Tests - bowling July 21, 2009

An in-depth analysis of Test bowlers

After the comprehensive analysis on Test batsmen, it's the turn of the bowlers to be put under the scanner
101

At last I have been able to finish the second part of the analytical review on great Test players. The three-part analysis on Test Batsmen generated well over 1000 comments and was, in general, well received and accepted. No analysis would satisfy all and this may also be true in the on-going analysis of Test bowlers.

I have learnt a lot through the Test Batsmen analysis. First and foremost is that doing a single comparison table over 134 years is not the correct method. Test cricket has changed probably 1080 degrees over the years and there cannot be a single yardstick for all the players. Hence I have separated the analysis into multiple periods.

Period Separation:

These periods have been identified with lot of thought and deliberation with inputs from a few interested readers. Many related factors have gone into this process. Separate tables will be prepared for different periods. In addition, I will show, in the follow-up article, two tables separating the bowlers by type of bowling. This will be only for information.

- The bowling era: 1877-1914 (134 Tests and 370 players)
- The batting era: 1920-1969 (535 Tests and 980 players)
- The balanced era: 1970-2009 (1251 Tests and 1220 players).

The first era is so different from the rest of the years that it is essential to separate it into a single one despite the paucity of Tests. Uncovered pitches, 3-day Test matches, 110+ overs bowled in a day, compulsory follow-ons, low average scores et al are some of the features.

The second era was where batting was king. However, the in-between wars period was lit up by the wonderful batting of Bradman, Hammond. Headley, McCabe et al and was the golden era of batting. Still the results were plentiful. What followed the WW-2 was unfortunate. These years were batting dominated. However the batting was defensive and the matches were driven by the desire not to lose, rather than to win. The new teams, India and Pakistan, the weaker New Zealand and the defensive strong teams contributed a lot to this situation. These 50 years form a separate era. There are lot of similarities within the two sub-periods in terms of numbers.

The third era is the most balanced era of all. This era saw great bowlers such as Lillee, Holding, Marshall, Hadlee, Imran, Wasim Akram, Waqar Younis, Kapil Dev, Muralitharan, Warne, Kumble et al. It also saw the presence of great batsmen such as Richards, Greg Chappell, Gavaskar, Tendulkar, Lara, Ponting, Miandad, Dravid, Gooch, Jayawardene et al. Thus there were great contests. As such this was a great balanced era and even though the number of Tests is quite high, this is a logical grouping.

As done for the Batting analysis, the analysis is done in two parts. The first is based on Match Performances and the second part is based on the Career achievements. Many people are under the misapprehension that Match Performance is based on team achievements. This is completely wrong. The Match Performance refers to the concerned bowlers' performances during the specific match and what happened in the match. The only team achievement considered is the result which, at the end of the day, is the most important aspect of any match.

A. Match Performances (Maximum 40 points)

The following factors are used to analyze the match performances of bowlers. The total points secured is divided by the number of innspells (my own term indicating a qualifying bowling stint, taking care to exclude bowling efforts such as 5-0-17-0 et al).

Base points
- Wickets captured
- Balls bowled - to recognize long spells
- Batsmen dismissed - based on his score at time of dismissal
Multiplicative factors
- Overall quality of batting team (primarily top-7 batsmen)
- Bowling accuracy (relative to the innings scoring rate)
- Match-related pitch characteristics
- An adjustment for pace bowlers bowling in the Asian subcontinent and spinners bowling outside
- Match situation
- Home/Away (incorporating relative team strengths)
- Result (incorporating relative team strengths)
- Series situation

B. Career Achievements (Maximum 40 points)

This is an equally important aspect of any such analysis. It also encompasses aspects of bowling which do not require consideration of the match conditions or situation. The only longevity measure is the "Career wickets captured" measure, carrying 5 points (6.2%). This will incorporate the following factors.

- Career wickets captured (5 points)
- Career wickets per innspell (5 points)
- Bowling Strike rate-BpW (10 points)
- Bowling accuracy-RpO (5 points)
- Average Quality of batsmen dismissed - based on CtD bat avge (5 points)
- Type of wickets captured - Top/Middle order/Late order (5 points)
- Performance ratio - % of wickets captured to % of balls bowled (5 points).

C. Match Performances(Maximum 40 points)

1.1. Wickets captured: Straightforward linear weight for wickets captured.
1.2. Balls bowled: This is to recognize the fact that a bowler might have bowled an innspell of 43-12-69-2 and provided great support to the main strike bowler(s). Around 25-over spell is considered as approximately equivalent to a wicket.
1.3. Batsmen dismissed: This is to take care of situations such as the Cardiff/Lord's Tests. The idea is to reward Anderson who dismissed Ponting at 0 as against Panesar who dismissed him at 150. Anderson gets almost complete credit while Panesar none. The importance of dismissing a top batsman at a low score cannot be over-emphasized. However it must be noted that in the Career Batsman quality measure, both Anderson and Panesar would get credit for 56.18.
2.1. Overall quality of batting team: This is based on the Career-todate batting averages of the first 7 batsmen and minimal weight to the late order batsmen.
2.2. Bowling accuracy: This is in relation to the bowling team's overall innings performance. three recent examples shown.
- Saf: 651 in 154.3 (Siddle 35-15-67-1)
- Nzl: 619 in 154 (Harbhajan 41-7-120-2)
- Ind: 379 in 92 (Franklin 14-4-38-1)
In each of these cases the bowler concerned has done very well as compared to his team mates and will be credited with the appropriate multiplicative factor, Siddle and Harbhajan more than Franklin because of the higher proportion of overs delivered.
2.3. Match-related pitch characteristics: Based on Arjun's suggestion of the 10 best scores. I have done an analysis of many matches of different periods and this measure has come out very well. The highest value is 1319 in the (in)famous Slk-Ind test in which 6 centuries, including Jayasuriya's 340, were scored. The lowest was in an Ashes test during 1888 with a figure of 181, the four innings scores being 116, 53, 60 and 62 (???). The higher this value is, the more difficult the bowlers' task is and vice versa.
2.4. Location based adjustment: All pace bowlers bowling in the sub-continent get a lift up and all spinners bowling outside get a lift up. There is no negative valuation. These are based on actual summary calculations.
2.5. Match situation: The innings type. In the second innings, what score was being defended, in the third innings, what is the deficit/advantage and what was the attempted target score and in the fourth innings, what was the score being defended and what was the margin of win, if there was one.
2.6. Home/Away: No blind computation. This takes into account the relative strengths of the two teams. Weaker teams, whether playing home or away will get additional weight and vice versa.
2.7. Result: Here also the relative strengths are taken into account.
2.8. Series situation: Is it a dead rubber, is the series still in the balance, what is the series score at mid points et al.

D. Career Achievements (Maximum 40 points)

1. Career wickets captured (5 points): Only longevity based measure. 5 points for 1000 wickets.
2. Career wickets per innspell (5 points): Performance based measure.
3. Bowling Strike rate-BpW (10 points): This generally favours the fast bowlers. And that is the way it should be.
4. Bowling accuracy-RpO (5 points): This generally favours the spinners.
5. Average Quality of batsmen dismissed - based on CtD bat avge (5 points): Averaged over all the wickets captured.
6. Type of wickets captured - Top/Middle order/Late order (5 points): The Top/Middle order gets clubbed together and gets much higher weight than the low order and then the average determined.
7. Performance ratio - % of wickets captured to % of balls bowled (5 points). This is to reward the bowlers who have delivered maximum while bowling less. Generally favours the fast bowlers although readers would be surprised to see Stuart Macgill in the top-10.

Let us now look at the tables. The same criteria is used for all periods so the tables are comparable, while exercising a degree of caution. The bowler should have reached the mark of 100 career wickets. The tables are current upto and inclusive of match no. 1924 (Second Sri Lnka - Pakistan Test completed recently).

Before readers rush off with comments let me outline below in a simple manner all factors which have been taken care of. Please do not make redundant comments on these factors.

1. Bowler perf points in stronger bowling teams have been increased.
2. Bowler perf points in weaker bowling teams have been decreased.
3. Bowler perf points against stronger batting lineups have been increased.
4. Bowler perf points weaker batting lineups have been decreased.
5. Pace bowler perf points in subcontinent matches have been increased.
6. Spin bowler perf points in outside-sc matches have been increased.
7. Batsman quality is career-to-date and adjusted based on period.
8. Longevity gets a weight of 6.25% and performance measures 93.75%.
9. Effort put in by bowlers, even supportive, has been recognized.

1. Current era (1970-2000): Table of top bowlers

SNo. Cty Bowler          BT Ratio Total Match  Wkt  Bow  Bow  Wkt  Wkt Perf
Pts  Perf   Pts StRt  Acc  Bat  Qty  Idx
Max Wt-> 80.0  40.0  10.0 10.0  5.0  5.0  5.0  5.0

1. Slk Muralitharan M ROB 1.28 51.30 23.85 6.49 6.74 3.89 4.02 3.81 2.51 2. Aus Lillee D.K RF 1.20 48.05 21.48 3.87 7.62 3.20 4.92 3.98 2.98 3. Aus Warne S.K RLB 1.20 48.00 22.52 5.57 6.47 3.64 3.69 3.61 2.52 4. Nzl Hadlee R.J RFM 1.20 47.97 21.16 4.37 7.69 3.38 4.73 3.88 2.76 5. Pak Imran Khan RF 1.20 47.90 21.41 3.87 7.37 3.46 5.15 3.92 2.72 6. Saf Steyn D.W RF 1.14 45.55 20.34 2.94 8.01 2.72 4.31 3.68 3.55 7. Win Marshall M.D RF 1.14 45.44 18.89 3.77 7.94 3.38 4.59 4.01 2.85 8. Aus McGrath G.D RFM 1.12 44.86 18.77 4.57 7.03 3.81 3.84 4.05 2.79 9. Ind Kumble A RLB 1.11 44.58 20.13 5.08 5.62 3.58 4.13 3.78 2.26 10. Pak Waqar Younis RFM 1.10 44.18 18.67 3.74 7.89 2.91 4.07 3.90 3.00

11. Saf Donald A.A RF 1.10 44.13 18.52 3.61 7.49 3.35 4.01 4.02 3.12 12. Win Ambrose C.E.L RF 1.09 43.55 18.76 3.81 6.90 3.83 4.01 3.96 2.27 13. Win Holding M.A RF 1.08 43.40 17.80 2.94 7.70 3.22 5.06 3.96 2.71 14. Pak Wasim Akram LFM 1.08 43.22 18.90 3.84 6.90 3.57 3.91 3.69 2.41 15. Pak Shoaib Akhtar RF 1.08 43.21 19.12 2.60 7.53 2.93 4.19 3.93 2.91 16. Aus Lawson G.F RF 1.08 43.20 19.26 2.70 6.40 3.12 5.18 4.17 2.37 17. Aus Reid B.A LFM 1.08 43.03 18.55 2.68 6.92 3.42 4.35 4.10 3.00 18. Win Croft C.E.H RF 1.07 42.97 18.20 2.43 7.86 3.15 4.61 4.10 2.61 19. Aus Thomson J.R RF 1.07 42.82 17.32 2.72 7.57 2.79 5.43 4.19 2.78 20. Ind Harbhajan Singh ROB 1.06 42.51 20.26 3.46 5.63 3.59 3.81 3.61 2.14

This is a galaxy of the best bowlers who have graced the grounds over the past 40 years. Not one of them does not deserve his place in this exclusive list. One might like minor moves amongst the top-10, but no one can say with any degree of conviction that there is even one undeserving candidate, including Dale Steyn.

Muralitharan is deservedly on top, that too by a margin of around 6%. The fact that he has played for Sri Lanka has only aided him slightly. His top-drawer performances, day in and day out, have given him the highest Match Performance points. His collection of wickets, wickets per innspell, good accuracy, quality of batsmen dismissed are all in the top 10%. Only in the last two measures does he lag behind others since he has taken a lion's share of his team's bowling efforts and has captured significant number of late order batsmen.

Lillee, who is in second place just ahead of Warne, was the first of the modern great fast bowlers. He formed a great team with Thomson and would have comfortably crossed 450 wickets barring the mid-career switch to Packer and injuries, because of which he missed 30 Tests. A sub-24 average and a 52+ strike rate tell the story.

Warne, in third position, is much more than the "ball of the century" and similar mind-blowing efforts. He had great variations and, barring against and in India, he was devastating everywhere. On dead pitches he had the ability to think out set batsmen. He gains slightly because he was in a strong bowling attack.

What does one say of Richard Hadlee, who is in fourth place. He might have played for a weak team but this works against him in the Match Performance analysis. However he has maintained 5 wickets per Test throughout his career. He was the single bowling star for his team for many years and deserves his second spot.

What Imran Khan would have done if he had bowled in those 8 batting-only Tests is anybody's guess. His 40-wickets performance against India in the 1982-83 series is one of the best series efforts ever and without any doubt the best performance by a pace bowler in the Asian sub-continent. A great captain and one of the greatest pace bowlers ever, as shown by this placement.

Before readers start sending torrents of mails asking why xyz is not ahead of pqr or something similar, please look at what separates the second to fifth placed bowlers, just 0.15 point. Kindly see them together as a band of equals.

Steyn comes in next. Do I see eyebrows raised at Steyn. If so, do not forget that his strike rate is 39.2, bettered only by the pre-WW1 figure of 34.1 by Lohmann (should be ignored for all purposes). He has captured 170 wickets in 33 Tests at an outstanding average of 23+. His Performance ratio (% of balls to % of wickets) is the highest for any bowler, standing at 1.78. His placement is also a vindication of the algorithms used in that a bowler with 170 wickets could be placed above bowlers who have captured in excess of 550 wickets.

Marshall, McGrath, Kumble and Waqar Younis complete this table of great bowlers. Each of these is a giant and could easily have graced the top-5. Alan Donald, the greatest South African pace bowler ever, just misses out.

Australia has three bowlers and Pakistan, as a tribute to their fast bowling skills, two bowlers. There are 3 spinners in this elite group, probably par for the period. Let me also add that only one more spinner, Harbhajan, that too just about, makes it to the top-20, making this a pace bowlers' era. Anyhow, other than, to a lesser extent, Saqlain Mushtaq and Abdul Qadir, there have not been very good spinners during these times.

As I am readying this for despatch, I get to view all-time best Australian XI. The three Australian bowlers in the Top-10 from this table and the no.2 from the Middle-era table have all found their place.

To view the complete list, please click here.

2.Current era (1970-2000): Table with support data

SNo. Cty Bowler          B/T Inn Rating Wkts Bow   Bow   Wkt  Wkt  B/W
Spls  Pts       StRt  RpO  Avge Qual Ratio

1. Slk Muralitharan M ROB 219 51.30 770 54.6 2.44 20.09 0.76 1.26 2. Aus Lillee D.K RF 127 48.05 355 52.0 2.76 24.58 0.80 1.49 3. Aus Warne S.K RLB 262 48.00 708 57.5 2.65 18.47 0.72 1.26 4. Nzl Hadlee R.J RFM 146 47.97 431 50.9 2.63 23.63 0.78 1.38 5. Pak Imran Khan RF 132 47.90 362 53.8 2.55 25.75 0.78 1.36 6. Saf Steyn D.W RF 61 45.55 170 39.3 3.62 21.55 0.74 1.78 7. Win Marshall M.D RF 149 45.44 376 46.8 2.69 22.97 0.80 1.42 8. Aus McGrath G.D RFM 241 44.86 563 52.0 2.50 19.22 0.81 1.40 9. Ind Kumble A RLB 234 44.58 619 66.0 2.70 20.66 0.76 1.13 10. Pak Waqar Younis RFM 149 44.18 373 43.5 3.25 20.33 0.78 1.50

11. Saf Donald A.A RF 126 44.13 330 47.0 2.84 20.06 0.80 1.56 12. Win Ambrose C.E.L RF 170 43.55 405 54.6 2.31 20.04 0.79 1.14 13. Win Holding M.A RF 110 43.40 249 50.9 2.79 25.28 0.79 1.36 14. Pak Wasim Akram LFM 175 43.22 414 54.7 2.59 19.56 0.74 1.21 15. Pak Shoaib Akhtar RF 78 43.21 178 45.7 3.37 20.94 0.79 1.46 16. Aus Lawson G.F RF 75 43.20 180 61.8 2.97 25.90 0.83 1.19 17. Aus Reid B.A LFM 40 43.03 113 55.3 2.68 21.75 0.82 1.50 18. Win Croft C.E.H RF 52 42.97 125 49.3 2.84 23.06 0.82 1.31 19. Aus Thomson J.R RF 87 42.82 200 52.7 3.19 27.17 0.84 1.39 20. Ind Harbhajan Singh ROB 137 42.51 330 65.1 2.81 19.07 0.72 1.07

To view the complete list, please click here.

3. Middle era (1920-1969): Table of top bowlers

SNo. Cty Bowler          BT Ratio Total Match  Wkt  Bow  Bow  Wkt  Wkt Perf
Pts  Perf   Pts StRt  Acc  Bat  Qty  Idx
Max Wt-> 80.0  40.0  10.0 10.0  5.0  5.0  5.0  5.0

1. Aus Grimmett C.V RLB 1.25 49.87 25.94 3.53 6.19 3.71 4.22 3.70 2.58 2. Aus O'Reilly W.J RLB 1.23 49.24 25.98 2.97 5.99 3.89 4.62 3.72 2.06 3. Saf Tayfield H.J ROB 1.13 45.20 23.10 2.94 5.12 3.46 4.93 3.76 1.87 4. Eng Trueman F.S RF 1.11 44.29 18.72 3.42 8.53 2.94 3.56 3.82 3.30 5. Pak Fazal Mahmood RFM 1.10 44.08 21.16 2.78 6.15 3.37 4.32 4.11 2.20 6. Eng Laker J.C ROB 1.09 43.46 19.32 2.75 7.00 3.36 4.33 3.99 2.70 7. Aus McKenzie G.D RF 1.07 42.84 19.91 2.97 6.07 3.19 4.39 4.07 2.24 8. Eng Bedser A.V RFM 1.07 42.68 19.51 3.13 6.45 3.22 3.85 3.94 2.60 9. Ind Chandrasekhar B RLB 1.06 42.23 18.86 3.12 6.57 3.09 4.50 3.90 2.20 10. Win Hall W.W RF 1.04 41.51 17.42 2.60 8.04 2.74 3.44 4.16 3.11

11. Aus Davidson A.K LFM 1.04 41.43 17.88 2.67 7.06 3.49 3.98 3.91 2.44 12. Eng Tate M.W RFM 1.03 41.19 19.80 2.51 4.80 3.93 4.09 3.89 2.18 13. Eng Snow J.A RFM 1.03 41.17 17.71 2.69 7.24 3.10 3.69 3.96 2.78 14. Ind Bedi B.S LSP 1.02 40.79 18.88 3.10 4.88 3.66 4.50 4.00 1.77 15. Saf Pollock P.M RF 1.02 40.75 16.81 2.32 7.80 3.09 3.68 3.97 3.09 16. Eng Underwood D.L LSP 1.02 40.74 17.52 3.02 5.58 3.74 4.62 4.12 2.15 17. Ind Gupte S.P RLB 1.01 40.57 19.36 2.74 5.58 3.14 3.59 3.84 2.32 18. Win Gibbs L.R ROB 1.01 40.54 20.00 3.19 4.21 3.67 4.01 3.65 1.79 19. Aus Lindwall R.R RF 1.00 40.17 15.75 2.67 7.35 3.15 4.67 3.90 2.69 20. Aus Johnston W.A LSP 1.00 40.06 17.54 2.40 6.22 3.33 4.10 4.09 2.38

The table is headed by two great leg-spinners from Australia, Grimmett and O'Reilly, two very different bowlers but were devastating wherever they played. They might have had the good fortune of having Bradman at slip rather than at the crease, but the England batting line-up was a pretty good one.

Tayfield, the South African off spinner is in third position, in a list where spin is king. His 9 for 113 off 37 consecutive overs against England remains the best bowling performance ever in this analysis.

Trueman, the fiery fast bowler and the first to reach 300 test wickets is in fourth position. He is also the best fast bowler in this middle era.

The fifth position is held by that master of seam, Fazal Mahmood, who troubled the batsmen on the matting wickets of Pakistan but outside also and allowed Pakistan to have a reasonable start to their test initiation. Unfortunately there was a lot of defensive thinking which meant that Fazal also had to act as the stock bowler.

The top-10 is completed by Laker, McKenzie, Alec Bedser, Chandrasekhar and Hall, an outstanding quintet. There are 5 spinners in this top-10 group indicating that this was an era which had a very strong spin presence.

To view the complete list, please click here.

4. Middle era (1920-1969): Table of support data

SNo. Cty Bowler          B/T Inn Rating Wkts Bow   Bow   Wkt  Wkt  B/W
Spls  Pts       StRt  RpO  Avge Qual Ratio

1. Aus Grimmett C.V RLB 66 49.87 216 67.2 2.16 21.10 0.74 1.29 2. Aus O'Reilly W.J RLB 48 49.24 144 69.6 1.95 23.08 0.74 1.03 3. Saf Tayfield H.J ROB 61 45.20 170 79.8 1.95 24.67 0.75 0.94 4. Eng Trueman F.S RF 122 44.29 307 49.4 2.62 17.80 0.76 1.65 5. Pak Fazal Mahmood RFM 50 44.08 139 70.7 2.10 21.58 0.82 1.10 6. Eng Laker J.C ROB 81 43.46 193 62.3 2.05 21.67 0.80 1.35 7. Aus McKenzie G.D RF 106 42.84 246 71.9 2.49 21.94 0.81 1.12 8. Eng Bedser A.V RFM 91 42.68 236 67.4 2.21 19.24 0.79 1.30 9. Ind Chandrasekhar B RLB 95 42.23 242 66.0 2.71 22.48 0.78 1.10 10. Win Hall W.W RF 88 41.51 192 54.3 2.92 17.20 0.83 1.55

11. Aus Davidson A.K LFM 80 41.43 186 62.3 1.98 19.92 0.78 1.22 12. Eng Tate M.W RFM 67 41.19 155 80.8 1.94 20.45 0.78 1.09 13. Eng Snow J.A RFM 90 41.17 202 59.5 2.69 18.44 0.79 1.39 14. Ind Bedi B.S LSP 113 40.79 266 80.3 2.14 22.50 0.80 0.89 15. Saf Pollock P.M RF 50 40.75 116 56.2 2.58 18.39 0.79 1.55 16. Eng Underwood D.L LSP 145 40.74 297 73.6 2.11 23.10 0.82 1.07 17. Ind Gupte S.P RLB 56 40.57 149 75.7 2.34 17.94 0.77 1.16 18. Win Gibbs L.R ROB 141 40.54 309 87.8 1.99 20.07 0.73 0.89 19. Aus Lindwall R.R RF 112 40.17 228 59.9 2.31 23.36 0.78 1.34 20. Aus Johnston W.A LSP 75 40.06 160 69.0 2.08 20.49 0.82 1.19

To view the complete list, please click here.

5. Pre-WW1 era (1877-1914): Table of top bowlers

SNo. Cty Bowler          BT Ratio Total Match  Wkt  Bow  Bow  Wkt  Wkt Perf
Pts  Perf   Pts StRt  Acc  Bat  Qty  Idx
Max Wt-> 80.0  40.0  10.0 10.0  5.0  5.0  5.0  5.0

SNo. Cty Bowler BT Ratio Total Match Wkt Bow Bow Wkt Wkt Perf Pts Perf Pts StRt Acc Bat Qty Idx Max Wt-> 80.0 40.0 10.0 10.0 5.0 5.0 5.0 5.0

1. Eng Barnes S.F RFM 1.27 50.71 26.15 3.90 6.72 3.53 3.37 3.97 3.08 2. Aus Turner C.T.B RFM 1.06 42.41 19.35 3.12 5.92 3.95 3.97 4.15 1.96 3. Eng Richardson T RF 1.05 41.87 19.84 3.44 5.92 2.95 3.26 3.93 2.52 4. Aus Spofforth F.R RFM 1.02 40.96 17.58 2.90 6.48 3.42 4.10 3.92 2.55 5. Aus Saunders J.V LSP 1.01 40.52 18.80 2.59 6.40 2.89 3.40 3.86 2.58 6. Eng Blythe C LSP 1.01 40.50 18.10 2.58 6.40 3.44 3.30 4.29 2.39 7. Aus Trumble H ROB 1.00 40.06 17.75 2.56 5.44 3.62 4.79 3.91 2.00 8. Eng Peel R LSP 1.00 39.85 18.77 2.70 5.92 3.91 2.46 3.90 2.20 9. Eng Lohmann G.A RFM 0.98 39.27 15.71 3.03 7.28 3.99 2.32 3.76 3.18 10. Aus Cotter A RFM 0.98 39.25 17.74 2.41 5.84 2.62 4.30 4.12 2.23

11. Aus Giffen G ROB 0.94 37.43 17.18 2.50 5.04 3.28 3.53 3.91 1.98 12. Aus Palmer G.E ROB 0.93 37.28 16.03 2.41 5.44 3.66 3.75 3.80 2.19 13. Eng Briggs J LSP 0.91 36.43 14.69 2.56 6.40 3.54 2.68 3.76 2.79 14. Aus Jones E RF 0.90 36.10 14.69 1.98 5.36 2.94 4.76 4.06 2.30 15. Aus Whitty W.J LFM 0.90 36.10 14.46 2.28 5.92 3.44 3.68 3.96 2.36 16. Saf Vogler A.E.E RLB 0.89 35.43 14.20 2.10 6.56 2.76 3.28 4.18 2.36 17. Nzl Cameron F.J RFM 0.88 35.04 13.45 1.63 5.76 4.33 4.24 3.79 1.84 18. Saf Faulkner G.A RLB 0.87 34.79 14.37 2.03 6.44 2.82 3.47 3.67 1.98 19. Aus Noble M.A ROB 0.87 34.67 13.51 2.05 5.28 3.37 4.85 3.70 1.92 20. Eng Ferris J.J LM 0.83 33.14 9.94 3.16 7.04 3.87 3.22 3.95 1.95 21. Eng Rhodes W LSP 0.81 32.24 12.47 1.90 5.44 3.40 3.54 3.49 2.00 22. Saf Sinclair J.H RLB 0.72 28.83 10.73 1.63 5.44 2.59 2.88 3.99 1.56 23. Aus Armstrong W.W RLB 0.69 27.57 10.47 1.37 2.73 3.70 4.00 3.97 1.34 24. Eng Woolley F.E LSP 0.69 27.52 8.79 1.28 4.49 3.30 4.10 3.83 1.75

Exactly 10 bowlers fulfill the criteria (Since changed cut-off to 60 wkts). The list is, as expected, headed by Sid Barnes, by the reckoning of many, the best fast-medium bowler ever. He is ahead of the next bowler by over 20%. Then come those deadly exponents of pace and spin who revelled on those uncovered deadly pitches.

Surprising thing is that Lohmann, despite his devastating strike rate and average, comes as low as fifth. His match performances have been below-par. The opposition has also been quite average. This list is dominated by spinners, 7 in all, but led by two great fast medium bowlers. Quite surprising that there is no leg spinner. Grimmett and O'Reilly started the tradition of great leg spinners, after the war.

6. Pre-WW1 era (1877-1914): Table with support data

SNo. Cty Bowler          B/T Inn Rating Wkts Bow   Bow   Wkt  Wkt  B/W
Spls  Pts       StRt  RpO  Avge Qual Ratio

SNo. Cty Bowler B/T Inn Rating Wkts Bow Bow Wkt Wkt B/W Spls Pts StRt RpO Avge Qual Ratio

1. Eng Barnes S.F RFM 48 50.71 189 41.7 2.37 16.83 0.79 1.54 2. Aus Turner C.T.B RFM 29 42.41 101 51.3 1.93 19.85 0.83 0.98 3. Eng Richardson T RF 22 41.87 88 51.1 2.96 16.31 0.79 1.26 4. Aus Spofforth F.R RFM 29 40.96 94 44.5 2.48 20.51 0.78 1.28 5. Aus Saunders J.V LSP 27 40.52 79 45.1 3.02 17.01 0.77 1.29 6. Eng Blythe C LSP 36 40.50 100 45.5 2.46 16.51 0.86 1.20 7. Aus Trumble H ROB 57 40.06 141 57.4 2.28 23.93 0.78 1.00 8. Eng Peel R LSP 35 39.85 102 51.1 1.97 12.30 0.78 1.10 9. Eng Lohmann G.A RFM 34 39.27 112 34.1 1.89 11.60 0.75 1.59 10. Aus Cotter A RFM 34 39.25 89 52.1 3.30 21.49 0.82 1.12

11. Aus Giffen G ROB 39 37.43 103 62.0 2.62 17.67 0.78 0.99 12. Aus Palmer G.E ROB 29 37.28 78 57.9 2.23 18.75 0.76 1.09 13. Eng Briggs J LSP 45 36.43 118 45.2 2.36 13.42 0.75 1.40 14. Aus Jones E RF 29 36.10 64 58.6 2.97 23.81 0.81 1.15 15. Aus Whitty W.J LFM 25 36.10 65 51.6 2.45 18.41 0.79 1.18 16. Saf Vogler A.E.E RLB 27 35.43 64 43.2 3.16 16.40 0.84 1.18 17. Nzl Cameron F.J RFM 36 35.04 63 77.7 2.39 21.21 0.76 0.92 18. Saf Faulkner G.A RLB 38 34.79 82 51.5 3.09 17.36 0.73 0.99 19. Aus Noble M.A ROB 63 34.67 121 59.2 2.54 24.24 0.74 0.96 20. Eng Ferris J.J LM 16 33.14 61 37.7 2.02 16.10 0.79 0.97 21. Eng Rhodes W LSP 75 32.24 127 64.8 2.50 17.68 0.70 1.00 22. Saf Sinclair J.H RLB 36 28.83 63 57.1 3.33 14.42 0.80 0.78 23. Aus Armstrong W.W RLB 70 27.57 87 92.2 2.19 20.00 0.79 0.67 24. Eng Woolley F.E LSP 72 27.52 83 78.3 2.60 20.50 0.77 0.87

I do not expect the readers to agree with all the placings. They have every right to disagree in a nice, positive, contributory manner. I have no problems if you express your disagreement supported by subjective, objective or figures-based arguments. Kindly stay away from rude, offensive or abusive comments. Also resist making mundane bare comments such as "abc is better than xyz.". Also all comments on batsmen have to be relevant to the topic under discussion. Otherwise, they are unlikely to see the light of the day.

One final note. Muralitharan's action has been analyzed and deemed to be perfectly acceptable by ICC. That is enough for me. That may not be enough for some readers, I have no problem with that. However please do not raise that issue in response to this article. One such comment I will ignore. If readers persist with such comments, I will have no other option but to ignore all their comments, however valid those might be. This is not the forum for such comments.

A reminder that the bowler-type tables will be brought out in the follow-up article.

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

Comments have now been closed for this article

  • Cabe on July 31, 2012, 14:56 GMT

    I bowl every once in awhile. I used to bowl a fair amonut in college. I use a two finger spin technique. I think this is where my problem is. I am fairly consistent but I never seem to improve. I want my average to go up. Do you have tips on improving my two finger spin technique? Also a couple of my buddies have changed over to a three finger (two fingers and a thumb) technique which has really helped them improve and be more consistent. I have tried it but do not know how to spin the ball using this technique. Can you help me out?

  • Zeeshan Ahmed on July 29, 2009, 11:26 GMT

    No doubt your another nice article. Almost same expected results are there. I think this one is better one to judge bowlers. Here one thing come in my knowledge that in All - Time XI of Australia, O Reilly selected instead of Grimmett as according to them he is perfect with Shane Warne. At one ocassion Don Bradman said that the best bowler of his era is O Reilly. Grimmett bowling average is 24.21 but against England his average is 32.44. O Reilly bowling average is 25.36 against England. I think both are two greatest bowlers among their contemporaries.

  • Mohsin Khan on July 28, 2009, 19:06 GMT

    Great effort,

    I think the analysis is really detailed and a lot of effort has gone into it but in the end this just shows how statistics are never a substitute for watching the game.

    Some very off rankings in there, Waqar>>Wasim, Waqar >> Holding, Garner....etc Kumble >> Wasim, Ambrose, Holding...etc

    Harbhajan in the top 20s (DOES NOT DESERVE IT AT ALL) no way is he better then Garner, Saqlain, Walsh and Kapil Dev and many others.

    Another thing which may require further thought is that instead of having the subcontinent vs outside distinction we need to introduce a bowlers adaptability to different conditions, best way I would think to do that would be to have some sort a standard deviation in a bowlers averages in different countries.

    I would suspect Warne and Muralitharan would drop a few points due to their ineffectiveness in India and Australia respectively.

  • Shankar Krishnan on July 27, 2009, 6:50 GMT

    I did not see the innings strike rate being used anywhere in the rating. Instead the career strike rate is used.I feel the innings strike rate is a key measure that is worth considering. A bowler,though not consistent, may produce devastating innings / match performances on and off. Harmison & Harbhajan, for example.In the current methodology these will get rolled up into a career number that may not reflect accurately his value / contribution to the teams' success.May be a good idea to see how the numbers fall by considering the innings strike rate. [[ Shankar It is an excellent idea but will not make sense if it is applied in an absolute manner. At the innings level it has to be relative to the team strike rate. Also it cannot be relative to the match strike rate in which case the poor bowling of the weak team (say, B'Desh) will over-value a stronger team's bowling. My idea is to do this as an additional base measure itself in view of the difficulty in getting the multiplicative measure correct. Let me look at it carefully. Overall a veru good idea. Ananth: ]]

  • Alex on July 27, 2009, 5:20 GMT

    Ananth - Thanks for the clarification. I am in investment business, and do something similar to the company and market data compared to what you have done with cricket database, which is my hobby. We still call it statistical analysis. The experience is that it is useful to create categories (alongwith variance or confidence intervals) which act as good filters. Beyond that, it is mostly seasoned intuition.

    I do not have a problem with BCL being No. 1. His peaks were spectacular, and he actually was quite consistent at a high level in his last 4 yrs. Likewise, barring Bradman, no one had a 5-6 year long purple patch that Ponting had over years 2000-06. I still prefer SRT over them but have no complaints about your metric. It's not like getting ranked #1 on such lists will help him pay for his kids' education, after all :)

  • Ananth on July 26, 2009, 16:07 GMT

    Alex/Yash First let me say that what I do is not a statistical analysis. It is Cricket Analysis. I rarely use the statistical methodology nor do I use statistical techniques. Once in a while I do some statistical derivations such as Mean, SD et al. What I do is to use my database, apply analytical techniques, cricketing knowledge, computer skills, logic and a bit of lateral thinking and finally go by consensus of the readers to derive common sense based tables. Most important, almost all my work is post-mortem, so what-if questions do not arise. There is no way I am going to be able to indicate what "variance" which is expected in my analysis since my analysis is not statistical. This is the main reason why a completely way-out, unbelievable, totally theoretical example does not excite me since I conclude that such an example will never happen and ignore it. What Yash should do is to pick up 2/3/4 real life bowlers of similar bowling averages, find out the assigned points in this analysis and point out discrepancies (as Arjun had done earlier re Bowling averages). Then I will look at that carefully and apply corrections, if warranted. If I start taking the "5-sixes-1-wicket-in-an-over" during a player's entire career seriously then someone will send me another extreme example of a Nadkarni type analysis of 32-27-5-0 repeated in 50 matches and raise hell that something is wrong. Or, as some readers did in the Batting analysis, talk about 7 0's and 1 x 400 with 8 x 50s, again repeated through a player's career. At the end of the analysis my tables must pass the test of reasonableness and be acceptable to most readers. Even in the batting analysis, all the fans of Tendulkar were fighting a bitter battle for one point and one place. If Lara and Tendulkar had exchanged places, all these would have been happy. My point is that all the questioning of my analysis was only to achieve this simple exchange. To summarize, use only real life examples to make your point. And do not think that this is a purely statistical analysis and start looking at statistical variations and discrepancies. And I suggest do not ignore the importance of including proper and valid cricketing ideas.

  • Anurag on July 26, 2009, 13:47 GMT

    have changed vastly over the years. While an economy rate of 2.5 is fantastic for a post-2000 bowler, it is average for the early bowling era or even the era of stonewalling batsmen like gavaskar. However, a 6-2 6-2 6-2 win with 25 aces is great for a laver as well as a federer.

    Finally inspite of everything said and done, a genuine legend like bradman or laver will stand at the helm of any list, whether statistical or one handwritten in gold by god.

    Take this analysis as a mere comparison of players' statistics and understand that a mere percentage point is no way to say imran khan was inferior to hadlee

  • Anurag on July 26, 2009, 13:37 GMT

    A lot of comments on this blog have expressed outrage at the wideness of the eras. While it is an undeniable fact that cricket, or any other sport for that matter, has been turned on its head in the 130-odd years of its existence, is it not interesting to see greats of different eras being pitted against each other in such an analysis? While acknowledging the greatness of a barnes or an o'reilly or a marshall or a murali, why not use the tool of statistics with different yardsticks for different eras to bring them onto a level playing field? It cerainly will not be unfair to do that since the standout bowlers of each era have about 50 rating points.

    I understand comparisons with tennis or golf where the debates to ascertain the greatest player ever are even more intense since those are individual sports and not team sports. This makes it much easier in those sports rather than in cricket, where more than equipments or playing fields, the mindsets and statistics

    to be continued

  • Alex on July 26, 2009, 12:39 GMT

    Ananth - Martin Crowe has made some nice suggestions so far, and hopefully he will give you the exact details of how those should be implemented. If he has the bandwidth, perhaps he could give us his Top 10 batsmen and Top 10 bowlers (as per his choice of eras), alongwith his reasons.

    Such statistical excercize make it possible to categorize but can never tell who is the best or second best etc ... which are more subjective matters. But with input from knowlegeable players and former players, this method could be improved by ICC etc. to award performances every year.

  • Alex on July 26, 2009, 12:07 GMT

    Ananth - I think Yash made a good point. Yours is a statistical interpretation of data, and he has concocted a hypothetical (but valid) scenario. Part of the problem will go away if you add a column of "variance" associated with your final score of a bowler ... I had requested this earlier. Most captains (and selection committees) would prefer a strike bowler who has slightly lower final score than someone else, provided he has a lower variance ... we are looking at an "investment" over 10-12 years in this exercize.

    To come back to Yash's example: the answer depends on the variance of this bowler. If he keeps producing such performances under all circumstances, sure he is better than Murali "except when the conditions really favor the bowlers". Because even on a featherbed wicket, this guy will ensure that the other side is bowled out twice in a match within 40 overs for a net total of approx. 600. His team can score those runs in 400 overs and win the match!

  • Cabe on July 31, 2012, 14:56 GMT

    I bowl every once in awhile. I used to bowl a fair amonut in college. I use a two finger spin technique. I think this is where my problem is. I am fairly consistent but I never seem to improve. I want my average to go up. Do you have tips on improving my two finger spin technique? Also a couple of my buddies have changed over to a three finger (two fingers and a thumb) technique which has really helped them improve and be more consistent. I have tried it but do not know how to spin the ball using this technique. Can you help me out?

  • Zeeshan Ahmed on July 29, 2009, 11:26 GMT

    No doubt your another nice article. Almost same expected results are there. I think this one is better one to judge bowlers. Here one thing come in my knowledge that in All - Time XI of Australia, O Reilly selected instead of Grimmett as according to them he is perfect with Shane Warne. At one ocassion Don Bradman said that the best bowler of his era is O Reilly. Grimmett bowling average is 24.21 but against England his average is 32.44. O Reilly bowling average is 25.36 against England. I think both are two greatest bowlers among their contemporaries.

  • Mohsin Khan on July 28, 2009, 19:06 GMT

    Great effort,

    I think the analysis is really detailed and a lot of effort has gone into it but in the end this just shows how statistics are never a substitute for watching the game.

    Some very off rankings in there, Waqar>>Wasim, Waqar >> Holding, Garner....etc Kumble >> Wasim, Ambrose, Holding...etc

    Harbhajan in the top 20s (DOES NOT DESERVE IT AT ALL) no way is he better then Garner, Saqlain, Walsh and Kapil Dev and many others.

    Another thing which may require further thought is that instead of having the subcontinent vs outside distinction we need to introduce a bowlers adaptability to different conditions, best way I would think to do that would be to have some sort a standard deviation in a bowlers averages in different countries.

    I would suspect Warne and Muralitharan would drop a few points due to their ineffectiveness in India and Australia respectively.

  • Shankar Krishnan on July 27, 2009, 6:50 GMT

    I did not see the innings strike rate being used anywhere in the rating. Instead the career strike rate is used.I feel the innings strike rate is a key measure that is worth considering. A bowler,though not consistent, may produce devastating innings / match performances on and off. Harmison & Harbhajan, for example.In the current methodology these will get rolled up into a career number that may not reflect accurately his value / contribution to the teams' success.May be a good idea to see how the numbers fall by considering the innings strike rate. [[ Shankar It is an excellent idea but will not make sense if it is applied in an absolute manner. At the innings level it has to be relative to the team strike rate. Also it cannot be relative to the match strike rate in which case the poor bowling of the weak team (say, B'Desh) will over-value a stronger team's bowling. My idea is to do this as an additional base measure itself in view of the difficulty in getting the multiplicative measure correct. Let me look at it carefully. Overall a veru good idea. Ananth: ]]

  • Alex on July 27, 2009, 5:20 GMT

    Ananth - Thanks for the clarification. I am in investment business, and do something similar to the company and market data compared to what you have done with cricket database, which is my hobby. We still call it statistical analysis. The experience is that it is useful to create categories (alongwith variance or confidence intervals) which act as good filters. Beyond that, it is mostly seasoned intuition.

    I do not have a problem with BCL being No. 1. His peaks were spectacular, and he actually was quite consistent at a high level in his last 4 yrs. Likewise, barring Bradman, no one had a 5-6 year long purple patch that Ponting had over years 2000-06. I still prefer SRT over them but have no complaints about your metric. It's not like getting ranked #1 on such lists will help him pay for his kids' education, after all :)

  • Ananth on July 26, 2009, 16:07 GMT

    Alex/Yash First let me say that what I do is not a statistical analysis. It is Cricket Analysis. I rarely use the statistical methodology nor do I use statistical techniques. Once in a while I do some statistical derivations such as Mean, SD et al. What I do is to use my database, apply analytical techniques, cricketing knowledge, computer skills, logic and a bit of lateral thinking and finally go by consensus of the readers to derive common sense based tables. Most important, almost all my work is post-mortem, so what-if questions do not arise. There is no way I am going to be able to indicate what "variance" which is expected in my analysis since my analysis is not statistical. This is the main reason why a completely way-out, unbelievable, totally theoretical example does not excite me since I conclude that such an example will never happen and ignore it. What Yash should do is to pick up 2/3/4 real life bowlers of similar bowling averages, find out the assigned points in this analysis and point out discrepancies (as Arjun had done earlier re Bowling averages). Then I will look at that carefully and apply corrections, if warranted. If I start taking the "5-sixes-1-wicket-in-an-over" during a player's entire career seriously then someone will send me another extreme example of a Nadkarni type analysis of 32-27-5-0 repeated in 50 matches and raise hell that something is wrong. Or, as some readers did in the Batting analysis, talk about 7 0's and 1 x 400 with 8 x 50s, again repeated through a player's career. At the end of the analysis my tables must pass the test of reasonableness and be acceptable to most readers. Even in the batting analysis, all the fans of Tendulkar were fighting a bitter battle for one point and one place. If Lara and Tendulkar had exchanged places, all these would have been happy. My point is that all the questioning of my analysis was only to achieve this simple exchange. To summarize, use only real life examples to make your point. And do not think that this is a purely statistical analysis and start looking at statistical variations and discrepancies. And I suggest do not ignore the importance of including proper and valid cricketing ideas.

  • Anurag on July 26, 2009, 13:47 GMT

    have changed vastly over the years. While an economy rate of 2.5 is fantastic for a post-2000 bowler, it is average for the early bowling era or even the era of stonewalling batsmen like gavaskar. However, a 6-2 6-2 6-2 win with 25 aces is great for a laver as well as a federer.

    Finally inspite of everything said and done, a genuine legend like bradman or laver will stand at the helm of any list, whether statistical or one handwritten in gold by god.

    Take this analysis as a mere comparison of players' statistics and understand that a mere percentage point is no way to say imran khan was inferior to hadlee

  • Anurag on July 26, 2009, 13:37 GMT

    A lot of comments on this blog have expressed outrage at the wideness of the eras. While it is an undeniable fact that cricket, or any other sport for that matter, has been turned on its head in the 130-odd years of its existence, is it not interesting to see greats of different eras being pitted against each other in such an analysis? While acknowledging the greatness of a barnes or an o'reilly or a marshall or a murali, why not use the tool of statistics with different yardsticks for different eras to bring them onto a level playing field? It cerainly will not be unfair to do that since the standout bowlers of each era have about 50 rating points.

    I understand comparisons with tennis or golf where the debates to ascertain the greatest player ever are even more intense since those are individual sports and not team sports. This makes it much easier in those sports rather than in cricket, where more than equipments or playing fields, the mindsets and statistics

    to be continued

  • Alex on July 26, 2009, 12:39 GMT

    Ananth - Martin Crowe has made some nice suggestions so far, and hopefully he will give you the exact details of how those should be implemented. If he has the bandwidth, perhaps he could give us his Top 10 batsmen and Top 10 bowlers (as per his choice of eras), alongwith his reasons.

    Such statistical excercize make it possible to categorize but can never tell who is the best or second best etc ... which are more subjective matters. But with input from knowlegeable players and former players, this method could be improved by ICC etc. to award performances every year.

  • Alex on July 26, 2009, 12:07 GMT

    Ananth - I think Yash made a good point. Yours is a statistical interpretation of data, and he has concocted a hypothetical (but valid) scenario. Part of the problem will go away if you add a column of "variance" associated with your final score of a bowler ... I had requested this earlier. Most captains (and selection committees) would prefer a strike bowler who has slightly lower final score than someone else, provided he has a lower variance ... we are looking at an "investment" over 10-12 years in this exercize.

    To come back to Yash's example: the answer depends on the variance of this bowler. If he keeps producing such performances under all circumstances, sure he is better than Murali "except when the conditions really favor the bowlers". Because even on a featherbed wicket, this guy will ensure that the other side is bowled out twice in a match within 40 overs for a net total of approx. 600. His team can score those runs in 400 overs and win the match!

  • Harsh Thakor on July 26, 2009, 9:01 GMT

    A bowler worth mentioning is Wasim Akram,the most versatile of all fast bowlers who performed outstandingly on flat sub-continent tracks and was the best ever old ball bowler.To me he was just a notch below the likes of Dennis Lillee,Glen Mcgrath Etc ,and better than Waqar younus and far ahead of Shoaib Akhtar.Andy Roberts to me was in the same class as the great bowlers discussed above and was a greater match-winner than Michael Holding,posessing a greater repertoire of deliveries.Sunil Gavaskar and Dennis Lillee rate Andy Roberts as the best fast bowler of their era.To me for his consistency and accuracy Joel Garner should have made the top 15 and for his great haul of wickets on dead tracks Kapil Dev should have been rated above Ian Botham or Shaun Pollock.Kapil Dev above all had no support from any bowlers and still troubled the greatest of batsman.

  • Marcus on July 26, 2009, 6:26 GMT

    Yash

    Looking at your Dale Steyn example, if his economy is a little high then his strike rate is low enough to cancel it out. Take another bowler with a SR of 50 and an average of 25. His SR is 3. Dale Steyn's SR is 3.6. It's higher than the second bowler, but he takes 10 fewer balls on average to take a wicket, and only gives away 6 runs more on average PER 10 OVERS. So I think that demonstrates that most captains will be prepared to live with a handful more runs for greater striking power.

  • Yash Rungta on July 26, 2009, 3:13 GMT

    Lets take an example. Lets assume there is a bowler who gives 5 sixes in an over and takes a wicket on the 6th ball throughout his career. His average would be 30 and a strike rate would be six. Lets say he plays 100 tests and takes 400 wickets. In the current analysis, he would get Strike Rate points almost 10 and the Perf Index Points almost 5. Bowling accuracy would be 0. This is equal to 15 out of 20. Whereas the best bowler, Murali, only has about 13/20.

    Would you really prefer the former bowler(where the opposition on an avg scores 300 off him) or Murali against whom the opposition on an avg scores only about 215? I would definately have Murali.

    Although a bowler like Steyn takes wicket every 40 balls, he also gives about 24 runs every 40 balls. [[ Yash I have never received a more outrageous and contrived comment as this. Instead of accepting a very commonly accepted dictum on bowlers, you have tried to create a totally artificial and unnnatural example which anyhow does not prove anything. If you want to retain your credibility and your comments taken seriously I suggest, stick to normal cricketing examples. Otherwise such contrived and totally worthless comments will not see the light of day. Ananth: ]]

  • Engle on July 25, 2009, 19:04 GMT

    I was watching some interview clips of C.Lloyd and V. Richards on the 75/76 series in Aus. They spoke of the pace generated by Thommo and the pall of apprehension he cast over the W.Indies.

    Is the 'fear factor' or intimidation relevant in assessing a bowlers worth ? [[ Bennett The fear factor is one of the most used (and abused) terms whether it related to Richards or Thomson or Roberts. Totally subjective. Also one batsman might be afraid, another would take it as a challenge to be faced and the third might might take it as an opportunity. Ananth: ]]

  • Saf on July 25, 2009, 13:40 GMT

    @yash rungta A lot of people would turn your argument around and say that actually the strike rate is by far the most important feature. It is well accepted that taking wickets is the best way to check the scoring. In fact if a bowler has a high career strike rate and high average, the average is probably skewed because of those matches wherein he bowled on dead pitches. Any captain would prefer a bowler who takes wickets in most situations. Admittedly there are situations where you simply want to contain runs, but so often we have seen that in these situations teams simply resort to negative bowling (wide of off stump with a loaded off side field, leg stump bowling etc)…but a bowler who can take wickets, he’s the one who is doing something very very right. In my opininon strike rate is almost everything. [[ Saf Well put. I would say that the two aspects of bowling are essential to any good bowler but the almost unanimous opinion is that taking wickets is more important, hence the higher weight. Ananth: ]]

  • Yash Rungta on July 25, 2009, 12:56 GMT

    If I'm not mis-understanding, the 'Performance ratio' measure is also Strike Rate or somewhat related to it. Can you elaborate? You're giving too much weight to strike rates which has resulted in bowlers like Steyn, Akhtar etc. come unnecessarily at the top. While a good strike rate is good for a strong team(West Indies in 1970s80s) and Australia(past 10 yrs), it is bad for a weak team. Maybe you could tweak this measure to take into account the team's strength. Although this wouldn't reward the skill of the bowlers, it would reward the utility of the bowlers to the team.

    To me, the Batting analysis was almost perfect. If you could just about mirror the criteria in the bowling, you'll surely get the best results.

    One request, can you please switch the links on so that we can just click on the player which will take us to his profile page? If not that blog, atleast form the page where we can see all the player's rankings?

  • Yash Rungta on July 25, 2009, 12:43 GMT

    Ananth,

    I can give a list of matches drawn by slender margins and point out bowlers who, if had bowled with a better strike rate might have lost the match for their team. Giving a higher weight to strike rate(10 points) than to economy rate(5 points) just diminishes the job done by people like say Vaas who tirelessly hold one end up and allow people like Murali to take wickets from the other end. Although its great to have people who can take wickets every 40 balls, people don't understand its also has a hidden disadvantage which is quite equal to the advantage. Having bowlers with low strike rates especially in weakers teams is better which actually enables them to draw matches than lose matches. You can have equal weights for economy and strike rates or better, just go by average and give it 15 points, the result will be the same.

    Ultimately, its the average that counts and ofcourse the quality of batsmen, the match situation, type of wickets, pitch etc. [[ Yash You are just talking with no supporting proof. I have no idea what you are trying to convey or what your grouse except that the high strike rate bowlers have done well in this analysis. Let me see one person who comes out supporting you that strike rates in test bowlers is less important than accuracy. Ananth: ]]

  • Bejaan on July 25, 2009, 5:14 GMT

    Ananth, Regret my umpteenth post. But I felt I must add…the intention is not at all to “put down modern greats”. Quite the opposite. In fact, if at all, by declaring every successive generation player as the “greatest ever” we are actually putting down previous generations greats. [[ Bejaan For the sake of argument let us say that Arnold Palmer was anointed in the 60s as the greatest. That spurred Nicklaus to reach farther than any golfer has gone and he was, and still is, considered to be the best ever. Now Woods is within 6 majors of his tally and he probably has no one but Nicklaus ahead of him. Let us say he finishes with 22 titles. He would be crowned as the best ever. It goes on. The great thing is that with each such announcement there is a new person who would come on and take over the mantle. that is the charm. Of course there are sportsmen, who are unlikely to be unseated, such as Bradman, Pele. Even today, 60 years after retirement, other than a few biased fans, Bradman is the greatest Test batsman ever, note I have not said player. So when someone sets the bar so high, people would not indiscriminately anoint a new all time best. Ananth: ]]

  • Harsh Thakor on July 25, 2009, 4:48 GMT

    Ananth,basically i respect your high rating of Imran Khan,who was outstanding on docile subcontinent surfaces and that too was a match-winner ,winning crucial tset matches in England,(7-40 at Leeds in 1987)Australia(6-63 and 6-102 at Sydney in 1977) and the West Indies.(7-83 at Georgetown in 1988),enabling Pakisatn to almost win the unoficial tset world title.I also like your placing Michael Holding high ,considering he was consistently the fastest of all bowlers. However just as you made the last improvised list I would like you to work on a rating that could perhaps do justice to Malcolm Marshall who to me was just a hair's breadth below Dennis Lillee.Remember he was a champion in all conditions with an outstanding strike rate and overshadowed the graetset West Indian paceman.For bearing the brunt of a weak taem and a series of outstanding match-winning efforts Curtly Ambrose could have been placed a few notches ahead.His acuracy was phenomenal.

  • Anwar-ul-Haque on July 25, 2009, 4:22 GMT

    Wow!!! Great analysis, I must say. One reservation from my side. From a bowler perspective, the wicket of a set batsman is difficult to take as compared to a new batsman; whereas you have given more weightage to the bowling efforts where bowlers take wicket of new batsmen. All the great batsmen (Lara, Tendulkar, Richards, Miandad, Gavasker etc.) are difficult to get out once they set on to the crease (this is more true for slow starter batsmen like Inzamam). So when a bowler take wicket of a set batsman, it must be given more weightage in my opinion. [[ Anwar While it may be, theoretically, more difficult to dismiss a set batsman as against a batsman who has come in, this analysis primarily about value to the team. Let me explain with a simple pair of sentences. If Ponting/Tendulkar/Lara/Sangakkara/Pietersen were dismissed 100 times at a score 0f 0-5, I am sure you will agree with me that no LESS than 90 of these dismissals would have been of great benefit to the bowling team. On the other hand, If Ponting/Tendulkar/Lara/Sangakkara/Pietersen were dismissed 100 times at a score 0f 100+, I am sure you will agree with me that no MORE than 10 of these dismissals would have been of great benefit to the bowling team. Pl don't point out dismissals at 136 or 195 or 221 which have been of great value to the bowling team. For each one of these examples I can point 10 times the number of low score dismissals which have been of great value. Ananth: ]]

  • Marcus on July 25, 2009, 4:14 GMT

    Bejaan- when I watched that great movie "Chariots of Fire" last year I was struck by the fact that the competitors basically had to run on fine gravel- not smooth modern running tracks. So perhaps Harold Abrahams would beat Bolt today on a smoother track. He'd almost certainly post a faster time.

    Maybe not stricly relevant, but my inner pedant must be satisfied! [[ Would he, Marcus, when the timings have advanced by hundredths of a second every year. But I will not counter any one of yours or Bejaan's arguments since I am as much of a person for history-is-great and the what if questions. The only dictum I follow is not to put down the modern greats using what if theories. Ananth: ]]

  • Bejaan on July 25, 2009, 4:11 GMT

    Ananth, All very true but: 1) Your reply itself contains what I meant. i.e. the ACTUAL performance level will go on increasing. So, over the last 100 yrs the actual performance level has increased almost linearly. This does not mean that every successive 10 yrs the fastest sprinter then is the greatest “ever”. Because you are pretty certain that his record will be broken say 10yrs down the line. As rgd. Federer/Nadal adjusting, they would but to what extent? Also, how do you factor in competition, importance to events etc. for eg. The “golden age” of tennis had 4/5 all time greats viz. Connors,Borg,Mac,Lendl etc all sharing the spoils. And no one even bothered about the Aus open then. Borg never played. Mac played one in his prime. So ,the actual amount of grand slams which could be won could be reduced to 75% of now available. Borg had 11( and lost 4 US finals to Mac/Connors). Also, then grass and clay were poles apart. Now, it is a common sight to see baseliners (Federer is predominantly one) play an almost exactly similar game on all surfaces. Even Borg used to serve/volley on first serve on grass. Mac had his career cruelly interrupted almost halfway by the sudden advent of graphite almost instantaneously rendering his game obsolete,Etc…ad infinitum,so a direct comparison is very very difficult, if not impossible. 2) My point of “longevity” was that if you take say only wickets/tests into account without incorporating time period…a bowler may do very well for a short “time” period for several reasons: situations conducive to all bowlers, or just to his particular type of bowling, he may just be on a hot streak, batsmen haven’t figured him out…so many things. But if a bowler is taking wickets after 5,10,15 yrs then several factors especially a period where things were just going good, a random hot streak ,conditions generally very conducive to bowling etc etc may be ruled out.

  • Yash Rungta on July 25, 2009, 4:07 GMT

    45 points for match perf, 15 points for runs(here wkts), 5 points for % of team's wkts etc. etc. The batting analysis was very very reliable.

    I don't why you need to take strike rate and economy rates into account at all. It should be 'average' only because average itself is a function of strike rate and economy rate. For the same average, a bowler having a higher economy rate will have a better strike rate and vice-versa. What difference will it make if a bolwer takes a wicket every 40 balls or 60 balls if he gives as many runs... A better strike rate would mean more results, but a result can be a loss or a win. This is in line with you dropping the strike rate criteria for batsmen. This will remove the undue advantage strike bowlers like Steyn, Waqar, Donald etc. have had in this analysis. [[ Yash The separation of Bowling Average into its components has worked well. You have to give me concrete evidence to show that the Bowling Average, as a single entity, would have been better. And this is almost the first time I have seen someone mentioning that better strike rates in bowlers should not be considered and that if bowlers take wickets quicker, the team might lose. Also I dropped batsman strike rates only because the information was not available for 70% of the matches. Ananth: ]]

  • Yash Rungta on July 25, 2009, 4:03 GMT

    Regarding cut-offs, I believe atleast in the modern era, i.e. post 1990, cut-off should be 50 matches. If you go by no. of wickets, to be called a great fast bowler, you need to keep a minimum of 250 wickets because you cannot just even consider players like Steyn(170 wkts), Pathan(100 wkts), Akhtar(178 wkts) etc. This is because you need to be consistent over a period of atleast 7-8 yrs to be called a great bowler. But a minimum of 25 wkts would leave bowlers like Cairns, Flintoff etc. Even Kallis just makes it with his 258 wickets. But these will be included if the cut-off is 50 matches. Another cut-off say a test playing span of 7 yrs would be good. You certainly don't want people like Steyn(who has not even reached half of his career) being compared with greats like Ambrose, Akram, McGrath etc.

    Another thing, can't you just mirror the points you had in the batting for the bowling analysis? Say 45 points for match performance, 15 points for runs(here wkts)

    To be cont...

  • Bejaan on July 25, 2009, 2:41 GMT

    Kartik,abdulaa,ananth

    1) Your point about longevity is valid. However 50 odd tests is a very good representative size when judging how good a bowler would be…ESPECIALLY if it is taken over a period of TIME. i.e. the time element is MORE (much more) important than the match volume! For e.g. a bowler may play 50 tests in 2 years (exaggerated, but as an e.g.) at his peak and do superbly well…does this count as longevity?? Certainly NOT. Longevity would be 50 tests over say 10 years, wherein different conditions, opponents, batsmen getting used to the bowlers “tricks”-whatever truly test the skill of a bowler. So, just as for the batting analysis Ananth has used a square root of striker rate AND average. Some method must be devised to incorporate BOTH number of tests AND years taken to complete them. 2) At a pinch a sample size of 50 tests spread over a long period is actually very good! Even the Don played only 52 tests….but over a LONG time…so we can relatively safely say that he would have been good all along. But if he had played the 52 tests ALL in a short period of say 4 years….then that would cast MANY many doubts about the validity of his achievements. 3) Ananth: Your Federer “fixation” (don’t mean to be rude) is fascinating. But we always cherish some recent grand performances more than those past. In my opinion you simply cannot compare. For.eg Federer would find it impossible to play most of his shots with wooden rackets…never mind his backhand flicks and squash shots. The training, diet, lifestyle, competition …what have you were so very different in different eras...that it is truly impossible to judge. For e.g. Usain Bolt (on pure performance) would thump any past sprinter. And if he does have another good Olympics (or 2) you can bet everyone will be jumping all over themselves to call him the greatest sprinter ever.

    [[ Bejaan The cut-off here is not number of tests but test wickets which are the cornerstone of any bowlers' performance. 175 is around 45 tests while 200 may be around 55 tests considering that only 49 bowlers have achieved the 4.00 mark. So even if I say that Federer should be treated as first among equals it does not gell with you. If there were wooden rackets today, great players such as Federer/Nadal would have adjusted, just as Borg would have adjusted to the graphite rackets. Finally if Bolt does indeed run in 9.60 secs and/or captures another gold with a sub-9.70 performance, pray what is wrong in calling him the greatest sprinter ever. Why this reluctance to shower a superlative. Maybe Bolt would retire and you may see, maybe not me, a Ralph Hammer, or whoever, run a sub-9.50 100 metres, ready to be called the greatest in 2030 and so on. Ananth: ]]

  • Kartik on July 25, 2009, 0:48 GMT

    Ananth,

    I am going to have to defend Kapil's final 16 (post-Australia tour) Tests here.

    It is true that his strike rate was terrible during this time, but NOT his average (which still stayed below 30 even for his final 34 wickets).

    http://stats.cricinfo.com/ci/engine/player/30028.html?class=1;template=results;type=bowling

    Needless to say, his batting average also did not fall from the efforts of his last 16 Tests.

    So his last 16 Tests saw him maintain both batting AND bowling averages equal to his career averages. ONLY his bowling strike rate suffered. Not to mention that 10 of his last 16 Tests were on the subcontinent in a spin-tailored era of wicket preparation.

    Compare that to Imran, who didn't bowl much at all in his final 2 years (his strike rate would have suffered if he did), and to Botham, who hardly played 15 Tests in his final SIX years of Test cricket.

    I'd say Kapil's all-round Test results in his last 2 years/16 Tests beats both. [[ Kartik What you say re Kapil's 17 tests on averages is true. But he fell off on his wicket tally and that is a factor here. Also we are comparing Kapil with a host of world class bowlers, not the battery of Indian fast bowlers. I have given credit for Kapil's (and others') wickets in the sub-continent. More than that it is difficult to do. Incidentally Imran does not lose because of his batting-only tests since he did not bowl at all. I don't think there is anything wrong with that. Ananth: ]]

  • Kartik on July 25, 2009, 0:33 GMT

    Bruce Reid and Colin Croft : Just 27 Tests each.

    Shoaib Akhtar and Geoff Lawson : Just 46 tests each.

    Yep, we absolutely need to raise the cutoff to 200 wickets. We cannot have people in the top 20 with just 27 Tests.

    Waqar Younis' performance in his first 27 Tests would have put him at #1 by far. Kapil in his first 50 Tests would have cracked the top-10.

    A higher cutoff is imperative.

  • Kartik on July 25, 2009, 0:23 GMT

    Ananth,

    Not to take the fun out of this analytical exercise....

    But we would probably get a similar ranking by taking the cricketratings.com scores (formerly PwC ratings), calculate the weighted moving average score over each bowler's career, and rank them accordingly (once you decide on a wicket cutoff like 200 or something).

    I doubt their formula for bowlers (and batsmen) is very different than yours, as they are known to take into account match result, batsman quality, etc. Longevity is the only factor not accounted for theren - again, the wicket cutoff captures that to a large degree.

    Reinventing the wheel and discussing it is fun for all of us, but the same result can be gotten off that preexisting source, I suspect.

  • Kartik on July 25, 2009, 0:17 GMT

    I agree with the others that longevity weight needs to be more. Unbreakables like Kapil and Walsh ultimately served their teams better through dependability than fragile bowlers like Bruce Reid and Shoaib Akhtar.

    However, this need not be a complicated exercise. The modern-era table needs to be cut off at 200 wickets. I think that is fair for the modern era, and includes Roberts and Thomson.

    Ananth, are you absolutely decided on not combining the list into an all-time one, like you did for batsmen and all-rounders? If you do, my earlier suggestion of a 189-wicket cutoff (to include Barnes) still has merit.

  • Abdullah on July 24, 2009, 19:40 GMT

    Hi Ananth

    Thanks for valuing my suggestion and acknowleding in your posts. The peer comparison does indeed look interesting, specially Steyn's position.

    I have a few points, which mostly all been raised earlier by other people.

    1. The weightage for no. of wickets should definitely be higher than 5%. It is an important measure for bowlers, who can remain fit and perform for a longer time, rather than bowlers like Akhtar, who are deadly in one match/innings and then break down. Longevity should be aknowledged

    2. Limits for bowlers in the balanced era should be raised to 175 or 200 while keeping the maintaing the cut-off for the other eras. This would assume that a "good" bowler will take around 3-4 wickets per match i.e. 50/60 tests which with the amount of cricket being played in this era is around 5 years.

    3. As mentioned by Marin Crowe and Kunal, any further division of eras should split the 1990s and 2000s as the conditions have changed alot.

  • love goel on July 24, 2009, 17:17 GMT

    1.Warne has more wickets overseas than Murali, but this is also because of the number of matches they played at home vs away. England plays a 5 matches series with Australia while they play only 2 matches series with Sri Lanka. If murali has more wickets against Zimbabwe and bangladesh , Warne has a whole lot of wickets against england. If Murali were to play similar number of matchs against England, his stats will be even better. 2. Murali performance has not been so good in Australia but that is also due to strong australian batting lineup 3.Also Australia have won more matches meaning they have bowled out the opposition twice meaning Warne had much more oppurtunities to pick wickets. And yes, warne picking up wickets is one reason for this , but lack of supporting bowlers and what total your batsmen put up on the board is a huge factor in this.

  • love goel on July 24, 2009, 13:48 GMT

    One reason for the low number of comments is that most of the factors have already been considered leading to a much refined and exhaustive rankings. I still feel this needs a little tweaking like 1. Longevity weight needs to be more. Most of mis-placements mentioned by the readers is due to bowlers with less number of wickets ranked above/along with bowlers with much higher number of wickets. Also , I feel that longevity is much more valuable for a bowler than a batsmen 2. Long bowling spells being equivalent to a wicket: Instead of taking overs bowled take percentage of overs bowled in a innings. If a team is dismmised in 200 overs, the bowlers gets more advantage by bowling more overs though it was because they were ineffective. 3. Different points for dismissing a batsmen at 0 or 100: This need not be considered. This is aptly captured in the bowler average/number of overs bowled/strike rate.Dismissal at 100 means more overs bowled which you are rewarding instead of penalising. [[ Goel 1. I started the wickets tally at 10 points, then changed it to 5 points and 5 points for wkts/innspell. Let us see if there is a case for recognizing this factor more. I am not very sure. Better to raise the qualification bar. 2. Your example of 200 overs is the specific effort which I want to reward. Take the case of the Slk innings of 952. Kumble, Chauhan and Kulkarni toiled for 220 overs for 3 wickets. Do you mean to tell me that they should each get credit for 1 wicket and nothing more. That is cruel. It was not their lack of effort but the concrete-like pitch at Premadasa. So let us be bit more considerate and reward the 20 hours of hard work. 3.I am convinced that dismissing Ponting at 0 cannot be treated the same as dismissing him at 150. This difference is not captured anywhere else but in this specific measure in Match performance. Your last sentence is quite confusing since you are mixing one point with a totally different point. Ananth: ]]

  • Zeeshan Ahmed Siddiqui on July 24, 2009, 12:43 GMT

    Another nice article!

    See some facts 1. Warne took 319 wickets at home and the remaining wickets are 389.

    Murli took 472 wickets at home and at away 298. In Asia his total wickets are 581. His total wickets are only 188 excluding Asia mens 75% wickets in Asia. 2. Warne took 17 wickets against Bangladesh and Zimbabwe.

    Murli took 176 wickets against them.

    3. Murli bowling average is 75.41 in Australia whereas

    Warne is champion in Australia due to home ground advantage. 4. Warne is more charismatic bowler as he can turn any match from any worst situation to his team’s favour. In batting this quality only found in Brian Lara. 5. Batting advantage to Warne, as he scored 3154 runs in test too. 6. I think Warne is more dangerous bowler on pitches which do not assist spin bowling

    7. The only blame against Warne is that his bowling average is 47 against India but almost all leg spinners are flopped against them. Murli is better as his average is 30.03 against them. [[ Zeeshan I do not deny any of the numbers you have given. These are all correct. However the fact is that Murali also captured -600+ wickets against top teams. Also every wicket he has captured against the weaker teams has been downvalued. The fact that he plays in a weak bowling team has reduced the points further. Also the wickets taken in the sub-continent do not get extra credit. Only the 188 wickets do get. Warne gets a lot of additional credit for the non-sub-continent wickets. His overall numbers, even downscaled by the Bng-Zim wickets, are outstanding. Also remember that Zimbabwe in many of the Sri Lanka matches was quite good with the Flower brothers at their peak. All said and done, two great bowlers, as good as Lara and Tendulkar, in batting. Let us savour and enjoy their great skills as I have been exhorting the other readers on the two great batsmen. Ananth: ]]

  • Harsh Thakor on July 24, 2009, 12:27 GMT

    Great Work,Ananth. I congragulate you for rating Dennis Lillee ,the best fast bowler of his era.He was the most versatile and complete fast bowler ,taking wickets aginst the best oponents.He deserves to be ahead of Marshall and Mcgrath,who have been rated higher by many judges in computer rankings.Deservedly Ricahard Hadlee has gained his place as he solely bore the brunt of his team's attack with an outstanding record.Murli and Warne too are correctly placed. However to me it is inconceivable that Malcolm Marshall is rated below Imran Khan or even Richard Hadlee.He was outstanding statistically in all conditions,against the best batsman .There is no greater challeger to Lillee for the title of the best fast bowler.Even Mcgrath should have edged Imran Khan,considering the conditions were hardly helpful in the modern era. Andy Roberts ,could have made the top 10,being the most versatile and complete of all West Indian fast bowlers,brilliant on the deadest surfaces.

  • Alex on July 24, 2009, 12:00 GMT

    Ananth - In addition to the "distructiveness" factor I suggested earlier, could you please add a "consistency" measure as well? This is on the lines of your test batsmen analysis where you took splices of 5 tests. As opposed to using the # wickets, you could use the bowler's own rating score (51.3 for Murali) and replicated your test batsman computations.

    Also, you could indicate the #matches (or 5 consecutive matches) over which a bowler had a rating exceeding, say, 60. This is a slight twist on #matches with 10 wicket haul etc.

  • V.J.Raghunath on July 24, 2009, 9:19 GMT

    amazing analysis-ananth the number of factors considered and the exhaustive research is truly amazing-need more time to go through in detail-saw it only today. i agree with your three periods of evaluation. where is Kapil Dev? to me,he is India's greatest bowler-nopacer before him had done anything comparable. and he never missed a test in his long career on fitness grounds-can you say that of any other pacemen-and he bowled on bone-hard pitches at home and in the sub-continent-and played enough one day cricket,to boot. should not fitness,not missing tests due to injuries or breaking down be a factor,especially for Fast bowlers. [[ Raghu With all these factors which are true, Kapil is not placed very high. His numbers are not that great, especially as he is pulled down by his last 17 tests. During this period he huffed and puffed his way, often painfully to the watchers also, with the sole aim of breaking Hadlee's record and gathered a measly 35 wickets. Ananth: ]]

  • Arjun on July 24, 2009, 7:49 GMT

    Hi Ananth,

    These changes could be made in follow-up post....

    1)As you have said, peer factor can replace Balls/wkts ratio. 2)Bow. accuracy index could be replaced by Own RPO/Team RPO(similar to balls/wkt ratio) it will mostly benefit spinners. 3)Bow. S/Rate could be replaced by SQRT of product of bow. avg. and s/rate. it will correctly benefit fast bowlers who have low avgs. eg.marshal, ambrose, mcgrath, donald etc. At present your analysis gives separate credit for S/rate and RPO but none to low bowling avgs. Point no. 2 takes care of Eco. Rate and point no.3 of avg. and s/rate.

    For 'base points for balls bowled' another alternative is to give credit only if bowler has taken 0 or 1 wkt in innspell. If he has taken two or more wkts in innspell no credit for long spell should be given.

    Arjun. [[ Arjun I will talk about the follow-up article in my interim but important post on "Test players - comparisons with peers". At this stage I will only talk about the Bowling average. You will see that the total of StRt + RpO points exceeds 11.00 for Marshall, Hadlee, Croft and Garner, all bowlers with great averages. So the split is working well. Ananth: ]]

  • Ramesh Kumar on July 24, 2009, 6:58 GMT

    One time periods, I have a slightly different view. The joy of statistics tool will be lost if we split it too much. It should not be an exercise to state the obvious. In my view, key change agents should be identified and used as the criteria for splitting the era. Issues related to playing conditions, rule changes, format changes etc should be the driving factors than the simplistic ones. Here people who are into history of the game can throw more light. Looks like Martin Crowe(What a pleasure to share the discussion with such a great cricketer) certainly feels that the game has changed. With his knowledge and his discussion with players of earlier era, he could probably detect/identify certain game changers which will get us going on this point. The criteria need not be same for bowlers and batsmen. Then we will have the pleasure of seeing how the numbers stack up for different players. [[ Ramesh There is room for both, an all-encompassing table and period-wise, customized analysis. The idea of a Tennis-style Hall of Fame has lot of good points. One is able to appreciate greatness without resorting to mundane comparisons. Even now I will only say that Federer is the first amongst equals, Laver, Sampras and himself. Let us consult in depth and move forward. Ananth: ]]

  • Martin Crowe on July 24, 2009, 5:17 GMT

    Kunal makes a good point. The 6 equal time periods I suggested looks fair but perhaps it should be split into 1990-99 and 2000-09. Tendulkar and Warne would feature heavily in both era's.

    MDC [[ Martin As I had indicated this will be a collaborative unhurried effort. Thanks for igniting the flame. Ananth: ]]

  • Ananth on July 24, 2009, 3:52 GMT

    Ed (in particular) and others who have raised similar comments: I think I have found out the one factor which takes care (to a great extent) of the great bowlers who have lost out because of playing in very strong bowling teams. I have been working on the Players' performance comparisons with peer players, the "Abdulla" suggestion. I had done this for ODI Strike Rates. Now I have extended this to Tests. Let us leave batting out for the time being. In bowling I have done the comparison of each bowlers' bowling average with his peers' (from first test to last test-maybe over a thousand bowlers) bowling averages. The preliminary results are great eye-openers, to say the least. Given below is the top-10. Note the preponderance of top performance players. This could very well replace the "Achievement/Effort factor", the last column in the table. Lohmann(1.94), Barnes, Marshall, McGrath, Muralitharan(1.50), Garner, Ambrose, Wardle, Hadlee, and Steyn(1.47). These are first cut numbers and yet to be checked but you should get the drift. THis is a fantastic peer-to-peer ratio and is not influenced by any negative factors. Also note that even in this table the incomparable Murali and Hadlee appear, indicating that their greatness is indeed genuine.

  • Youvi on July 23, 2009, 22:00 GMT

    Ananth- I always wonder about the weightage given to home versus away games. Is it primarily a subcontinental issue ? For example, Lillee and Thomson who played most of their international cricket in their home country. Obviously both were superb fast bowlers irrespective of whether they bowled mostly at home or away. Statistically or otherwise, is it as important as we from the subcontinent think it is ? Obviously, individual regions/grounds do vary but that is a distinct entity from home versus away test matches. I am no statistician but I find your analyses fascinating and it is a pleasure to look at it in conjunction with the visual experience/memory of watching soem awesome cricketers or reading about them whether from current or past eras. [[ Youvi The home advantage benefit is minimal and is also team-strength related. Similarly the pace-at-subcontinent and spin-at-non-subcontinent is only a multiplication factor. These are all based on numbers. Ananth: ]]

  • Eddy on July 23, 2009, 20:12 GMT

    @ Salim....your last comment made me smile. @ Aanath... i must say your findings did disapoint me. Not that your analysis was incorrect but i had a strong feeling i knew the top 5, of was wrong. I thought Murali would be top and Marshall second. I was shocked to see Lillie, steyn,Khan and Haddlee(quicks) all above him. I think im right in saying Marshall still has the lowest avg for a test bowler that has taken 300+ wickets? Perhaps i now know how the Tendulkar fans felt when your findings put Lara on top. There's no analysis that will change my mind that Marshall wasnt the greatest seam/fast bowler, just as the Sachin fans refused to acknowledge Lara was 2nd after the Don. Funny, now the shoe is on the other foot i dont feel so good.

    Great work again Aanath, regards as always Eddy. [[ Eddy The only difference is that you have the sagacity to accept that this is only a number-crunching exercise. You do not think that I have done it deliberately and you do not throw mud at me doubting my parentage, patriotism, nationality et al. Where Marshall (and other greats like him) have suffered is that they had to share the spoils. This is reflected by Match perf points and the Perf Idx points. I have tried to address this partly by increasing the Match Perf if the own bowling strength is very good. But not enough. But in reality I am helpless and cannot think of anything more. The other basic problem is that there is no limit to what a batsman can do. He can theoretically score a lot of runs out of very few. In bowling there is a limit. Once again thanks for understanding. Maybe the follow-up will improve this particlular situation, if I get any pointers on this specific 3/4-strong-bowlers-in-a-team situation. Can you put your mind to it and suggest a simple idea. You will notice that in Bow Strt + Bow Avge combined Marshall has 11.32 points, the best, by a mile indicating that he has been credited for the fantastic average. Ananth: ]]

  • Alex on July 23, 2009, 18:02 GMT

    Salim - part of the reason must be that bowling is difficult to analyze so that it becomes difficult to take issues. For example, the fans and the press had field days when none of SRT's test innings featured in WISDEN 100 but how many of them noticed that none of Lillee's performances was included in the WISDEN 100 either? I feel Andy Roberts should be in Top 15 but cannot find a flaw in Ananth's metric to correct this situation; I just thank my stars that Holding made the Top 15.

    It's also got to do with the audience. Maybe the old-timer Aussies are not taking a look at Ananth's article. If they did, they might have issues with how Sarfraz is rated close to Lindwall and above Miller.

  • Kunal on July 23, 2009, 13:38 GMT

    Martin Crowes comment on various periods is pretty good. However I have my reservations about both bowlers and batsmen who made most of their “gains” in the 2000s. Clearly the 2000s have been a boom time for a few bowlers and several batsmen. So a player who got most of his wickets/runs in the 90s would fare quite poorly as compared to a player who reveled in the 2000s. Especially hurt “points wise” would be the ones who started around the late 80s and retired in the late 90s, early 2000s.

    For e.g. none of the very fine batsmen(Gooch,anwar,mark Waugh,azharrudin etc etc) or bowlers(waqar,Donald,akram,Ambrose,Walsh,saqlain etc etc) who retired in the mid/late 90s to early 2000s come in very high up.

    They are simply blown away by the excesses of the 2000s.

    If you take players such as Murali,Warne,Tendulkar,Lara etc and compare it to others such as Donald,Akram,Waqar,Saqlain,Gooch,Athers,Azhar etc in the 90s only then does it do justice to the “90s guys”….otherwise once you get into the 2000s they get blown away. I am pretty sure if you compare the 90s performances of say Tendulkar, lara and steve waugh they will be pretty similar. But the 2000s part gets them miles above Steve. Also it is almost certain that if steve was around in the 2000s he too would have piled on loads of runs like so many others.

    So, perhaps the 2000s require different treatment. Some more thought is due on how to make it a less raw deal for this period for the players who performed either only or very well in the 90s and then got bludgeoned away statistically by the players in the 2000s.

    So, if we take 1990-2009 as a single period we will end up with a “lost generation” of 90s performers.

  • Salim on July 23, 2009, 13:13 GMT

    Hello Ananth, its quite amazing that when you produced the first batting analysis you had over 400 posts and then the reworked version has over 120. It really stir up emotion and debate (much was around the Lara/Tendulkar debate and Viv's low positioning). Now we have the Bowling analysis and at the time of me wrinting this two days after you presented your findings there are 50 odd comments. 1. Is this because cricket is a batsmens game? 2. Is this because of the Lara/Tendulkar debate? Or 3. is this because the Indian fans do not have a player they can argue a case for being No1 unlike the batting analysis?

    thank you Salim [[ Salim You have hit the nail on the head. I also noticed the anomaly but waited for a reader to make a comment. But I am not too unhappy. I am at peace. The answer, to the points raised by you (numbered by me), is 20%, 50% and 30% respectively. Ananth: ]]

  • Ramesh Kumar on July 23, 2009, 9:46 GMT

    Ananth,

    excellent work and a lot of thought has gone into this. Some random thoughts on this subject.

    1.If you do a sensitivity analysis of Match performance components, from the bowlers data, do you see a big change in the overall rankings by changing a specific parameter(s)? The intention is not to tweak it for some players sake, but more to see if there is any criteria changing the dynamics for certain players. If yes, one can probably review that factor's role in this analysis.

    2. Bowlers with high rankings, but played in less matches in the current era--any specific insight? In the current era, if Reid & Lawson are so good, why did they play less no. of matches when there is an opportunity for more.

    3. One of the comments was leading upto a request for another analysis-best batsmen against best bowlers. I think the statistics should not be stretched to do that. The players perform to match situation, they perform certain roles as a team member and they play against best available 11 from opposition. The proposed analysis will throw some numbers from which correct inferences cannot be made.

    4. On the eras, I think we need to look at specific conditions of play as the criteria if they can be identified. Some of the thoughts could be timeless tests, uncovered pitches, usage of helmets etc. bouncer restrictions, neutral umpires etc. Also the era of surplus of one day matches(1980 onwards) is my favourite criteria as I believe even while evaluating test batsmen/bowlers, we need to appreciate the role of one day matches. The players need to play more matches, use different skills, more injury prone etc.

    Once again congrats to Ananth on his excellent work.

    regards,

    Ramesh Kumar

  • kalyan on July 23, 2009, 8:47 GMT

    And i want to see your analysis on most valuable player for the team, like karthik said

  • Arjun on July 23, 2009, 8:34 GMT

    Hi Ananth,

    Is J.J.Ferris's matchperf. average 9.94 ? it is equal to carl hooper (bow. ave. 49.0) whearas ferris ave. is 12.0 ?

    Arjun. [[ My minimum number of innspells is 30. Ferris has bowled only 16 times so 30 is used as the divisor. Possibly I may have to change this minimum for the pre-ww1 era considering that the minimum wickets is reduced to 60. No great impact however. Will change in the follow-up post. Hooper's, on the other hand, is a genuine number !!! Ananth: ]]

  • Alex on July 23, 2009, 6:54 GMT

    Ananth - is it possible for you to add a column on "distructiveness (or cumulative distructiveness)"? The idea is as follows: look at the inspells that (i) exceed a certain threshold score, and (ii) include 4 or more wickets. Add these innspell scores and divide the total by the number of matches the player played (or by only the # matches in which these spells were bowled ... to bring out "when he got hot, he really got hot"). I added a filter of 4 wickets (as opposed to 5 or 6 wickets) to be fair to bowlers like Roberts, Croft, Garner, and Holding ... Marshall suffered less from the in-competition after 1986. [[ Alex Fascinating request. Will look at your comment and do what I can. Ananth: ]]

  • Arjun on July 23, 2009, 6:48 GMT

    Hi Ananth,

    I feel that the weightage you have given to 'balls bowled in innspell' (for base points) is too high, as you have mentioned 25 overs = 1 wkt. However this will give double credit to spinners who usually bowl long spells of 40-50 overs which fetch them more wkts (five wkt hauls eg. warne, murli). On pitches which are conducive to fast bowlers, generally fast bowlers will pick more wkts in less overs (teams getting all out in 40-50 overs) and will not get credit of bowling more overs as compared to spinners. It will be lot better if weight given to balls bowled is reduced (say 0.5 wkt for 25 overs).

    Arjun. [[ Arjun What you say makes sense. I am certain that I want to recognize those long but quite low-yield spells since these act as the base for the more attacking bowlers. Possibly giving credit for 25 overs (an approximate figure) as equivalent to a wicket might be too high. I will look at this and will evolve a more equitable allocation in the follow-up article. Thanks Ananth: ]]

  • Martin Crowe on July 23, 2009, 4:41 GMT

    Ananth,

    The more I think about it the more it would be wonderful for the game if we did break up the eras and celebrate the best of each. As you can see I am very much into the history of the game, have been since age 8.

    Anyway just to improve on what I originally suggested I believe the best way to set out the eras might read like this;

    Grace-Trumper era (1877-1909) 23 yrs Hobbs era (1910-1929) 20 yrs Bradman era (1930-1949) 20 yrs Sobers era (1950-1969) 20 yrs Lillee-Richards era (1970-1989) 20 yrs Warne-Tendulkar era (1990-2009) 20 yrs

    MDC (Martin Crowe) [[ Martin Excellent idea. The only ryder I would add is that I am not going to rush into it but will do it in a consultative and unhurried manner with inputs from interested knowledgeable readers, including you. Once again, thanks for the insight. Ananth: ]]

  • sushant singh on July 22, 2009, 18:02 GMT

    Mr. chandu_sistla on any given day & in every rating I am looking for past 15 years IMRAN KHAN is bigger bowler than anil kumble in every aspect.You have your personal opinion but in every rating IMRAN is there among the tops & by the way it will be interesting to know from where got to know that IMRAN is not a big bowler. MR. ANANTH -- another nice work by you although warne is my personal favourite but your article satisfies me good to see akhtar in top because he has bowled brilliantly whenever he is fully fit & waqar also. your work will not satisfy some but will satisfy many like me .GREAT WORK KEEP IT UP [[ The more I see it the more I like Martin Crowe's suggestions. There is no way we can compare two great bowlers of different generations. I myself have only brought them to a common platform, based on what I percieve are a number of valid measures, without ever saying that one is better than the other. I would only say that, based on these parameters, the two bowlers secure these many points. Let me put it on record. Imran and Kumble were the giants of their eras. Period. Ananth: ]]

  • Alex on July 22, 2009, 16:52 GMT

    Ananth - What is the degree of uncertainty associated with the final score? With your test batsmen's analysis, I think was 5% (you reduced it to 5% from 10%, I believe).

    Also, with the exception of Shane Warne, everybody else in the Top 5 has been a great mainstay of his side's bowling for most of his career. So, it appears that your metric has a positive bias for such bowlers. Is that true? I suspect Marshall, Akram, and Roberts would probably move up in your list if this bias were somehow removed.

    You mentioned All Time Australian XI. So, I checked Keith Miller's ranking in your tables. He is below Sarfaraz (and Mankad), who is almost on par with Lindwall!! If such details get ironed out, this analysis can be built to categorize the "best" duos/trios/quartet, and the best bowling attacks over 5/10/20 tests etc. [[ Alex Lillee shared the spoils with Thomson, Lawson. But I agree that he was the mainstay of the Aussies than McGrath. Anyhow, how does one penalize because they took a greater load than their compatriots. Pl remember that I also have a measure to specifically take care of own team's bowling strength. Ananth: ]]

  • Vivek on July 22, 2009, 16:48 GMT

    Sorry, I couldn't send in all my comments in one post, but I think it would be a good idea if you got the percentage wickets at home and away. I don't think that Harbhajan Singh should even make it in the top 20, because of his ineffectiveness on pitches outside India. Thanks a lot, Vivek [[ The home/away differential is implemented in the Match Perf Ratings. I had the home/away wickets in the career calculations also. However took it away when I reduced the weight from 10 to 5. the 5 going to the performance based measure wickets/innspell. So now the difference will not be much. Anyhow Harbhajan is 42.51, McDermott 42.50 and Alderman 42.49. You cannot get these any closer. Ananth: ]]

  • romel on July 22, 2009, 15:40 GMT

    Great work ananth..good to see spinners get their due instead of their lower strike rate than the pacers..a request...would you please post another kind of ranking of the test batsman solely based on their performance against the bowlers passed your criteria (minimum no of wickets and longivity)...the title may not be the 'great ones'...but could be 'performance of good batsman against the good bowlers'....my suggestion is to ignore the not out figures and even the match situation (which are already counted in bowler's ranking)...make the ranking solely based on the career performance on batsmen against these bowlers from respective era...my point is if we count your metric ideal theoratically (which is very close in this article i believe)...we can measure other against this

  • Gopal on July 22, 2009, 15:12 GMT

    Hi Ananth,

    Another great article. I see taht you've taken into consideration the state of the series, but I was wondering if you've taken into consideration the match situation as well. For eg. if you take Lord's 2005, after the Aussies were bowled out for 190 it was definitely advantage England, but McGrath proved what a great bowler he is and within his spell of 7-8 overs he'd taken 5 for just 2 runs!! I feel such performances should be ranked higher than ones where the bowlers have runs on the board to defend. Also, could you please elaborate on the 'innspells' factor which is a little confusing to be honest.

    Thanks and Regards, Gopal [[ Gopal In the match you mention, McGrath gets a significantly high additional weight as his team was defending a low total of 190. Incidentally McGrath's innspell is rated quite high because he also dismissed the first 5 wickets. Contrast this with Kumble's wonderful spell of 8 for 141 against Australia. He gets a very low weight on this measure since he was defending a 700+ total. However his effort is rated to be in the top-50 ever. I define an "innspell" as the complete bowling done by a bowler in an innings. I used to use the term "spell" but a few people argued that this should be one continuous stint of bowling. I love the term since it is innings+spell. Ananth: ]]

  • Salim on July 22, 2009, 14:42 GMT

    Your amazing analysis have shown Lara and Murali to be statisically the greatest modern players. Murali admits that Lara was the greatest he bowled to. Lara's record against Sri Lan and Murali is excellent. Do your findings suggest that Lara was the greatest modern cricket player?

    thank you Salim

  • Marcus on July 22, 2009, 14:10 GMT

    Thank you for the clarification Ananth. It would be interesting though to see some set of standardised statistics for these bowlers (ie. average and strike rate) such as what you did for the "Greatest Batsman" analysis.

  • Clinton on July 22, 2009, 14:04 GMT

    Ananth, “Martin Crowes Theme” is the most “sensible” comment we have here. We are so consumed by tweaking a stat here and there that we simply cannot see the big picture. It is a touch arrogant to assume that we can use numbers to “equalize” different eras and conditions into some “level playing field”…we cannot. Also the assumption behind someone being ranked above another is that the particular person would have done better. Which is why perhaps the initial perplexity about Richards lowly position. Even the latest table implies that lara/tendulkar/ponting etc were “better batsmen” than Richards and Gavaskar. A similar argument has been applied to tennis, wherein the point (very valid) is made that if Federer cannot even beat Nadal , how can he be the best of “all time”-the term implicitly means that Federer would beat ANY past player – but he cannot beat a player of his own era (so we have a conundrum)…No such complication existed for Sampras ,who on his day walloped everyone in his era (and had the good sense to retire at the appropriate time).So the best that may be asked of any sportsman is- how did he do in his own “era”?…Please note , the players era also denotes a particular playing period. Federer may go on playing till 35,with a dodgy back and thus take the risk of being beaten by every new kid on the block. As Martin Crowe says we have to consider individual eras.

    So, similarly the best batsmen on your batting list ,viz.Lara would be expected to do better than the others even in the prehelmet ,no bounce rule etc period. This ,to most cricket observers, having observed laras discomfort against several top class pacers is extremely difficult to believe.

    So ,before we get back into our “box” and tinker around endlessly with the stats…lets get the big picture straight.

  • Atif Iqbal on July 22, 2009, 13:27 GMT

    great analysis. couple of observations from a Pakistani perspective:

    the order of Pakistani fast bowlers is accurate but the spinners is not that accurate (re: Danish Kaneria is rated higher than Abdul Qadir)the performance ranges between all spinners Saqlain, Mushtaq, Kaneria and Qadir are too much.

    I do not want to start the debate of whether Saqlain at 25 makes sense (compared to Harbajan at 20) because that is more of an intiutive argument.

    Some suggestions that I want to make are

    1. type of dismissal should be considered i.e. lbw/bowled should have a higher weightage than caught/stumped (its more of a forced and un-forced error kind of an argument)

    2. bowlers hunt in pairs and although you have partially addressed this by taking into consideration the length of spells, I believe if try and capture the impact of Wasim/Waqar or Saqlain/Wasim or Waqar/Mushtaq we will see an improvement in bowling performances (applying pressure from both ends)

    appreciate it. [[ Atif Did you miss my article published a few months back titled "Good in isolation, great as a pair" which specifically looked at the bowling pairs. As far as I am concerned Harbhajan and Saqlain are almost perfectly equal. The 5-place difference is neither here nor there. I personally feel Qadir is fantastic classical leg spinner. Kaneria's figures are better and Qadir's away record was average. However please see that they are only 5 places apart. Ananth: ]]

  • Sandeep on July 22, 2009, 13:18 GMT

    Continuing from my previous post

    Spinners getting a lift up, when bowling outside the sub continent- now does that include turning pitches like Sydney as well?

    But, as I said- great job- none will dispute that the top 50 need to be there- some may suggest exchanging places, but then that is what makes this debate so interesting [[ Sandeep Spinners might get undue credit at Sydney and Pace bowlers undue credit at Eden Gardens during 1999. These things will happen. But in most cases the premise is facts-based and correct. Ananth: ]]

  • Sandeep on July 22, 2009, 13:14 GMT

    Hi Ananth- first of all, congratulations for coming up with another excellent analysis. You have taken almost every variable into consideration- well done. Just a few observations- not to be taken as criticisms. Probably Steyn and Bhajji are a bit highly ranked in the top 20- they r just midway through their careers while the rest in the top 20 have finished their careers (I know Murali is still playing, but he is special)- surely they might end up with great careers, but then nobody knows. And I was surprised that Messrs Shoaib Akhtar, Geoff Lawson, Bruce Reid and Colin Croft r ranked that high- good bowlers no doubt, but I am not sure if they played long enough to guide their teams to victories. And for batsmen dismissed, you say that Anderson gets more points for dimissing Ponting for 0 rather than Panesar who does it when he is on 150. How about Swann, who dismissed Clarke for 136- surely that should rank very high, as it dashed all Aussie hopes of a great win. [[ Sandeep It is impossible to work out all algorithms and reward the specific 136 dismissal (you could as well have referred to the little master's 136 at Chennai - credit to Saqlain) as against some other similar dismissal in a match where the bowling team might have lost. Pl remember that it is easy for you to pinpoint Clarke's dismissal. My program has to work correctly across 7700 innings. The complexity level goes way high to decide that the 136 was indeed a match-winning dismissal. But Swann will certainly get the deserved credit because of the won match, quality of wickets, the supporting level played and so on. So I do the next best thing. Reward low dismissals and give credit at career level all for dismissals. Not a perfect situation but I have never claimed so. Also pl remember that this is one of the 11 or so measures. Ananth: ]]

  • Alex on July 22, 2009, 13:13 GMT

    MDC - As per Ananth, you are indeed Martin Crowe, the NZL batsman. If so, it is super to have you comment on such statistics. I am not a kiwi but, as a kid, I used to get up at 3AM and midnights to watch the NZL top order bat just so to ensure that I don't miss out on one ball of your innings. You were my favorite batsman of the 80's! Wish you the best in all endeavors!

  • Yash Rungta on July 22, 2009, 12:43 GMT

    Cont. from previous post

    Lets take Kumble vs. Ambrose. Both are great bowlers in their own regard, but I'm sure everyone will agree that Ambrose needs to higher than Kumble. Looking at stats, I don't see any reason why Kumble is better than Ambrose.

    Average: Kumble 30 and Ambrose 21 Strike rate: Kumble 66 and Ambrose 55 Economy Rate: Kumble 2.7 and Ambrose 2.3 Quality of Opposition: Kumble has about 18% of his wickets against Australia whereas Ambrose has about 30% of wickets against Australia. But then South Africa was dominant from 95-2000 so this evens out fo Kumble. Home/Away: Kumble has 350:269 whereas Ambrose has 203:202 Win/Loss: Kumble has won 43(out of 132) matches whereas Amborse has won 44(out of 98) matches. This has a bearing on match performance.

    I've seen a LOT of matches where Ambrose has single handedly defended targets of 120-200 in the 4th innings. Kumble, though a genius, hasn't been able to match Ambrose.

    Something needs to be done to the criteria. What say?

  • Yash Rungta on July 22, 2009, 12:30 GMT

    I myself was thinking of a Most Valuable Player like 'Kartik' suggested.

    It might make sense to bifurcate 2 eras within 1970-2009. I believe it is essential to separate the Kapil Devs,Bothams, Imran Khan, Lilee, Hadlee etc. from the Ambrose, Akram, McGrath, Warne, Murali, Kumble etc. All of the latter made their debut 1985 or later and all of them played in the 21st century, so you could use any of the 2 criteria to separate them.

    To be considered a great, you've atleast got to have 200-250 wickets. 100/125/139/150 is just not enough. Maybe 150 wickets in the 1970-1985 era was good enough, but in the current modern era, 200 is the minimum.

    Also, 5 points for the 'no. of wickets' factor is very less. In batsmen, you had 15 points for runs and that is perhaps why you got the actual good batsmen at the top. Why not have the same weightage in the bowlers as well for longetivity factor?

    Anyways, having Kumble(avg about 30) higher than Ambrose(avg about 21) is not understandable.

  • Alex on July 22, 2009, 12:06 GMT

    Ananth - A few conundrums. Bruce Reid at #13 ... a top class bowler but, injury plagued, his career was not that different from Ian Bishop's, who is 39th. What caused such a large gap between them? Also, I understand your metric, but the low ranking of Andy Roberts is a bit tough to look at. To many, including the WI bowlers, he was the leader of the WI quartet. This probably illustrates that stats, no matter how exhaustive, can never tell the full story.

    MDC - I do not believe that you are indeed Martin Crowe. But if you are, it was such a joy to watch you bat! [[ Alex The main difference is the low number of innspells bowled by Reid, much better Wkt Avge and a very high Performance factor of 1.50. He is one of the most consistent bowlers ever, having only 2 unsuccessful innspells. It is indeed Martin, wrongly placed by me as the second best NZL batsmen ever, but in reality the best by a few points. Ananth: ]]

  • amit sinha on July 22, 2009, 11:17 GMT

    i would suggest simpler measures for both batsmen and bowlers. for batsmen - take the averages and multiply with the strike rate to get the effective average. this gives a good measure of the effectiveness of the batsman. for instance a batsman with an average of 40 and a strike rate of 90 will have an effective avg of 36, which is superior contribution to a batsman with an average of 25 and a strike rate of 110 (effective average - 27.5)

    For bowlers lower averages and lower strike rates (balls per wicket) are considered desirable. again multiply the two to get effective average. a bowler with an average of 22 and a strike rate of 60 will have an effective average of 13.2, whereas a bowler with an average of 25 and a strike rate of 50 will have an effective average of 12.5 which is superior. you want lower average and faster wickets - marry the two measures.

  • Arjun on July 22, 2009, 11:09 GMT

    Hi Ananth,

    What are the numbers of Mitchel Johnson(now since he has completed 100 wkts), Stuat clark(91) and shane bond(79) ?. Bond and clark average around 22 in this batting era.

    Arjun. [[ Have not updated the last test to my database since I wanted to preserve the report-base database for a couple of days. Will post an update for you within a couple of days. Ananth: ]]

  • Arjun on July 22, 2009, 11:03 GMT

    Hi Ananth,

    Warne's Matchperf. points is 22.52, it is 20 % better than Mcgrath(18.77). I am sure there isn't any computation mistake. 1.Mcgrath(0.81) has captured more top order wkts. as against warne(0.72). 2.Mcgrath's batsmen dismissed ctd avgs. is 19.22, warne's 18.47. This means mcgrath dismissed better quality batsmen as comp. to warne. 3.He has better average, s/rate and Eco. rate than warne.

    How can warne's match perf. be better than mcgrath (I am not saying that method is wrong but the diff. of 20% is too big) ? Is it something to do with long spells warne bowled as compared to mcgrath ? [[ The key factor is that McGrath has averaged only 2.33 wkts per innspell which is 16% below Warne and this has, in turn, caused a MP average reduction of around 22%. Arjun. [[ Arjun I was also intrigued after seeing your comment. The difference is that McGrath has bowled (proportionately) more innspells than Warne. The following table will explain this. ISP Wkts W/I MPPts MP/I 262 708 2.70 5900 22.52 241 563 2.33 4523 18.77 Ananth: ]]

  • chandu_sistla on July 22, 2009, 9:41 GMT

    With due regards to your analysis, how can the greatest bowler of all time be placed in third position? And only today I found that the likes of Imran Khan are much bigger than Anil Kumble. Also thanks for enlightening me that Kapil Dev and Courtney Walsh are not good enough to be in Top 20 list. [[ Chandu Someone else might say Lillee was the greatest bowler ever, someone else, Holding, some one else, Kumble, maybe even Syd Barnes, mebe even Muralitharan and yourself, Warne. And so on. I am sure you get the point. Ananth: ]]

  • Al on July 22, 2009, 9:18 GMT

    That may or may not be the case. But its not just me.Several ppl have posted numerous comments along similar lines about the absurdity of directly comparing different periods. Also your batting tables themselves in the 80s,80s,2000s blatantly point out to the massive increases in runscoring and 50 plus avgs achieved in the 2000s. On a purely statistical basis...there are over a dozen batsmen avg.50+ now as compared to just 2/3 in previous periods.

  • Jeff on July 22, 2009, 8:57 GMT

    Continued...

    Then the world discovered fast bowling as a weapon – Lillee/Thompson & the Windies revolutionized cricket. This is the 4th distinct era.

    4. 1974- 2000/1 – Strike Rates suddenly dropped to the lowest level since WW1 (SR 70) but batsmen also started scoring more quickly than ever (2.70 rpo)

    This seems like a long time for one era (over 800 tests) but when you try and split it up into smaller groups, the figures hardly change at all (eg the period from 74 to 89 has almost identical figures to that from 89 to 2001) suggesting that it’s pretty fair to group bowlers from this era together.

    Finally, things seem to change pretty suddenly again in 2001, starting the final era

    5. 2001- present – Batsmen suddenly start scoring much quicker (3.16 rpo) in fact every season since 2001, rates have been above 3 (vs only 9 in 176 seasons before then.) BUT, bowlers also find it easier to take wkts (SR 66) – the lowest since WW1.

    (Btw - great job Ananth !)

  • Jeff on July 22, 2009, 8:56 GMT

    I think I’m the reader that Ananth mentions in his reply to Martin Crowe. I suggested splitting the analysis into 5 distinct groups. I came to this conclusion by analysing 2 factors – bowling strike rate and economy rate (which combine to produce bowling average.) These factors tell us what conditions were like to bowl in and should help to judge what years to compare with each other.

    By looking at the change in these over time, I noticed 5 distinct periods:

    1. Pre WW1 – slow scoring (2.57 rpo) and lots of wkts (SR 58) make this (as Ananth says) a compelety different era to any other.

    2. Inter-war years – still slow scoring (2.56 rpo) but suddenly bowlers found it difficult to take wickets (SR 79)

    3. 1945-1973/74 – very defensive cricket – scoring rates even slower (2.34 rpo) but wickets even harder to come by (SR 82)

    To be continued...

  • Ben on July 22, 2009, 8:52 GMT

    In an expansion on love goel's 5th point, regarding dismissing a batsman on a high score. Say, hypothetically, that panesar was first brought on to bowl when Ponting was on 150, and removed him with his first ball. Surely the fact that Ponting had scored all his runs off of all the other bowlers, and couldn't beat one ball by Panesar, shouldn't count against him? [[ Ben You have missed the point. We are looking at bowlers here. Anderson got the Match perf credit for dismissing Ponting at 0 and Panesar nothing for dismissing Ponting at 150. That is all. Ananth: ]]

  • Arjun on July 22, 2009, 8:26 GMT

    Hi Ananth,

    In Top order/Middle order/Late order ratio, Top four bowlers are Zaheer Khan, Irfan Pathan, R Collinge and Karsan Ghavri; even Vaas is at no.10. All these are Left arm Medium fast bowlers. From this can we safely say that tailenders play LFM bowlers better(barring Wasim Akram) or that LFM bowlers usually dismiss Top order often with their incoming balls(lbw) or that LFM bowlers are less effective against late order. Maybe captians can get something valuable information from this analysis.

    Arjun. [[ Possibly the case of top order dismissals. Might also be the fact that the late order batsmen do not put bat to ball against LFM bowlers. Ananth: ]]

  • Al on July 22, 2009, 7:50 GMT

    Yes sir, we've been going on about the batting periods for analysis for several blogs now. The 2000s cannot be compared to the 80s or even the 90s. A decade wise comparison was suggested long ago...and reveals much more, especially with players with greater intersection of playing careers within the decades. [[ Al Your purpose was very limited, to show that one player was better than another. What Martin has suggested is from an overall point of view and has more credibility Ananth: ]]

  • Kartik on July 22, 2009, 7:26 GMT

    Had Imran bowled in those 8 tests (which were in years where his bowling form was otherwise prime), he would have taken 48 wickets in those 8 tests, and easily be the clear and solid #2 on this list.

    Nevermind also being the best batsman AND captain, among this list of Top-20 bowlers. An amazing cricketer.

    This is why, Ananth, that I hope in the back of your mind you are planning to combine all prior work done into the mother of all analyses - the Most Valuable Player (MVP). This will tell us everything - is Tendulkar more valuable than Warne? Gilchrist vs. Botham? Lara vs. Muralitharan?

  • Kartik on July 22, 2009, 7:16 GMT

    Bowling, unlike batting, has no 'outlier' player. Bradman has an average 65% higher than the next highest batsman with over 2000 runs, but there is no bowler (yet) who is so clearly ahead.

    Barnes might be the closest, but he was too early in Test history to really count. Grimmet and Muralitharan are not ahead of the second best by such a large margin.

    If, someday, we get a modern bowler who manages to sustain an average better than 15.00 over a full career (thus taking over 7 wickets per test and 2.5 wickets per ODI), it would reorganize the whole cricket world. Whichever team that bowler belonged to would probably rocket to the #1 ICC ranking position unless it was Bangladesh or West Indies.

    At the start, Ajantha Mendis had a strike rate that flattered to deceive. But he fell back to Earth. Two decades earlier, Narenda Hirwani sustained this for his first 4 tests, and not a moment longer.

    Perhaps we will see an 'outlier' bowler in our lifetimes.

  • Kartik on July 22, 2009, 6:59 GMT

    Ananth,

    I did say 189 wickets as a cutoff if there was to be a merged all-time list (in order to include Barnes). But as the list has be split into eras, the 1970-2009 should have a higher cutoff than the others. 150 is fine, but 200 would be ideal.

    The main thing is that some of the 15-20 ranked bowlers were not automatic selections for their respective teams, even when fit. Hence the preference of a 200 cutoff. Steyn will cross that soon.

  • Tarun on July 22, 2009, 6:35 GMT

    Could u break up the performance as home and away? [[ Tarun The Home/Away performances have been considered separately in the analysis. However the bowlers' careers have to be evaluated in their entirety. Ananth: ]]

  • Martin Crowe on July 22, 2009, 6:24 GMT

    Ananth,

    Hey you do great work and while it can't be perfect it's worthy of respect and acknowledgement. Yes it would apply to batting as well. On that note I would change Warne-McGrath era to Tendulkar-Warne era (1995-2005). Even to say the Lillee-Richards era (1975-85).

    The idea of a top 5 or 10 of those era's has real merit. Thats when you start to see the Hall of Fame revealed.

    MDC [[ Martin Your phrase "Hall of Fame" is the right approach. That is the only way Bradman's AND McCabe's widely varying contributions will be recognized. I have come to realize that it is very essential to compare great players with their peers. You would have seen the first attempt at this in the previous article on ODI strike rates, as suggested by Abdulla. I have started the work to get these database segments going for Tests. From your point of view you would like to be compared with Richards, Gavaskar, Javed, Greg Chappell and so on. That is the true comparison. The bowling was comparable and playing conditions similar. Once again many thanks. Ananth: ]]

  • Marcus on July 22, 2009, 6:09 GMT

    Very good analasis as always. I especially like the fact that some of the more obscure (but still outstanding) bowlers like Tayfield, McKenzie and so on get due recognition in this analysis. I also find it interesting that both Grimmett and O'Reilly have a higher ratio than Warne. Despite everyone saying Warne is the greatest leggie of all time (which might be true- I don't know) there's no doubt that all three would be very closely matched if they were able to play in the same area.

    I do have one question. Why don't the "Wkt Avg" figures in the support tables match their career averages? Are they weighted against the top 7 batsmen? Because that's the only reason I can think of. Sorry to ask, it's just that I can't find any reference to "Wkt Avg" in the article itself. [[ Marcus Wkt Avge is not the Bowling average. In fact I do not have the Bowling average at all. Instead it is split into its two components, the Bowling strike rate and Bowling Rpb since these are two distnict measures requiring independent evaluation. The Wkt Avge is the sum of the batting averages of batsmen dismissed divided by the number of wickets. It is slightly low for Warne since he has dismissed quite a number of average-average batsmen. Note the figures for Imran Khan or Holding. Very high values.

    Ananth: ]]

  • Adeel on July 22, 2009, 6:08 GMT

    hi. I am pretty much a silent reader of ur statistics articles. I always enjoy them. One thing i always don't understand is the criticism that is sent this way. I think most ppl need to think of these things as purely statistical data. They surely are passionate about their fav bowlers for their obvious charisma and presence on their field and while playing the way they intimidated and all. But they have to remember lol the lists don't have points for presence , charisma, stature. Freddie does not have great statistics but yet he is compared to the likes fo botham. Statistics are just a way to reflect a bit of ap layers ability. Plus this is not the last list of all time great bowlers lol so relax and let the dude breath :) ... ability wise i do agree that akram and akhtar are a world apart. but its statistics... just take it as another step in analysis which is not final. PeaXe~! [[ Adeel Many thanks. Flintoff bowled one of the best fast bowling spells of the past 5 years on Monday. Hostile, 90+ mph almost always and at the batsmen every ball. However when the analysis is done what I see (or the program sees) is the fact that he has captured below 3 wickets per test in 75+ tests and has had only 2 x 5-wicket hauls in an innings in 120+ spells and so on. Ananth: ]]

  • Al on July 22, 2009, 5:41 GMT

    @big gunz "Or are they penalized for having to share the spoils among the other great bowlers on the team?"

    Same problem with the batsmen.Certain team configurations yield more points.

  • Martin Crowe on July 22, 2009, 5:21 GMT

    Ananth,

    I was fascinated to hear the comments of Rod Laver following Federers 15th Slam win. He said its ridiculous to compare eras, instead its better to honour each era in its own right. He had no hesitation in saying Federer is the greatest of his era, since the Sampras era.

    I note you have created only 3 periods. Over 130 odd years of Tests can't surely be enough?

    Maybe to define what Laver rightly points out, you need to identify these eras, not giant periods. Eg its impossible to throw in the last 5 yrs with the mid 70's and 80's. Ie you can't include that great Windies era, with Bangladesh now beating the Windies, which is a major blow to Test cricket's integrity.

    The way I say it;

    Pre WW1 (1877-1915) Post WW1 (1916-1929), Bradman era (1930-1939), Post WW2 era (1945-1956), Sobers era (1957-1974), Lloyd-Windies era (1975-1985), Border-Aust era (1985-1995), Warne/McGrath era (1995-2005) & Now (2006-2009)

    Food for thought.

    MDC [[ Martin As you have correctly mentioned, food for thought. Maybe we place too much emphasis on one (at least I have now gone to 3) table(s) of great bowlers or batsmen. What you are suggesting makes a lot of sense. I had a similar comment from another reader earlier on similar lines. This applies to batting also. What bowling you and your peers faced was a world of difference to what the current batsmen must be facing. My fear was the number of tables and the length of posts. If I can find a way past it, this could be implemented. Maybe I should look at doing the work, as you have mentioned and then only provide the top-5 of each era and the links. Let me look at it carefully. The other significant benefit is that there need not be a single set of criteria and measures across the eras. The pre-WW1 era bears as much resemblance to today's era as Renshaw to Federer. At a pinch the Federer era could be compared to the Sampras era, but surely the Laver or Borg years were totally different. Many thanks for taking your time to come out with a top-player's view of an analytical conundrum. We are all indebted to you for that. Ananth: ]]

  • EAMiran on July 22, 2009, 5:21 GMT

    Off the topic - since when did Wasim and Waqar become "fast-medium" bowlers as denoted (LFM/RFM). In their pomp both were lightning quick, just like Lillee and Imran. [[ Miran These were picked up from Cricinfo database. Probably they might since have corrected. Will look at it. Ananth: ]]

  • Al on July 22, 2009, 5:17 GMT

    Another point which strikes me: Batsmen dismissed - based on his score at time of dismissal The new ball strike bowlers then suffer loss of points. [[ Al Not really.The new ball bowlers have a pretty good chance of getting a top order batsman quite early in his innings. Ananth: ]]

  • Al on July 22, 2009, 4:58 GMT

    A few immediate thoughts:

    1) An adjustment for pace bowlers bowling in the Asian subcontinent and spinners bowling outside: Warne will “suffer” for this since he never did well in India.

    2) Also, I feel “own team batting” strength is also an important factor. A better lineup putting up good scores makes it easier for bowlers to attack/defend totals. Similarly, in the batting analysis too this factor was missing .i.e. “own team bowlers”…and hence all the Indian batsmen inc. Gavaskar,Tendulkar,Dravid,Sehwag etc fall more than 10% short of the top batsmen on this one parameter alone.

    3) Balls bowled –to recognize bowling spells: killer blow for the express pace bowlers. Perhaps helps the “donkey” bowlers...Pathan and co?

    4) Home/away –incorporating relative strengths? Could you shed some light on how you are comparing whole “team” strengths? Does this include batting, fielding, bowling “individual” player stats?

    5) Result – again same as point 4. Have u done this for the batsmen analysis as well?

    6) Bowling accuracy- relative to the innings scoring rate: for this measure you will require to incorporate the strike rate of batsmen as well in your “overall quality of batting team [[ Al 1. Warne gains for the wickets he captures, including at home. 2. Recognition of the effort put in. 45 overs, whether by a pace bowler or spinner is a lot of effort. 3. The Team strength is the sum of batting, bowling and fielding (Dismissals based). 4. Has been done for the batting analysis also, in the follow-up article. 5. This has nothing to do with the batting strike rate. It is a relative measure. Pl study the examples shown. Ananth: ]]

  • Yash Rungta on July 22, 2009, 4:36 GMT

    I cannot believe Ambrose is not in the top 10. In my opinion, he should be in the top 5. You need to either increase the eligibility criteria from 100 wickets to say atleast 125 wickets or 150 wickets and/or you need to increase the weightage for the longetivity(no. of wickets) factor to say 10(from the present 5).

    People like Steyn(although a very very good bowler) and Pathan(a decent bowler but I cannot imagine he is better than even people like Cairns, Gillespie etc.) are still developing and they should take atleast about 100-150 wickets more before they're even considered being modern day greats. [[ Yash First Steyn has taken 170 wickets in 33 tests. An outstanding career so far. Pl see my response to Kartik's comment. Even a cut-off of 139 is fine. Ananth: ]]

  • Big Gunz on July 22, 2009, 4:29 GMT

    An era from the 70s to 2000 and there is only ONE Windies bowler in the top 10? and only 4 in the top 20? Mate I am as Aussie as a transplanted WACA rat can be in London but no way could that period not have at least 3 windies in the top 10 and maybe 7 in the top 20 even if it meant bumping an Aussie or two! Or are they penalized for having to share the spoils among the other great bowlers on the team? In such a case then Walsh should be a shoo in given he carried the team for so long on his own! As my Jamaican friend would say " Mr. DJ rewind dat back"

  • Kartik on July 22, 2009, 4:29 GMT

    As this is the first pass, there are bound to be tweaks. I would be surprised if the final list had Lawson, Reid, and Croft ahead of Garner, Pollock, etc.

    Perhaps a cutoff of 200 wickets is needed for the current 1970-2009 era? Bowlers with under 150 wickets are surprisingly high up. Some of the 11-20 players were not regular selections for their teams even when fit to play. [[ Kartik 150 will leave out O'Reilly, Gupte, Verity, Fazal (139), Croft et al. The idea probably is to pin our number, say as 139, as you yourself have said earlier, there is no need to have a round number. Let me wait for the responses. Afterthought: All the bowlers I have mentioned are in the middle era. So there should be no problems in increasing the cut-off to 150 for the current era. Ananth: ]]

  • Big Gunz on July 22, 2009, 4:28 GMT

    An era from the 70s to 2000 and there is only ONE Windies bowler in the top 10? and only 4 in the top 20? Mate I am as Aussie as a transplanted WACA rat can be in London but no way could that period not have at least 3 windies in the top 10 and maybe 7 in the top 20 even if it meant bumping an Aussie or two! Or are they penalized for having to share the spoils among the other great bowlers on the team? In such a case then Walsh should be a shoo in given he carried the team for so long on his own! As my Jamaican friend would say " Mr. DJ rewind dat back" [[ It is true that they had to share the spoils amongst them as many Australian greats had to do. It has been partly taken care of by having the strength of own attack built in. I am afraid you have to live with that. Ananth: ]]

  • Kartik on July 22, 2009, 4:20 GMT

    Ananth,

    Regarding the type of wicket captured :

    The quality of a fielding side makes a big difference in the success of a bowler. Warne had a much better fielding side than Kumble. If both had equal fielders, their career average and strike rate would probably be the same.

    The better fielding side pervades most of the parameters of the analysis (total wickets, match result, etc.)

    The only easy way to account for this is to put more weightage to wickets taken bowled or lbw, and less to stumped, caught behind, or caught in field. Hence, the fielder factor becomes smaller in the analysis. [[ Kartik That would be unfair to the spinners and to many medium pacers. McGrath and Hadlee specialized in finding the edge and that was a special skill. I considered and decided not to have the same. A wicket is a wicket. Inducing a mishit off a slower ball requires as much skill as finding the pads. Ananth: ]]

  • Qasim on July 22, 2009, 4:14 GMT

    Again an impressive table Anand well done. I have just one question which I couldn't clarify myself, have you accounted for the strength of a bowlers own attack?? For example Wasim Akram in his career had Imran Khan, Waqar Younis, Shoaib Akhtar (all in top 20) and other good bowlers like Saqlain Mushtaq with him fighting for wickets whereas someone like Muralitharan had only Vaas with him, who isn't that high either.... Obiuosly it would've been harder for Akram to take wickets as compared to Murali, perhaps this should also be added??? [[ Qasim If you read the article carefully, this factor has been clearly taken care of and let me confess that only after this measure was introduced did the great fast bowlers like Younis/Akram/Holding/Imran/Marshall move up. Ananth: ]]

  • Michael on July 22, 2009, 3:19 GMT

    Hi Ananth, good work. No real issues with most of the lists - very fine bowlers who all deserve to be there. I wonder, though, if there is a case for relaxing the eligibility criteria for the Pre-WW1 era, based on the admitted paucity of Tests that may have contributed to less bowlers reaching 100 wickets? Example is Fred Spofforth, who claimed 94 wickets in 18 matches over 10 years while missing very few matches. Or Joey Palmer (78 wickets from 17 matches) and JJ Ferris (61 wickets, 9 matches), both of whom have records that suggest they were fine bowlers who dominated the opposition but who were only able to play a few Tests a year. Every analysis needs a cut-off but to have 10 bowlers meet the criteria for a whole 37-year period seems a bit restrictive. Perhaps a cut off of 50/60 wickets, rather than 100, would be more representative of bowlers from this era. [[ Michael Good point. I must admit I was fascinated by the exact number of 10 reached. Will look at lowering the qualification criteria and update the article itself instead of waiting for the follow-up article. Michael Have since added the revised tables with 60 wickets as the cut-off. Ananth: ]]

  • Yogesh on July 22, 2009, 2:58 GMT

    But isn't Pathan above Lee, Gillespie, Srinath an obvious misplacement ?? And location based adjustment is something i always emphasise on incorporating but only if it can be in a just manner. Pace bowlers being given extra points in sub continent is quite a few times a false advantage like Ind vs Sa in Ahmedabad 2008 or Ind vs Pak Karachi 2005. Both the pitches were pace friendly. And again on a crumbling wicket, a pace bowler is as dangerous as a spinner. Rather what might be more indicative of the pitch is the avg. of the pace and spin bowlers in the match. But this could sometimes be erroneous when a particular bowler has an exceptional spell like Kumble in Sydney 2004.

    In bowling accuracy, shouldn't something be there for the bowler who outperforms not only his team-mates but even the opposition easily ? Again, the ex. being Kumble in Sydney 2004. [[ Yogesh Good idea to avoid unnecessary accusatory comments. I certainly respect your cricketing comments. I agree Pathan is somewhat high placed having just crossed the tape at 100 wkts. However he is ahead of most of the bowlers you have mentioned on the following points. - B/W Ratio - much higher figure, almost a spinner's number. - Wkt quality (lot less late order wkts). - Avge quality of batsmen dismissed (again a reflection of the top order batsmen dismissed). Maybe one must increase the lower limit especially for the current era. Re sub-continent wickets, I am waiting for comments on Kapil Dev. How he toiled on the sub-continent pitches and so on. Pl remember that I have gone on actual pace/spin distribution of wickets in sub-continent. Again it is easy to pick up one or two tests in either end of the world. I have gone on overall figures. Ananth: ]]

  • David on July 22, 2009, 1:25 GMT

    I would hardly call the the period from 2000-2009 part of the "balanced" era. Flatter pitches, bigger bats, restrictions on agressive bowling, pretty much any expert on the game would say that the balance has been in the favour of the bat. Add that to the fact that were very few attacks consisting of 4 better than average bowlers (usually one good bowler with a few average ones) meants less pressure on batsmen. Just look what happened when Australia ran into a four pronged pace attack in England in 05!

    Fortunately we are starting to see better bowling attacks (which is why SA and India have now beaten Australia) but I think you may want to reconsider the idea of a "balanced" period. [[ A better balanced era than the previous one, especially the period between 1940 and 1969. Certainly you cannot deny that the bat and ball have both have their continuous moments during this period. Ananth: ]]

  • love goel on July 21, 2009, 20:27 GMT

    great analysis ,with so many factors and in depth analysis.Some questions on the calculations 1. Why linear weight for career longevity. Certainly the number of bowlers with 400 wickets is less than half of bowlers with 200 wickets. 2.Also why 5 points for 1000 wickets when no bowler has done that? why not for the maximnum wickets at present(murlitharan) 3. Similarly when the cutoff is 100 wickets , then why not zero longevity points for 100 wickets and not .5 points(100/1000) 4.Same applies for match performance. Much easier to get 3&3 in both innings than 6 in one inning(whether 3&3 is as valuable as 6&0 is for debate) 5.Also the example of Panesar and Ricky Ponting.I think you need to calculate how many more runs Ponting score once he has scored 150 and compare that to what he scores when he is on 0. Batsmen like virendar sehwag are much much harder to dismiss when at 100 than at 0. [[ Point 2 is valid. Only problem is that the points, then will keep on changing with each wicket Murali takes. Pt 3. 100 wickets are 100 wickets, not 0. Pt 4. Debatable. Also if the 3+3 carried the Match perf pts as 6+0, there is no change in the calculations. Pt 5. Don't take just this example. Try to understand the general point. The theme is "dismissing very good batsmen at low scores", that is all. Anyhow thanks for some nice comments. Ananth: ]]

  • Vivek on July 21, 2009, 19:36 GMT

    Sorry I couldn't add this in my previous post, but is there any way you could get the statistics of the bowler's average, strike rate and percentage of wickets for batsman from #1-6, and batsman from #7-11. Interesting article though. [[ Yes I have that information. Will try and add another link. Ananth: ]]

  • Vivek on July 21, 2009, 19:31 GMT

    Interesting article!!! Nice work

  • Yogesh on July 21, 2009, 19:23 GMT

    Ananth, At first look at the complete list, i see this serious anomaly. Irfan Pathan at 40 ahead of Gillespie, Lee, Srinath, Zaheer, Flintoff. Now, i wonder even if ardent fans of Irfan would rank him above these ?? The effort and thought is appreciable but not statements like "Muralitharan is deservedly on top" which is a bit like saying that your rankings are "the rankings" to assess bowlers. As awesome as they might be, performances over 30-40 matches cannot be preferred over consitent performance over a career for 15 years all over the world. Thats why Mcgrath and Kumble are called Legends and Steyn has another few years to go before he can be even compared with them. While there is a world that seperates Akhtar from Akram, your points seem to indicate a difference of 0.01. [[ In any table there will be anomalies, that too in the beholders' eyes. Anyone can pick two bowlers and say "Hey what". I am not going to explain any such placements other than obvious misplacements. Again, just to find fault, you are saying "your rankings are "the rankings". I never said it You said that. Ananth: ]]

  • No featured comments at the moment.

  • Yogesh on July 21, 2009, 19:23 GMT

    Ananth, At first look at the complete list, i see this serious anomaly. Irfan Pathan at 40 ahead of Gillespie, Lee, Srinath, Zaheer, Flintoff. Now, i wonder even if ardent fans of Irfan would rank him above these ?? The effort and thought is appreciable but not statements like "Muralitharan is deservedly on top" which is a bit like saying that your rankings are "the rankings" to assess bowlers. As awesome as they might be, performances over 30-40 matches cannot be preferred over consitent performance over a career for 15 years all over the world. Thats why Mcgrath and Kumble are called Legends and Steyn has another few years to go before he can be even compared with them. While there is a world that seperates Akhtar from Akram, your points seem to indicate a difference of 0.01. [[ In any table there will be anomalies, that too in the beholders' eyes. Anyone can pick two bowlers and say "Hey what". I am not going to explain any such placements other than obvious misplacements. Again, just to find fault, you are saying "your rankings are "the rankings". I never said it You said that. Ananth: ]]

  • Vivek on July 21, 2009, 19:31 GMT

    Interesting article!!! Nice work

  • Vivek on July 21, 2009, 19:36 GMT

    Sorry I couldn't add this in my previous post, but is there any way you could get the statistics of the bowler's average, strike rate and percentage of wickets for batsman from #1-6, and batsman from #7-11. Interesting article though. [[ Yes I have that information. Will try and add another link. Ananth: ]]

  • love goel on July 21, 2009, 20:27 GMT

    great analysis ,with so many factors and in depth analysis.Some questions on the calculations 1. Why linear weight for career longevity. Certainly the number of bowlers with 400 wickets is less than half of bowlers with 200 wickets. 2.Also why 5 points for 1000 wickets when no bowler has done that? why not for the maximnum wickets at present(murlitharan) 3. Similarly when the cutoff is 100 wickets , then why not zero longevity points for 100 wickets and not .5 points(100/1000) 4.Same applies for match performance. Much easier to get 3&3 in both innings than 6 in one inning(whether 3&3 is as valuable as 6&0 is for debate) 5.Also the example of Panesar and Ricky Ponting.I think you need to calculate how many more runs Ponting score once he has scored 150 and compare that to what he scores when he is on 0. Batsmen like virendar sehwag are much much harder to dismiss when at 100 than at 0. [[ Point 2 is valid. Only problem is that the points, then will keep on changing with each wicket Murali takes. Pt 3. 100 wickets are 100 wickets, not 0. Pt 4. Debatable. Also if the 3+3 carried the Match perf pts as 6+0, there is no change in the calculations. Pt 5. Don't take just this example. Try to understand the general point. The theme is "dismissing very good batsmen at low scores", that is all. Anyhow thanks for some nice comments. Ananth: ]]

  • David on July 22, 2009, 1:25 GMT

    I would hardly call the the period from 2000-2009 part of the "balanced" era. Flatter pitches, bigger bats, restrictions on agressive bowling, pretty much any expert on the game would say that the balance has been in the favour of the bat. Add that to the fact that were very few attacks consisting of 4 better than average bowlers (usually one good bowler with a few average ones) meants less pressure on batsmen. Just look what happened when Australia ran into a four pronged pace attack in England in 05!

    Fortunately we are starting to see better bowling attacks (which is why SA and India have now beaten Australia) but I think you may want to reconsider the idea of a "balanced" period. [[ A better balanced era than the previous one, especially the period between 1940 and 1969. Certainly you cannot deny that the bat and ball have both have their continuous moments during this period. Ananth: ]]

  • Yogesh on July 22, 2009, 2:58 GMT

    But isn't Pathan above Lee, Gillespie, Srinath an obvious misplacement ?? And location based adjustment is something i always emphasise on incorporating but only if it can be in a just manner. Pace bowlers being given extra points in sub continent is quite a few times a false advantage like Ind vs Sa in Ahmedabad 2008 or Ind vs Pak Karachi 2005. Both the pitches were pace friendly. And again on a crumbling wicket, a pace bowler is as dangerous as a spinner. Rather what might be more indicative of the pitch is the avg. of the pace and spin bowlers in the match. But this could sometimes be erroneous when a particular bowler has an exceptional spell like Kumble in Sydney 2004.

    In bowling accuracy, shouldn't something be there for the bowler who outperforms not only his team-mates but even the opposition easily ? Again, the ex. being Kumble in Sydney 2004. [[ Yogesh Good idea to avoid unnecessary accusatory comments. I certainly respect your cricketing comments. I agree Pathan is somewhat high placed having just crossed the tape at 100 wkts. However he is ahead of most of the bowlers you have mentioned on the following points. - B/W Ratio - much higher figure, almost a spinner's number. - Wkt quality (lot less late order wkts). - Avge quality of batsmen dismissed (again a reflection of the top order batsmen dismissed). Maybe one must increase the lower limit especially for the current era. Re sub-continent wickets, I am waiting for comments on Kapil Dev. How he toiled on the sub-continent pitches and so on. Pl remember that I have gone on actual pace/spin distribution of wickets in sub-continent. Again it is easy to pick up one or two tests in either end of the world. I have gone on overall figures. Ananth: ]]

  • Michael on July 22, 2009, 3:19 GMT

    Hi Ananth, good work. No real issues with most of the lists - very fine bowlers who all deserve to be there. I wonder, though, if there is a case for relaxing the eligibility criteria for the Pre-WW1 era, based on the admitted paucity of Tests that may have contributed to less bowlers reaching 100 wickets? Example is Fred Spofforth, who claimed 94 wickets in 18 matches over 10 years while missing very few matches. Or Joey Palmer (78 wickets from 17 matches) and JJ Ferris (61 wickets, 9 matches), both of whom have records that suggest they were fine bowlers who dominated the opposition but who were only able to play a few Tests a year. Every analysis needs a cut-off but to have 10 bowlers meet the criteria for a whole 37-year period seems a bit restrictive. Perhaps a cut off of 50/60 wickets, rather than 100, would be more representative of bowlers from this era. [[ Michael Good point. I must admit I was fascinated by the exact number of 10 reached. Will look at lowering the qualification criteria and update the article itself instead of waiting for the follow-up article. Michael Have since added the revised tables with 60 wickets as the cut-off. Ananth: ]]

  • Qasim on July 22, 2009, 4:14 GMT

    Again an impressive table Anand well done. I have just one question which I couldn't clarify myself, have you accounted for the strength of a bowlers own attack?? For example Wasim Akram in his career had Imran Khan, Waqar Younis, Shoaib Akhtar (all in top 20) and other good bowlers like Saqlain Mushtaq with him fighting for wickets whereas someone like Muralitharan had only Vaas with him, who isn't that high either.... Obiuosly it would've been harder for Akram to take wickets as compared to Murali, perhaps this should also be added??? [[ Qasim If you read the article carefully, this factor has been clearly taken care of and let me confess that only after this measure was introduced did the great fast bowlers like Younis/Akram/Holding/Imran/Marshall move up. Ananth: ]]

  • Kartik on July 22, 2009, 4:20 GMT

    Ananth,

    Regarding the type of wicket captured :

    The quality of a fielding side makes a big difference in the success of a bowler. Warne had a much better fielding side than Kumble. If both had equal fielders, their career average and strike rate would probably be the same.

    The better fielding side pervades most of the parameters of the analysis (total wickets, match result, etc.)

    The only easy way to account for this is to put more weightage to wickets taken bowled or lbw, and less to stumped, caught behind, or caught in field. Hence, the fielder factor becomes smaller in the analysis. [[ Kartik That would be unfair to the spinners and to many medium pacers. McGrath and Hadlee specialized in finding the edge and that was a special skill. I considered and decided not to have the same. A wicket is a wicket. Inducing a mishit off a slower ball requires as much skill as finding the pads. Ananth: ]]

  • Big Gunz on July 22, 2009, 4:28 GMT

    An era from the 70s to 2000 and there is only ONE Windies bowler in the top 10? and only 4 in the top 20? Mate I am as Aussie as a transplanted WACA rat can be in London but no way could that period not have at least 3 windies in the top 10 and maybe 7 in the top 20 even if it meant bumping an Aussie or two! Or are they penalized for having to share the spoils among the other great bowlers on the team? In such a case then Walsh should be a shoo in given he carried the team for so long on his own! As my Jamaican friend would say " Mr. DJ rewind dat back" [[ It is true that they had to share the spoils amongst them as many Australian greats had to do. It has been partly taken care of by having the strength of own attack built in. I am afraid you have to live with that. Ananth: ]]