agarkar used to get bagful wickets (with albeit bit high number of runs).For me,he is India's top ODI bowler with his outstanding strike-rate (32.9 - beaten only by Sreesanth's 32.6) and average (27.85 - beaten only by Kapil's 27.45).

    One surprise is Sachin ahead of Kapil & Harbhajan. However,no one can deny those 6 times Sachin had bowled magnificently.

    Any views or comment on this analysis for India?

  • nigel on December 1, 2008, 11:18 GMT

    The ODI spell that stands out in my mind is when Bichel took 7 for 20 in the 2003 World Cup. Every time Ponting turned to him he seemed to get a wicket whereas no one else even looked like getting a wicket. Where does that rank in this analysis?? [[ Right on top. It is 26th (out of over 1000 entries0. Bichel's collection of batsmen was outstanding (171 pts). Only the fact that he dismissed most batsmen after they scored runs (see below) pushed it slightly lower. Otherwise this would have been in the top-10. Total: 89 points. Knight: 30 Vaughan: 2 Hussain: 1 Stewart: 46 Vollingwood: 10 Flintoff: 45 and Giles.

    Ananth: ]]

  • karthik on November 30, 2008, 8:36 GMT

    i only have a small doubt .. u ve considered the average of the batsman dismissed ... wat i want to know is it his avg b4 tat match or is it his current average ...tat is if sachin played a game in 1994 his avg may only be around 38 while now its 44 + .. so i think if u are going to consider a bowling spell in 1994 then the avg during tat match must be considered .. if u have taken tat average only then i am sorry

  • Vidhya on November 23, 2008, 13:43 GMT

    About the suggestion that Keyur made, ranked 325 is Richard Hadlee's 9-1-32-5 against India at Perth in 1980-81. Hadlee dismissed Gavaskar, Viswanath, Sandeep Patil, Kapil Dev and Ghavri.

    The interesting thing about this is that Hadlee bowled four spells, and got a wicket with the first ball of each one, in addition to getting Kapil for a second ball duck. If you consider how early the bowler dismissed the batsman, it would be hard to beat this one. [[ Given below are the details for this match. 1. Gavaskar S.M 0 [Rpi:30.31] (Avg:35.14 - +35.14) 4. Viswanath G.R-23-10 [Rpi:15.00] (Avg:19.95 - + 9.95) 6. Patil S.M 39 [Rpi:23.93] (Avg:24.51) 7. Kapil Dev N 0 [Rpi:19.11] (Avg:23.79 - +23.79) There are many reasons why this is not necessarily that great an analysis. 1. Only 4 top-order and 3 below-avge. 2. None of the batsman was top-drawer (Gavaskar included) Vishewanath had a low average anyhow and was adjusted downwards to an RPI of 15.00. The information you have provided is interesting but not available and cannot anyhow be used. Thanks for identifying an interesting instance. Ananth: ]]

  • Sumit Sanghai on November 23, 2008, 3:30 GMT

    I think you should use the very good D/L method (along with batsman's averages and SRs as input) to do this calculation. DL can take situations into account and predict what's the team score would be before and after the wicket. For example, assume India is batting with SRT as one of the openers. At 0/0 DL can take all the team player's averages and S/Rs to determine the avg. team score. If SRT fails, i.e. at 0/1 you can recalibrate D/L with the remaining batsmen's avg and SRs to determine how much a team gained by SRT's dismissal. This way you can decide how much a bowler helped his team's cause by taking wickets. Finally, subtract the number of runs he gave and you have the final metric. [[ Sunit, I think what you have suggested is too complex for what is a simple calculation. I am not also sure whether D/L is the best and totally correct. Let us accept that the spinners will have a raw deal as far this sub-analysis is concerned. Hopefully some of these problems will be redressed in the complete analysis. Thanks Ananth: ]]

  • keyur on November 21, 2008, 15:13 GMT

    There has been a lot of controversy regarding the "when" factor in your analysis disfavouring the spinners as compared to the fast bowlers because the batsman mostly start at 0 against fast bowlers while they have made up some runs before the spinners come on reducing the "when" points for the spinner. How about considering the score of batsman at the time of the bowler starting his spell? This means that the opening bowlers get the same points as earlier, but the spinner too gets a chance to rake up the same points if he gets the batsman out early in his spell without allowing them to add any runs DURING HIS SPELL. [[ I am very appreciative of the originality of the suggestion. It is not often that we get a suggestion of this quality. There is no doubt that, if implemented, this will considerably improve the "when" factor. However, as I have pointed out earlier, the problem is that this will require complete ball-by-ball data which is either not available for most matches or availabe in an inaccessible proprietary fomr. This does not take away anything from the excellent suggestion. Thank you. Ananth: ]]

  • Ram on November 21, 2008, 11:28 GMT

    Ananth,

    grd analysis.. The last 10 years the nature of the pitches used for one day matches have been made batsmen friendly.. So an average batsman's average of this period will be more inflated compared to a greater batsman of the earlier years... suggest taking this into consideration when you adjust the averages over the years..How is big question...

  • Ross on January 17, 2009, 9:55 GMT

    The BCs had 2 for them and none against, we must be ok... :P

  • Allen Johnston on January 2, 2009, 4:16 GMT

    In terms of the great ODI bowling spells, it must surely be hard to beat the ridiculous one by Phil Simmons vs Pakistan in 1992. Simmons took 4 wickets, for 3 runs in ten overs. 3 Runs in 10 overs! Simmons scalps consisted of Javed Miandad, Aamer Sohail, Saleem Malik and Asif Mujtaba - all in the top 5 of the batting order.

    The West Indies managed to score 214 runs on the pitch.

    (3 runs!!) [[ Yes. Symond's performance is great. It is in the Top-50 with very respectable tally of 235 points. Ananth: ]]

  • Allan Pinchen on December 4, 2008, 9:18 GMT

    Great stuff!! Waqar Younis is about the best I`v`e ever seen,I used to love those inswinging yorkers.By the way I`m Aussie`[58yrs] and loved D K Lillee,he was the best!! Waqar at his peak was unplayable and a great bloke..loved to watch him!

  • Uday Trivedi on December 1, 2008, 11:32 GMT

    One interesting observation on full list data. for India, Agarkar is the top most bowler with 12 entries in the list. In bracket I have put total wks the bowler has taken till now.

    Agarkar - 12 (288) Kumble - 10 (334) Srinath - 10 (315) Zaheer - 8 (205) Prabhakar - 6 (157) Sachin - 6 (152) Pathan - 5 (148) Kapil - 4 (253) Harbhajan - 4 (200) Prasad - 4 (196) Nehra - 4 (90)

    Agarkar has 12 outstanding spells for India,he is no. 3 is all time ODI bowlers list for India, still he didn't get the credit he deserves. I remember during 1997 to early 2000s, Agarkar used to get bagful wickets (with albeit bit high number of runs).For me,he is India's top ODI bowler with his outstanding strike-rate (32.9 - beaten only by Sreesanth's 32.6) and average (27.85 - beaten only by Kapil's 27.45).

    One surprise is Sachin ahead of Kapil & Harbhajan. However,no one can deny those 6 times Sachin had bowled magnificently.

    Any views or comment on this analysis for India?

  • nigel on December 1, 2008, 11:18 GMT

    The ODI spell that stands out in my mind is when Bichel took 7 for 20 in the 2003 World Cup. Every time Ponting turned to him he seemed to get a wicket whereas no one else even looked like getting a wicket. Where does that rank in this analysis?? [[ Right on top. It is 26th (out of over 1000 entries0. Bichel's collection of batsmen was outstanding (171 pts). Only the fact that he dismissed most batsmen after they scored runs (see below) pushed it slightly lower. Otherwise this would have been in the top-10. Total: 89 points. Knight: 30 Vaughan: 2 Hussain: 1 Stewart: 46 Vollingwood: 10 Flintoff: 45 and Giles.

    Ananth: ]]

  • karthik on November 30, 2008, 8:36 GMT

    i only have a small doubt .. u ve considered the average of the batsman dismissed ... wat i want to know is it his avg b4 tat match or is it his current average ...tat is if sachin played a game in 1994 his avg may only be around 38 while now its 44 + .. so i think if u are going to consider a bowling spell in 1994 then the avg during tat match must be considered .. if u have taken tat average only then i am sorry

  • Vidhya on November 23, 2008, 13:43 GMT

    About the suggestion that Keyur made, ranked 325 is Richard Hadlee's 9-1-32-5 against India at Perth in 1980-81. Hadlee dismissed Gavaskar, Viswanath, Sandeep Patil, Kapil Dev and Ghavri.

    The interesting thing about this is that Hadlee bowled four spells, and got a wicket with the first ball of each one, in addition to getting Kapil for a second ball duck. If you consider how early the bowler dismissed the batsman, it would be hard to beat this one. [[ Given below are the details for this match. 1. Gavaskar S.M 0 [Rpi:30.31] (Avg:35.14 - +35.14) 4. Viswanath G.R-23-10 [Rpi:15.00] (Avg:19.95 - + 9.95) 6. Patil S.M 39 [Rpi:23.93] (Avg:24.51) 7. Kapil Dev N 0 [Rpi:19.11] (Avg:23.79 - +23.79) There are many reasons why this is not necessarily that great an analysis. 1. Only 4 top-order and 3 below-avge. 2. None of the batsman was top-drawer (Gavaskar included) Vishewanath had a low average anyhow and was adjusted downwards to an RPI of 15.00. The information you have provided is interesting but not available and cannot anyhow be used. Thanks for identifying an interesting instance. Ananth: ]]

  • Sumit Sanghai on November 23, 2008, 3:30 GMT

    I think you should use the very good D/L method (along with batsman's averages and SRs as input) to do this calculation. DL can take situations into account and predict what's the team score would be before and after the wicket. For example, assume India is batting with SRT as one of the openers. At 0/0 DL can take all the team player's averages and S/Rs to determine the avg. team score. If SRT fails, i.e. at 0/1 you can recalibrate D/L with the remaining batsmen's avg and SRs to determine how much a team gained by SRT's dismissal. This way you can decide how much a bowler helped his team's cause by taking wickets. Finally, subtract the number of runs he gave and you have the final metric. [[ Sunit, I think what you have suggested is too complex for what is a simple calculation. I am not also sure whether D/L is the best and totally correct. Let us accept that the spinners will have a raw deal as far this sub-analysis is concerned. Hopefully some of these problems will be redressed in the complete analysis. Thanks Ananth: ]]

  • keyur on November 21, 2008, 15:13 GMT

    There has been a lot of controversy regarding the "when" factor in your analysis disfavouring the spinners as compared to the fast bowlers because the batsman mostly start at 0 against fast bowlers while they have made up some runs before the spinners come on reducing the "when" points for the spinner. How about considering the score of batsman at the time of the bowler starting his spell? This means that the opening bowlers get the same points as earlier, but the spinner too gets a chance to rake up the same points if he gets the batsman out early in his spell without allowing them to add any runs DURING HIS SPELL. [[ I am very appreciative of the originality of the suggestion. It is not often that we get a suggestion of this quality. There is no doubt that, if implemented, this will considerably improve the "when" factor. However, as I have pointed out earlier, the problem is that this will require complete ball-by-ball data which is either not available for most matches or availabe in an inaccessible proprietary fomr. This does not take away anything from the excellent suggestion. Thank you. Ananth: ]]

  • Ram on November 21, 2008, 11:28 GMT

    Ananth,

    grd analysis.. The last 10 years the nature of the pitches used for one day matches have been made batsmen friendly.. So an average batsman's average of this period will be more inflated compared to a greater batsman of the earlier years... suggest taking this into consideration when you adjust the averages over the years..How is big question...

  • ABC on November 20, 2008, 9:01 GMT

    Ananth, What a reply to Mr Rana's Comment.

  • Mudassar Rana on November 20, 2008, 4:28 GMT

    Whatever additions that could be added to make this research more accurate cant take away that Waqar Younis is probably the finest bowler of the modern era. With so many spells in the top 50 this research is confirmation of that. The shocker was that s.joshi made the list! Factors that possibly we could never account for are such as what type of pitch it was. for example for waqar to get so many wickets on sub continent pitches is a greater acheievment than say him getting them on australian pitches.A great effort and very enlightening.look forward to the next model! [[ Joshi's inclusion will not be a shocker if enough faith is vested in the analysis, which has placed Pakistani bowlers on top, and accept that Joshi's collection of top South African wickets allows him to be placed at the top. Ananth: ]]

  • Ananth on November 20, 2008, 1:42 GMT

    A common response: 1. Possible adjustment of Batting Averages will be needed and will be incorporated in the complete analysis. 2. I expected to see this table dominated by pace bowlers but was surprised to see the extent of domination. However readers should note that there was no determined effort to keep spinners out. 3.Taken in context, a dismissal of a low-avge batsman at 100 would be far more relevant than a high-avge batsman at 0. However this analysis does not, as deliberately planned, take into account context. 4.The idea was to do this major sub-analysis, invite reader's comments and then do the full-blown analysis of ODI bowling innspells. It is working out as I already have a good pulse of the readers.

  • pritham on November 20, 2008, 0:56 GMT

    Hi Ananth, Gr8 analysis. As is always the case...u rock. it would have been helpful if you had the batsman's strike rate into account and also if possible remove the easy runs scored against associates and weak attacks. I believe Vaas at 2 is bit over hyped, not your fault clearly but sometimes we feel Mark Twain had some truth. Thanks Again

  • David Barry on November 20, 2008, 0:01 GMT

    Ananth, in your reply to Vidhya you say that you're not convinced that averages have increased over the years. This one's pretty easy to show - Statsguru can give you overall averages by year. The trend is obviously upward, from mid- to high-twenties through the 1980's, about 28 by 1990, and then a fairly steady rise to 31 or 32 today.

  • Tushar on November 19, 2008, 23:53 GMT

    As, you rightly said, lets keep the result and rpo aside. 1. Averages Sachin 44.33 Robin 25.95 Kambli 32.59 Yuvraj 37.19 Badani 33.34 Dahia 16.61 ------ Total:190

    Thats the numbers, and they should speak better than the names (at least in this column). Having said that we can anyday pit them against the zimbambeans who took Vaas to No.2 in the list. or even against the England batting order including Night, Stewart, Shah and Cork that put Waqar on top.

    2. Sachin's 61 - Vaas, Zoysa and Upashantha had allowed him to race to (if my assumtions and calculations are right) around 55, so Murali allowed him to score only 6 more before getting him out. Should this not give him some good 'When' points.

    Its not fast v/s spin, but I cannot stop myself from repeating that spinners are more likely to miss out on When points (using the method you have used) just because they rarely share the new ball.

    Thanks for this really good analysis

    - Tushar Kardile

  • Charles Davis on November 19, 2008, 23:36 GMT

    While analysis that rates bowlers according to batsmen dismissed can be interesting, be aware that this loads the dice massively in favour of pace bowlers over spinners. I think there is only one (relatively little-known) spinner in the Top 50. Murali, with 479 ODI wickets, comes in at # 89, and Warne not in the Top 100, although both these bowlers are in the all-time Top 10 in terms of wickets per match (min. 100 wickets).

    In short, spinners get less chance to bowl to top-order batsmen who are out of form when they are fresh to the crease. You will probably find that most of the wickets in the above lists fell in the first 15 overs, when spinners are little-used because of fielding restrictions.

  • Usman on November 19, 2008, 23:02 GMT

    I really appreciate your efforts. One interesting factor that i observed that Waqar Younis has quite a few spells in this list which certainly is amazing achievement.

  • Rishi on November 19, 2008, 23:00 GMT

    Hi Ananth, Enjoyed it. and remembered some good and bad times while watching those performances. Although you seem to think that batting averages havent changed, the fact that there are only 3 bowling performances from the 1980s in top 90 results seem to indicate that it is not true. Unless (all other things being equal) total one day matches played in 80s are 1/30th of the games played since 1990-2008. perhaps..that is possible..can you look that up? If that's true, i would buy that batting avgs havent changed over time.

  • saurabh somani on November 19, 2008, 18:09 GMT

    agree with the first comment made by sushamna and jeff's two points above... but then ananth has stated that this is but a sub-analysis and these points will be incorporated in the full analysis - so am looking forward to the full analysis! for the first time i'm also able to whole-heartedly agree with ananth on a point - the picking of batsmen from 1-8!

  • Tushar on November 19, 2008, 16:00 GMT

    "Who" and "Real When" - Addendum to my previous post

    In a IND-SL game, for which I could not find detailed commentary, but following extract from the wisden report proves the case:

    [ Muralitharan began his opening spell of four for 21 in seven overs, including Tendulkar for a 54-ball 61. While he rested, Badani and Dahiya rallied India's reply, but his second spell, culminating in a double-wicket maiden, reduced them to 211 for nine, leaving Jayasuriya to complete the victory. ]

    6 of his 7 wickets were top 8, giving him 190 "Who" points and I guess not more than 50 runs were given (by both ends) since his first spell started and while he was bowling TILL THE TIME HE CLAIMED HIS 6th WICKET. Giving him about 140 "When" points.

    This will put him very high in the list, but unfortunately this grand performance lies 89th in your list

    --Tushar Kardile

    [[ Tushar, Let us agree to leave the result, rpo et al out, and look at Murali's wickets. Tendulkar (after scoring 61), Robin Singh, Kambli, Yuvraj, Badani, Dhahia abd Joshi. Certainly not an impressive collection, barring Tendulkar. Compare this with the collection of any of the top 25 bowlers in the table. Ananth: ]]

  • Tushar on November 19, 2008, 15:33 GMT

    [ Ananth : At least the early bowlers did not get credit for these batsmen !!! ] - Thanks mate, but my point was - A spinner coming in to bowl 11th over of the game with openers playing on 40 each and team on 80/0, and claiming 3 wickets in 3 overs. coming back in 25th over with score at 150/3 and no.4 and no.5 batting on 35* each and claiming both of them in the first over he bowled, and not taking any more wickets... he will get (suppose) 40 * 5 = 200 "Who" points (assuming top 5 averaging 40 each), and only 50 "When" points (assuming no 3 dismissed on 0). Now, this is no mean feat if you compare it to an opening bowler claiming top 5 for 10 apiece in his first 5 overs, but the opening bowler in this equation collects 200 "Who" points and 150 "When" points.

    So even if "At least the early bowlers did not get credit for these batsmen" in that particular game, in general early bowlers claiming wickets in their first few overs always have (unfair???) advantage over later bowlers. [[ Your illustration is very good. However there is no denying that a fast bowler capturing 2/3 early wickets helps his team a lot. Pl remember this is not a fast vs spin analysis. This is only an analysis of bowlers who took top order wickets and when they took these wickets. To do a complete job one would have to analyze each dismissal from various points. Suppose Pakistan is chasing 300 and is at 290 for 9 in the 49th over with Afridi batting on 100. The bowler who takes Afridi's wicket in the 49th over deserves a million points but would get very little in this analysis. Ananth: ]]

  • sridhar on November 19, 2008, 14:21 GMT

    would i be right in saying that this analysis is only for fast bowlers. Considering sunil joshis figures as an exception. I could not spot one other spinner in the top 50 in your list. I think thats an interesting point. Given that the top three highest wicket takers are all world class spinners. Not one of their spells figure in the top hundred when the "quality of wickets" is considered!!!! [[ This certainly did not start as an analysis of fast bowlers. It just panned out that way. At no stage have I mentioned that these are the best bowling innspells of all time. To arraive at that requires lot more work including bowling accuracy. As and when I do a complete analysis I am sure more spinners will move up. Ananth: ]]

  • Jeff on November 19, 2008, 13:44 GMT

    Apologies for over-looking your original comment on wickets in the last over.

    I understand that, for the scope of this particular analysis, what you have done is fine.

    My comments were purely suggestions for what you might want to consider in any follow up study.

    Having thought a bit more, if you wanted to be really thorough, you might also want to consider the batting partners available to batsmen when they are dismissed.

    For example, if you dismiss Tendulkar for 20 runs, then on average you might expect him to score another 30 or 40 (if he had enough overs remaining to do this in.)

    But what if Tendulkar was the 9th wkt of the inns and the only man left to bat with him was Ishant Sharma (or whoever the relevent Indian number 11 is/was?)

    If Sharma lasts on average 10 balls per innings (not sure in reality what this figure is) then how many balls would Tendulkar have the chance of facing? 10? 20 if he farms the strike? Not much chance of scoring 30 or 40 in that case.

  • Jeff on November 19, 2008, 12:57 GMT

    (Cont.)

    2. Consider "When" in terms of which overs the wickets were taken. For example, it's no good dismissing Hussey for a duck off the last ball of the innings and then saying this is worth 50-odd runs. Obviously the most it could be worth is 6 and probably not much more than 1 run on average.

    One way to do this is to factor in strike rate as well as average.

    For example, if you dismiss Hussey for nought in the 40th over, then he would probably have missed out on 30 balls (roughly half of those remaining.) His SR is about 85, so the value of the wicket is about 26.

    Cheers, Jeff [[ Ah I knew this and maybe that is why I correctly decided to do the more comprehensive analysis later after posting a subset analysis and getting the readers' comments. Both are excellent points. I have already touched on picking up Tendulkar's wicket in the last over. The problem, Jeff, is that for a large number of early matches, only FoW information is available and not "Who was dismissed". However let me see what can be done. Ananth: ]]

  • Jeff on November 19, 2008, 12:49 GMT

    Hi Ananth,

    I enjoyed the analysis - in fact, this is one of those occasions when I think there is real merit in doing one of your "full-blown" meta analyses..

    I have a couple of comments (which you might want to consider if expanding the work...)

    1. Consider not the difference between the actual score and the average, but how many MORE runs the batsman is likely to score. For example, a batsman who averages 40, will have no doubt made lots of low scores (as David mentioned, it's easier to get out early) which bring that average down. But once the player has made (say) 20 runs, then how many more is he likley to make - ie. in every inns that he's reached 20, what is the average number of extra runs that he goes on to score?

    I realise this is complicated work - I also seem to remember another blog on this site (Charles Davis?) doing something similar with test match batting.

    I will post my 2nd comment in a further message...

  • Tushar on November 19, 2008, 12:27 GMT

    Ananth, This one is a really meaningful analysis. This addresses most of the gripes I have had about the bowling figures. Only one situation this does not handle (IMHO) is when a bowler starts his (first or subsequent) spell mid-innings and a pair of batsmen have already accumulated some runs against other ordinary (or unfortunate) bowlers, and gets quick wickets. Almost as effective as an opening bowler taking wickets early on in the innings. But losing out on a lot of "When" points just because there has been some runs on the board already. [[ At least the early bowlers did not get credit for these batsmen !!! Ananth: ]]

  • Richard Mackey on November 19, 2008, 12:27 GMT

    Interesting analysis - but it doesn't take much account of the runs conceded by the bowler, which is key to the final match result.

    For example, if in the first result Waqar had gone for an additional 30 runs, but they were all scored by a batsman he didn't dismiss (or wides or no-balls), the analysis wouldn't be affected at all. Perhaps you could tweak the analysis to take account of economy rate? [[ If you re-visit the article, I have specifically mentioned that I have not considered the bowling accuracy et al. As and when I do a complete analysis of innspells to determine the greatest bowling innsspell ever, this factor will certainly be included. Ananth: ]]

  • David Barry on November 19, 2008, 12:19 GMT

    While it's certainly more useful to dismiss batsmen early in their innings, it should be noted that it's much easier to get them out early, and so it is more impressive to dismiss a batsman who's already set.

    In particular, it's more than twice as easy to dismiss a batsman on nought than when he's got past the score of 3, by which time most of the 'getting your eye in' has taken place. (Based on ODI batsmen who average > 35 in the last ten years.) [[ David, it is true what you say. However the benefit of taking a top batsman or two or seven early cannot be denied. Ananth: ]]

  • Vidhya on November 19, 2008, 12:18 GMT

    Interesting analysis and not much to complain about.

    As for the "when" factor - this is one place where some fairness could be added. The spinners come on to bowl usually when the batsmen facing them already have a few runs against their name, unlike the opening bowlers. I hope that over the long run this gets adjusted because spinners tend to bowl one or two spells, while the faster men bowl two or three.

    Still this advantage is visible at the top - Waqar and Vaas opened the bowling and took all their wickets in a single spell.

    Due to the flatness of the modern wickets, the batsmen tend to average more. Shastri-Srikkanth-Vengsarkar-Gavaskar-Amarnath is a far better line up Dippenaar-Gibbs-Cronje-Rhodes-Pollock, yet Joshi's spell is rated higher than Imran's. Can some adjustment be done for this ? [[ I am not that convinced that the averages have increased over the years. I agree that the strike rates have increased. If you recall my earlier analysis the strike rates showed an increase of 15% while the averages less than 5%. The early batsmen, barring a few, tended to bat longer albiet slowly. It is also true that the pace bowlers will get the benefit in this analysis. And if they dismiss the top order batsmen for low scores, they shoul. Ananth: ]]

  • Shafiq on November 19, 2008, 11:08 GMT

    Nice to see 6 Pakistani's spells in top 15. i personally rate IMRAN's magic in sharjah dust very high agaist quality Indians. Excellent work----and it is amazing wasim is not in top 10. [[ Ananth: Pakistan bowling over the past 20 years has been the most effective and incisive of all. That is reflected in these tables. Wasim Akram's best has been his 5 for 15 against (a rather weak) Australia. 5 top order wickets. But these are Wessels, Kerr, Jones, Border and Hughes. Hence this has been placed at no.35, still a very respectable position. ]]

  • Sushamna Deodhar on November 19, 2008, 10:36 GMT

    I've a doubt with Rpi / Avg stats used here. The averages considered are player's average to date. However, for instance, when Aaqib Javed took Sachin's wicket in 1991, his average was much less than 44.xx (was under 30s I guess, that's why he used to come at 3-down). The same goes with Rpi. To incorporate this for all batsmen is quite a difficult and complex task, but that will include player's actual statistics when the match took place, rather than what it is now, say after 17 years. How about that? [[ Ananth: Your point is valid. However as you yourself have observed it is a rather complex task and I would attempt it if I was doing serious Rating work. But this is a one-off analysis. The other point is to look at it from the bowler's point of view, not the batsman's. Aaqib Javed 7 wickets, one of which was Tendulkar, average then 35 and now 44. Similarly there could be some other batsman who might have moved the other way. Consider this only as a pointer to top class incisive spells. ]]

  • No featured comments at the moment.

  • Sushamna Deodhar on November 19, 2008, 10:36 GMT

    I've a doubt with Rpi / Avg stats used here. The averages considered are player's average to date. However, for instance, when Aaqib Javed took Sachin's wicket in 1991, his average was much less than 44.xx (was under 30s I guess, that's why he used to come at 3-down). The same goes with Rpi. To incorporate this for all batsmen is quite a difficult and complex task, but that will include player's actual statistics when the match took place, rather than what it is now, say after 17 years. How about that? [[ Ananth: Your point is valid. However as you yourself have observed it is a rather complex task and I would attempt it if I was doing serious Rating work. But this is a one-off analysis. The other point is to look at it from the bowler's point of view, not the batsman's. Aaqib Javed 7 wickets, one of which was Tendulkar, average then 35 and now 44. Similarly there could be some other batsman who might have moved the other way. Consider this only as a pointer to top class incisive spells. ]]

  • Shafiq on November 19, 2008, 11:08 GMT

    Nice to see 6 Pakistani's spells in top 15. i personally rate IMRAN's magic in sharjah dust very high agaist quality Indians. Excellent work----and it is amazing wasim is not in top 10. [[ Ananth: Pakistan bowling over the past 20 years has been the most effective and incisive of all. That is reflected in these tables. Wasim Akram's best has been his 5 for 15 against (a rather weak) Australia. 5 top order wickets. But these are Wessels, Kerr, Jones, Border and Hughes. Hence this has been placed at no.35, still a very respectable position. ]]

  • Vidhya on November 19, 2008, 12:18 GMT

    Interesting analysis and not much to complain about.

    As for the "when" factor - this is one place where some fairness could be added. The spinners come on to bowl usually when the batsmen facing them already have a few runs against their name, unlike the opening bowlers. I hope that over the long run this gets adjusted because spinners tend to bowl one or two spells, while the faster men bowl two or three.

    Still this advantage is visible at the top - Waqar and Vaas opened the bowling and took all their wickets in a single spell.

    Due to the flatness of the modern wickets, the batsmen tend to average more. Shastri-Srikkanth-Vengsarkar-Gavaskar-Amarnath is a far better line up Dippenaar-Gibbs-Cronje-Rhodes-Pollock, yet Joshi's spell is rated higher than Imran's. Can some adjustment be done for this ? [[ I am not that convinced that the averages have increased over the years. I agree that the strike rates have increased. If you recall my earlier analysis the strike rates showed an increase of 15% while the averages less than 5%. The early batsmen, barring a few, tended to bat longer albiet slowly. It is also true that the pace bowlers will get the benefit in this analysis. And if they dismiss the top order batsmen for low scores, they shoul. Ananth: ]]

  • David Barry on November 19, 2008, 12:19 GMT

    While it's certainly more useful to dismiss batsmen early in their innings, it should be noted that it's much easier to get them out early, and so it is more impressive to dismiss a batsman who's already set.

    In particular, it's more than twice as easy to dismiss a batsman on nought than when he's got past the score of 3, by which time most of the 'getting your eye in' has taken place. (Based on ODI batsmen who average > 35 in the last ten years.) [[ David, it is true what you say. However the benefit of taking a top batsman or two or seven early cannot be denied. Ananth: ]]

  • Richard Mackey on November 19, 2008, 12:27 GMT

    Interesting analysis - but it doesn't take much account of the runs conceded by the bowler, which is key to the final match result.

    For example, if in the first result Waqar had gone for an additional 30 runs, but they were all scored by a batsman he didn't dismiss (or wides or no-balls), the analysis wouldn't be affected at all. Perhaps you could tweak the analysis to take account of economy rate? [[ If you re-visit the article, I have specifically mentioned that I have not considered the bowling accuracy et al. As and when I do a complete analysis of innspells to determine the greatest bowling innsspell ever, this factor will certainly be included. Ananth: ]]

  • Tushar on November 19, 2008, 12:27 GMT

    Ananth, This one is a really meaningful analysis. This addresses most of the gripes I have had about the bowling figures. Only one situation this does not handle (IMHO) is when a bowler starts his (first or subsequent) spell mid-innings and a pair of batsmen have already accumulated some runs against other ordinary (or unfortunate) bowlers, and gets quick wickets. Almost as effective as an opening bowler taking wickets early on in the innings. But losing out on a lot of "When" points just because there has been some runs on the board already. [[ At least the early bowlers did not get credit for these batsmen !!! Ananth: ]]

  • Jeff on November 19, 2008, 12:49 GMT

    Hi Ananth,

    I enjoyed the analysis - in fact, this is one of those occasions when I think there is real merit in doing one of your "full-blown" meta analyses..

    I have a couple of comments (which you might want to consider if expanding the work...)

    1. Consider not the difference between the actual score and the average, but how many MORE runs the batsman is likely to score. For example, a batsman who averages 40, will have no doubt made lots of low scores (as David mentioned, it's easier to get out early) which bring that average down. But once the player has made (say) 20 runs, then how many more is he likley to make - ie. in every inns that he's reached 20, what is the average number of extra runs that he goes on to score?

    I realise this is complicated work - I also seem to remember another blog on this site (Charles Davis?) doing something similar with test match batting.

    I will post my 2nd comment in a further message...

  • Jeff on November 19, 2008, 12:57 GMT

    (Cont.)

    2. Consider "When" in terms of which overs the wickets were taken. For example, it's no good dismissing Hussey for a duck off the last ball of the innings and then saying this is worth 50-odd runs. Obviously the most it could be worth is 6 and probably not much more than 1 run on average.

    One way to do this is to factor in strike rate as well as average.

    For example, if you dismiss Hussey for nought in the 40th over, then he would probably have missed out on 30 balls (roughly half of those remaining.) His SR is about 85, so the value of the wicket is about 26.

    Cheers, Jeff [[ Ah I knew this and maybe that is why I correctly decided to do the more comprehensive analysis later after posting a subset analysis and getting the readers' comments. Both are excellent points. I have already touched on picking up Tendulkar's wicket in the last over. The problem, Jeff, is that for a large number of early matches, only FoW information is available and not "Who was dismissed". However let me see what can be done. Ananth: ]]

  • Jeff on November 19, 2008, 13:44 GMT

    Apologies for over-looking your original comment on wickets in the last over.

    I understand that, for the scope of this particular analysis, what you have done is fine.

    My comments were purely suggestions for what you might want to consider in any follow up study.

    Having thought a bit more, if you wanted to be really thorough, you might also want to consider the batting partners available to batsmen when they are dismissed.

    For example, if you dismiss Tendulkar for 20 runs, then on average you might expect him to score another 30 or 40 (if he had enough overs remaining to do this in.)

    But what if Tendulkar was the 9th wkt of the inns and the only man left to bat with him was Ishant Sharma (or whoever the relevent Indian number 11 is/was?)

    If Sharma lasts on average 10 balls per innings (not sure in reality what this figure is) then how many balls would Tendulkar have the chance of facing? 10? 20 if he farms the strike? Not much chance of scoring 30 or 40 in that case.

  • sridhar on November 19, 2008, 14:21 GMT

    would i be right in saying that this analysis is only for fast bowlers. Considering sunil joshis figures as an exception. I could not spot one other spinner in the top 50 in your list. I think thats an interesting point. Given that the top three highest wicket takers are all world class spinners. Not one of their spells figure in the top hundred when the "quality of wickets" is considered!!!! [[ This certainly did not start as an analysis of fast bowlers. It just panned out that way. At no stage have I mentioned that these are the best bowling innspells of all time. To arraive at that requires lot more work including bowling accuracy. As and when I do a complete analysis I am sure more spinners will move up. Ananth: ]]