September 15, 2012

Test players: a look into their best decades

Performance analysis of Test batting and bowling careers of players by decade
243

Imran Khan has the best adjusted average in a decade © Getty Images

In a passing comment, Sarosh Nayyar requested that I do an analysis of the best decade for a player and compare the results across the players. It seemed a very intriguing request. I realized that this had the makings of a, hitherto not attempted, unique analysis, when I started looking at it deeply. The possibilities were immense. There was a scenario in which we could get a batsman-decade which was closer than the 40% gap between Bradman and Sutcliffe, which itself was a great incentive.

And ten years is a long enough period to draw conclusions with sufficient weight. We also agreed that the raw numbers had to be adjusted for the specific decade's peer numbers, during our exchange of views. The complex mechanism to determine the peer averages between any two Tests was already in place. It is indeed a tricky exercise and I have been working on-and-off for nearly a month now. The complexities are summarised below.

1. How does one handle the six-year voids which were caused by the two World Wars? The careers of some great players like Barnes and Verity were curtailed and the careers of top players such as Hobbs, Bradman, Hammond, Hutton, Compton et al got disrupted. I had to find a satisfactory methodology which handled this effectively.

2. This is predominantly a performance analysis: hence the decade has to be strictly defined. At the same time the careers of Pietersen (Twitter-gaffes notwithstanding), Cook, Gambhir et al, which might fall short of the qualifying ten-year periods, have to be recognized.

3. The peer adjustment has to be crisp and clear and has to make a telling impact. While the bowling comparisons have to be across the board, the batting peer comparisons have to be restricted to the Top-7 batsmen.

Keeping these in mind, I have worked on the following criteria and methodology.

1. The decade is strictly and exactly defined, match to match. There are very few exceptions to the 10-year requirement which are explained below.

2. The war-years, as far this analysis is concerned, are just wiped off. How I wish I could move the clock back and do it in real life. Millions of lives could have been saved!!! Enough of this wishful thinking. Tests 134 and 135 are, in real life, around six-plus years apart. As do the Tests 274 and 275. But this period is reduced to nothing as far as this analysis is concerned. In other words, the period 1908-1922 becomes eight years or thereabout. Similarly 1938-1954 is equivalent to ten years.

3. S.F.Barnes played his last Test during 1914 and did not play thereafter. And his best years were between 1907 and 1914. Hence the period of seven years has been accepted for him. Verity played his last Test in 1939 and then died on the battle-fields. His entire career lasted eight years from 1931 to 1939. Hence this has been considered valid. Note the subtle differences between the careers of these two great bowlers.

4. It was indeed fortunate that Bradman's best decade was between 1929 and 1939. So no problems. However the best years for Hobbs, Hammond and Hutton, spanned the wars and their "decades" are allowed to go past 10 years.

5. For the best performance analysis, no concession is given to anyone else, including the recent players. However for the total runs/wickets analysis it really does not matter if the career has lasted below 10 years.

6. The Top-7 Batting average for the specific range of Tests is considered for peer comparisons. This is compared to the all-time Top-7 Batting average of 35.92 and then adjusted. Couple of examples will explain this.

The average for Kallis is a very high 66.13, between 2001 and 2011, when the T7-Avge for all players, other than Kallis, was 39.03. So his average is adjusted to 60.86 (66.13 x 35.92 / 39.03). On the other side, Harvey averaged 54.40 between 1948 and 1958 when the T7-Avge for all players, except Harvey, was 33.80. So his average is adjusted to 57.82 (54.40 x 35.92 / 33.80).

7. For peer comparisons the all-inclusive Bowling average for the specific range of Tests is taken. This is compared to the all-time all-inclusive Bowling average which is 32.0. This is necessary because the wickets are captured across all batting positions.

Harbhajan's best decade average was a somewhat high 29.92 between 1999 and 2009 when the Bowling Avge for all bowlers, sans Harbhajan, was at a higher level of 34.08. So his average is adjusted to 28.09 (29.92 x 32.0 / 34.08). On the other hand, Laker averaged 18.56 at his peak between 1949 and 1959, a bowler-dominated decade, during which the Bowling Avge for all bowlers, other than Laker, was a low 29.4. So his average is adjusted to 20.20 (18.56 x 32.0 / 29.4)

Batsman best decade analysis: Based on adjusted average

Batsman Adjust <<<< Best Decade >>>> # of Inns NO Runs Avge Peer Values
Name Avge StMat StYear EndMat EnYear Yrs T7-Avg T7-Oth
 
Bradman D.G 103.93 180 1929 271 1939 10 51 5 4785 104.02 37.37 35.95
Sobers 73.75 450 1958 631 1968 10 78 12 4821 73.04 35.97 35.57
Hobbs J.B 70.65 117 1912 167 1926 14 55 4 3326 65.21 34.47 33.15
Walcott C.L 66.43 304 1948 458 1958 10 64 6 3581 61.74 33.75 33.38
EdeC Weekes 64.97 297 1948 451 1958 10 73 4 4222 61.18 34.30 33.83
Tendulkar 63.19 1201 1992 1628 2002 10 138 14 7744 62.45 35.75 35.50
Barrington 61.81 409 1955 592 1965 10 90 11 4710 59.62 34.94 34.65
Sutcliffe H 61.60 153 1924 233 1934 10 76 8 4235 62.27 37.01 36.31
Hutton L 61.25 263 1938 387 1954 16 119 15 6538 62.86 37.46 36.86
Kallis J.H 60.86 1562 2001 2006 2011 10 165 29 8995 66.13 39.18 39.03
Ponting R.T 60.13 1426 1998 1887 2008 10 172 25 9192 62.53 37.53 37.35
Richards 59.39 768 1976 1034 1986 10 97 5 5372 58.39 35.62 35.31
Hammond W.R 58.77 244 1935 284 1947 12 54 7 3025 64.36 40.07 39.34
Waugh S.R 58.13 1236 1993 1669 2003 10 163 30 7697 57.87 35.84 35.76
Dravid R 58.08 1344 1996 1820 2006 10 168 22 8710 59.65 37.06 36.89
Hayden M.L 57.96 1252 1994 1684 2004 10 85 8 4489 58.29 36.24 36.12
Harvey R.N 57.82 295 1948 448 1958 10 85 6 4298 54.40 34.16 33.80
Boycott G 57.78 627 1968 815 1978 10 74 10 3737 58.39 36.46 36.30
Simpson R.B 57.69 553 1964 729 1974 10 50 3 2768 58.89 36.88 36.67
Hill C 57.39 53 1897 96 1907 10 57 2 2411 43.83 28.21 27.43
Chappell 56.41 679 1971 895 1981 10 113 14 5571 56.27 36.04 35.83
Compton 55.71 282 1947 436 1957 10 105 10 5029 52.93 34.45 34.13
Border A.R 55.48 944 1983 1209 1993 10 147 27 6733 56.10 36.40 36.32
Mohd Yousuf 55.43 1502 2000 1962 2010 10 111 10 6037 59.77 38.87 38.74
J Miandad 55.42 937 1982 1196 1992 10 101 8 5205 55.96 36.50 36.27
Hendren E.H 55.22 153 1924 233 1934 10 57 8 2771 56.55 37.01 36.78
Sangakkara 54.88 1537 2001 1991 2011 10 142 12 7756 59.66 39.22 39.05
Worrell 54.77 296 1948 449 1958 10 57 5 2691 51.75 34.11 33.94
Gavaskar 54.10 683 1971 898 1981 10 123 8 6180 53.73 35.99 35.68
Crowe M.D 54.01 1003 1984 1278 1994 10 97 10 4756 54.66 36.50 36.35

Bradman's best decade was between 1929 and 1939, as expected. The adjusted batting average is 103.93 (unadjusted avge 104.02), which is around 4% above his career batting average. This indicates that virtually any decade of Bradman is likely to produce a number around 100. Sobers, whose adjusted average is a magnificent 73.75 (73.04), between 1958 and 1968 is in second place. It is the first time a player, over a long period of 10 years, comes within 30% of Bradman. Hobbs, with a best decade spread across the WW1 (1912 to 1926) is next with 70.65 (65.21). Walcott and Weekes follow, with adjusted averages either side of 65.

Tendulkar, with a best decade of 1992 to 2002, appears next in the list with 63.19 (62.45), followed by the old-timers, Barrington, Sutcliffe and Hutton, comfortably in excess of 60. Jacques Kallis rounds off the top-10 with an adjusted average of 60.86, between 2001 and 2011, seriously adjusted downwards from 66.13. He is in fact third in the unadjusted table. This confirms the significant impact of adjusting against peer values.

Bowler best decade analysis: Based on adjusted average

Batsman Adjust <<<< Best Decade >>>> # of Wkts Runs Avge Peer Values
Name Avge StMat StYear EndMat EnYear Yrs Avg Oth
 
Imran Khan 18.26 861 1979 1127 1989 10 250 4593 18.37 31.74 32.20
Barnes S.F 18.57 96 1907 133 1914 7 163 2684 16.46 26.73 28.36
Muralitharan 18.60 1376 1997 1842 2007 10 575 11231 19.53 33.02 33.60
Hadlee R.J 19.07 848 1979 1115 1989 10 293 5618 19.17 31.66 32.16
Marshall 19.09 884 1980 1148 1990 10 316 6194 19.60 32.29 32.86
McGrath G.D 19.53 1313 1995 1771 2005 10 484 9792 20.23 32.69 33.15
Ambrose 20.04 1142 1990 1490 2000 10 316 6273 19.85 31.33 31.69
O'Reilly 20.04 233 1934 275 1946 12 110 2356 21.41 33.01 34.19
Laker J.C 20.20 317 1949 468 1959 10 166 3081 18.56 28.99 29.40
Garner J 20.66 803 1977 1072 1987 10 238 4874 20.47 31.35 31.70
Wasim Akram 20.90 1086 1987 1391 1997 10 271 5747 21.20 32.12 32.47
Pollock S.M 21.20 1318 1995 1776 2005 10 371 8122 21.89 32.74 33.04
Waqar Younis 21.32 1137 1990 1483 2000 10 270 5724 21.20 31.54 31.82
Davidson 21.83 375 1953 539 1963 10 181 3692 20.39 29.54 29.90
Holding M.A 21.98 775 1976 1044 1986 10 230 4955 21.54 31.07 31.36
Lindwall 22.02 275 1946 424 1956 10 192 4197 21.85 31.31 31.75
Donald A.A 22.14 1206 1992 1590 2002 10 318 7010 22.04 31.58 31.86
Trueman F.S 22.32 386 1954 560 1964 10 247 5220 21.13 29.85 30.30
Lock G.A.R 22.70 412 1955 595 1965 10 139 3067 22.06 30.87 31.10
Miller K.R 23.29 275 1946 424 1956 10 147 3381 23.00 31.31 31.60
Walsh C.A 23.63 1016 1985 1297 1995 10 262 6316 24.10 32.35 32.63
Lillee D.K 23.66 698 1972 927 1982 10 320 7369 23.02 30.76 31.14
Underwood 23.77 622 1967 806 1977 10 263 6310 23.99 31.90 32.29
Gibbs L.R 23.86 448 1958 628 1968 10 154 3692 23.97 31.91 32.15
ShoaibAkhtar 24.10 1389 1997 1849 2007 10 175 4377 25.01 33.10 33.20
Verity H 24.12 210 1931 272 1939 8 144 3510 24.37 31.69 32.33
Willis 24.43 739 1974 988 1984 10 296 7052 23.82 30.90 31.20
Grimmett 24.98 167 1926 251 1936 10 197 4917 24.95 31.39 31.96
Warne S.K 24.99 1215 1993 1637 2003 10 477 11925 25.00 31.74 32.01
Statham J.B 25.19 332 1951 506 1961 10 196 4514 23.03 29.02 29.26

For obvious reasons I have ignored Lohmann who had a best decade with 100+ wickets at 10+. These were days about which we really cannot come to a decision and are not comparable with later periods. Hence the cut-off is 1900.

There is a real, real surprise at the top. This should make people sit up and take notice. Imran Khan, between 1979 and 1989 had an adjusted bowling average of 18.37 (18.26). He gained almost nothing through peer adjustment. Who could have imagined that Imran Khan would have leapt over many other fancied bowlers? He captured 250 wickets during this period and improved his bowling average from 31.72 to 22.21. That single sentence is sufficient.

Barnes comes in second with an average of 18.57 (16.46). There was a significant adjustment for the period. Readers should note that the period for Barnes has been accepted at seven years owing to the WW1 intervention. Muralitharan, between 1997 and 2007, captured 575 wickets at 18.60 (19.53). Hadlee, who captured 293 wickets at 19.07 (19.17) between 1979 and 1989 follows next. Marshall completes the top-5 with 316 wickets at 19.09.

The next five bowlers represent the very best bowling talent that ever played the game - McGrath, Ambrose, O'Reilly, Laker and Garner. These 10 are probably on anyone's list of the top dozen bowlers. Laker's average has had a major downward adjustment. Wasim and Waqar follow in 11th and 13th positions.

Batsmen best decade analysis: Top Run-scorers in a decade

Batsman Career Yrs NO <<<< Best Decade >>>> # of Runs Runs/Yr
Name St End Yrs StMat StYear EndMat EnYear Yrs
 
Ponting R.T 1995 2012 17 1523 2000 1983 2010 10 10058 1005.8
Kallis J.H 1995 2012 17 1528 2001 1988 2011 10 9318 931.8
Dravid R 1996 2012 16 1589 2002 2031 2012 10 9031 903.1
Lara B.C 1990 2006 16 1340 1996 1816 2006 10 8707 870.7
Sangakkara 2000 2012 12 1567 2001 2016 2011 10 8384 838.4
Hayden M.L 1994 2009 15 1493 2000 1904 2009 9 8365 929.4
Tendulkar 1989 2012 23 1531 2001 1991 2011 10 8276 827.6
Jayawardene 1997 2012 15 1473 1999 1936 2009 10 8248 824.8
Smith G.C 2002 2012 10 1619 2002 2053 2012 10 8201 820.1
Sehwag V 2001 2012 11 1591 2002 2034 2012 10 7943 794.3
Waugh S.R 1985 2004 19 1208 1993 1637 2003 10 7873 787.3
Border A.R 1979 1994 15 855 1979 1126 1989 10 7851 785.1
Gavaskar 1971 1987 16 769 1976 1034 1986 10 7725 772.5
Taylor M.A 1989 1999 10 1113 1989 1439 1999 10 7525 752.5
Langer J.L 1993 2007 14 1352 1997 1826 2007 10 7443 744.3
Waugh M.E 1991 2002 11 1202 1992 1620 2002 10 7331 733.1
Laxman 1996 2012 16 1589 2002 2031 2012 10 7078 707.8
Pietersen 2005 2012 7 1756 2005 2051 2012 7 7076 1010.9
Strauss A.J 2004 2012 8 1700 2004 2053 2012 8 7037 879.6
Kirsten G 1993 2004 11 1243 1994 1681 2004 10 7013 701.3
Boon D.C 1984 1996 12 1017 1985 1298 1995 10 6979 697.9
Chanderpaul 1994 2012 18 1601 2002 2042 2012 10 6913 691.3
Gower D.I 1978 1992 14 828 1978 1101 1988 10 6847 684.7
Stewart A.J 1990 2003 13 1176 1991 1557 2001 10 6798 679.8
Mohd Yousuf 1998 2010 12 1403 1998 1864 2008 10 6770 677.0
Barrington 1955 1968 13 474 1959 640 1968 9 6754 750.4
Atherton 1989 2001 12 1191 1992 1558 2001 9 6562 729.1
Cook A.N 2006 2012 6 1785 2006 2053 2012 6 6555 1092.5
Hutton L 1937 1955 18 263 1938 387 1954 16 6538 408.6
Inzamam 1992 2007 15 1366 1997 1830 2007 10 6431 643.1

This is a simple run-based analysis. As such there are no restrictions and the current players also qualify. After all if Cook accumulates over 6000 runs in 6 years and counting, why should he not get ahead of Gooch who accumulated nearly as many runs in 10 years?

Ponting is the only batsman in history to accumulate over 10000 runs in 10 years. This is a stand-out achievement and has to be recognized. He did this between 2000 and 2010. Kallis accumulated 9318 runs between 2001 and 2011. Dravid, Lara and Sangakkara amassed either side of 9000 runs in 10 years.

Hayden accumulated 8365 runs, but in 9 years. This was after his break and return. Hence 9 years is fine. Then Tendulkar and Jayawardene follow.

Cook leads in the matter of Runs per year, having accumulated 6555 runs in 6 years at an average of 1092. Pietersen is the only other batsman to average more than 1000 runs per year along with Ponting.

Readers should realize the difficulty in maintaining a 1000 runs per year average over 10 years. Mohammad Yousuf had the greatest of years during 2006 when he scored 1788 runs but could accumulate only 5000 runs during the other 9 years of his best decade. In three years, 2003, 2005 and 2006, Ponting accumulated nearly 4500 runs, but had to bat outstandingly well to compile 5500 runs in 7 years.

Bowler best decade analysis: Top Wicket-takers in a decade

Batsman Career Yrs # of <<<< Best Decade >>>> # of Wkts Wkts/Yr
Name St End Yrs StMat StYear EndMat EnYear Yrs
 
Muralitharan 1992 2010 18 1394 1998 1859 2008 10 588 58.8
Warne S.K 1992 2007 15 1346 1996 1825 2006 10 488 48.8
McGrath G.D 1993 2007 14 1313 1995 1771 2005 10 484 48.4
Kumble A 1990 2008 18 1405 1998 1866 2008 10 430 43.0
Pollock S.M 1995 2008 13 1356 1997 1830 2007 10 390 39.0
Ntini M 1998 2009 11 1504 2000 1944 2009 9 380 42.2
Harbhajan 1998 2011 13 1531 2001 1991 2011 10 372 37.2
Walsh C.A 1984 2001 17 1169 1991 1544 2001 10 372 37.2
Botham I.T 1977 1992 15 806 1977 1078 1987 10 370 37.0
Marshall 1978 1991 13 949 1983 1175 1991 8 342 42.8
Ambrose 1988 2000 12 1095 1988 1414 1998 10 337 33.7
Donald A.A 1992 2002 10 1188 1992 1590 2002 10 330 33.0
Kapil Dev N 1978 1994 16 851 1979 1123 1989 10 323 32.3
Lillee D.K 1971 1984 13 698 1972 927 1982 10 320 32.0
Hadlee R.J 1973 1990 17 817 1978 1090 1988 10 312 31.2
Lee B 1999 2008 9 1479 1999 1902 2008 9 310 34.4
Wasim Akram 1985 2002 17 1127 1989 1468 1999 10 304 30.4
Waqar Younis 1989 2003 14 1151 1990 1511 2000 10 298 29.8
Willis 1971 1984 13 732 1974 977 1984 10 298 29.8
Steyn D.W 2004 2012 8 1728 2004 2053 2012 8 287 35.9
Anderson 2003 2012 9 1646 2003 2053 2012 9 276 30.7
Imran Khan 1971 1992 21 831 1978 1106 1988 10 272 27.2
Vaas WPUJC 1994 2009 15 1445 1999 1909 2009 10 271 27.1
Zaheer Khan 2000 2012 12 1535 2001 1991 2011 10 266 26.6
Trueman F.S 1952 1965 13 409 1955 592 1965 10 265 26.5
Underwood 1966 1982 16 622 1967 806 1977 10 263 26.3
D Kaneria 2000 2010 10 1518 2000 1967 2010 10 261 26.1
McDermott 1984 1996 12 1017 1985 1298 1995 10 260 26.0
Gillespie 1996 2006 10 1343 1996 1799 2006 10 259 25.9
Garner J 1977 1987 10 801 1977 1072 1987 10 249 24.9

Muralitharan had a golden decade from 1998 and 2008 during which he accumulated 588 wickets at an average of over 58 per year. He is followed by the wonderful bowling pair of Warne and McGrath who accumulated 480 odd wickets between around 1995 and 2005. Kumble is next with 430 wickets in the decade beginning 1998. Pollock, Ntini and Harbhajan appear next in this list. Botham is the only non-modern bowler, indicating the recent profusion of Test matches.

Let us think for a second. In a period of 10 years, Muralitharan captured nearly 600 wickets in 10 years. Only 42 bowlers in 135 years have captured more than Muralitharan's average of 58.8 wickets per year. The highest tally is by Warne, during 2005, with 96 wickets, followed by Muralitharan, with 90 wickets in 2006.

The top batsmen in a decade

Graph of adjusted averages for batsmen (best decade)
© Anantha Narayanan

The graphs are self-explanatory. The idea is not to show the relative position of the averages but rather the positioning across the years. The 1950s-60s seem to be the golden years of batting since there are four batsmen present in the top-10 with high averages.

The top bowlers in a decade

Graph of adjusted averages for bowlers (best decade)
© Anantha Narayanan

It must be remembered that the Y-axis is not really to exact scale in view of the close bunching of players with similar values. The bowling decades are the 1980s-90s.

To download/view the comprehensive Excel sheet containing all the tables related to Test Decades analysis, please CLICK HERE.

Thanks, Sarosh, for providing a good spark.

The stars of this special analysis are Sobers and Hobbs, whose batting average exceeded 70 during their best decades and Imran Khan and Muralitharan who had bowling averages around 18 during their respective best decades.

When I heard that Lara was going to be inducted into the ICC Hall-of-Fame, I almost shelved this article, post-editing, to be replaced by a tribute to Lara. Then I decided not to rush the same. Since ICC have taken their own time, let me take a few more days to come out with a well-rounded tribute to the batting genius, in the process fine-tuning the single-player analysis program. I will close this with Milind's telling comment.

Sometimes a talent requires the recognition of being an awardee. Then there are cases when the award needs the prestige by associating itself with a talent that transcends awards.

Well said, Milind.

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

  • Vikram on November 16, 2012, 13:41 GMT

    Very interesting article Ananth. Here's an attempt at an alternate team. There are some personal preferences (like yours) especially when there was a close call. After all, as the manager, I should be comfortable with the players in the team, right? Same rules as yours, players selected for being specialists, split of 7-1-7. However, I have chosen 3 openers with one of them (Sehwag) having the ability to cover the middle order if needed. Openers: Hayden, Gooch, Sehwag Other batsmen: Dravid, Harvey, Headley, Crowe Wicketkeeper: Sangakkara Fast bowlers: Imran, Waqar, Donald, Barnes, Pollock Spinners: Saqlain, Chandrashekhar [[ Why do you people insist on doing the very thing I do not want you to do. That too, sensible readers like you. Your criteria is "excluding my XV". To me that seems quite silly. How do I take this. Is it really the best world XV. Pl remember we are not playing a match. I am trying to get the readers' responses into an Excel sheet and work out a wonderful Readers' XV. Are you helping me in doing that. Ananth: ]]

  • Gerry_the_Merry on September 30, 2012, 17:16 GMT

    Between 1995 and end of career, Murali and Warne, including only away matches in England+WI+South Africa+NZ+Pak (but nor Z, BD), and home matches, or matches against India, where batsmen have been extraordinary against spin, Murali and Warne have following stats - Matches, Wkts, Avg, Strike Rate

    Murali - 33/200/22.5/54 Warne - 38/190/24/50

    Murali definitely has delivered the numbers. I only have a problem with his action. [[ Yes, I guessed so. Once the "action" comes in we should stop further discussions. At least in this blogspace. Ananth: ]]

  • Waspsting on September 30, 2012, 12:51 GMT

    "May I know where this "outclassing" took place." (Murali & Warne)

    there was a very amusing comment in an interview with Stuart MacGill along the same lines of "Warne completely outbowling Murali on a tour of SL" (paraphrased).

    He also added that to him, you measure how players play relative to each other in the same conditions.

    My reading of this comment supports my theory of players giving opinion that indirectly reflects well on themselves. (MacGill has a better record than Warne in matches they played together - and I bet he knows it!)

    ----

    On an unrelated note going back to a debate we had on this blog about the relative merits of economy rate vs strike rate in ODI cricket - note John Buchanan's position (according to Macgill) - a guy who presumably had the biggest say in how those all conquering sides approached the ODI game

    http://www.espncricinfo.com/magazine/content/story/565775.html

  • Waspsting on September 30, 2012, 12:30 GMT

    was also full of praise. Pace and spin seem to have come alike to him (assuming "hype" bias' would work against rather for Headley, the high regard he's held in by Eng and Aus players is especially telling)

    - noted bad wicket player, by any standard

    Compare him to the legendary Wally Hammond. they played 3 series' together -

    33 in Eng, Hammond 24.66, Headley 55.40 34/35 in WI, Hammond 25, Headley 97 39 in Eng, Hammond 55.8, Headley 66.8

    Didn't play much Headley, but his career record, at every level, is near flawless. I rate him VERY HIGH up there, amongst the top 10 batsman certainly.

  • Waspsting on September 30, 2012, 12:19 GMT

    "Coming back to the team on his day Lara was a better match-winner than Sachin"

    @Harsh - what is the specific thinking behind this?

    re: Headley, I agree with you, for a # of reasons, despite as Ananth points out, the "too few games" point against him.

    - he played over a long period of time, which is categorically different than averaging 60 for 3 consecutive years ("hot streak"). That he didn't play much... it is what it is.

    -successful in all conditions. Eng, Aus and WI

    - the 10 100s to 5 50s in tests. The ability to convert 50s to 100s, i think is the hallmark of a really classy player (in 1st class, the record is a very a good 44 - 100 too)

    - ave 70 for 10,000 runs at 1st class. very impressive

    - impressed all about him (i assume for WI player in that time, that'd be particularly difficult and if anything, subjective bias would work AGAINST him - much like it does for Pak players even now)

    - Hutton called him "one of the best players of fast bowling I saw", while Grimmett...

  • Waspsting on September 30, 2012, 11:55 GMT

    re: Simpson - I was thinking of as a strong contender for a place as an opener for Aus 11 - not a place in the middle order competing against guys like Bradman, Chappell, Ponting etc.

    I'd go with Ponsford and Hayden - but Simpson's worthy of a more thought than he usually gets.

    As for his bowling, i was thinking of him as a useful bowler, not an all-rounder. As a bonus bowler, he's VERY USEFUL.

    He dismissed Gary Sobers 4 times in test cricket as well as Barrington, Cowdrey and Graeme Pollock twice.

    not bad for a guy in the team for his batting.

    Not bad for a batsman!

  • Nitin Gautam on September 30, 2012, 9:53 GMT

    Lastly like all great fast bowlers Mcgraw, donald, akram, & to some extent steyn too had exceptional record everywhere & it didnt matter for them if they are bowling at dustbowls of India or low slow pitches of SL or fast hard boucny Australian tracks of seaming eng & NZ.they excelled everywhere similarly I dont have any doubt had Murali been an Australian he would have been more accepted with equally gud record..

  • Nitin Gautam on September 30, 2012, 9:48 GMT

    @Shrikanth While I agree that a 10wk haul @ galle against BD is not same as 10 for at perts/m'bourne etc however Murli has pretty much proved himself everywhere vs everyone..Just for a comparison leaving out home country (aus for warne & SL for Murli) & waterloo countries for both (India for warne & Aus for Murli) although murli in Aus is much worse that warne in India, here are the stats. Warne-67tests, 355 wickets @23.93 Murli-55 tests, 255 wickets@22.84 Thing to notice here is warne lost 313@27 while murli lost whopping 493@19..low count of murli's wickets elsewhere is due to less tests & considering his more than 5wk/test he would hv matched warne's tally in equal no.of tests. If we remove WI for warne & India for murli (the next worst place for both respectively) their record says Warne-338 wickets@19.5 & Murli 255@20 very slight diff & really nothing to chose from..both are such an exceptional bowlers that leaving out any1 will raise few eyebrows.though personally I prefer warne

  • shrikanthk on September 30, 2012, 4:44 GMT

    also ppl use BD, ZIm to downgrade the wickets captured by him to glorify warne & I dont patronize this too..every test wicket & every test run & every test win is as important as any other..

    Nitin: I don't like the slightly dogmatic tone in this comment. I won't downgrade Murali for any reason. But surely a 20 wicket haul against Bdesh on pitches like Galle isn't quite the same as a 20 wicket haul over 3 tests held at Perth, Melbourne and Sydney.

    Let's face it. Had Warne been a Sri Lankan, I am sure he'd have an even better record. His success over a very long period on unfavourable wickets for spin bowling is without a parallel in cricket history. Australia has always been a graveyard for spinners. To average 25-26 in that continent is truly remarkable. [[ And if Murali had been born in Australia he would have found better acceptance, better support on the field and possibly 100 wickets less. And why should Australia be portrayed as a graveyard for spinners. These are not dead batting wickets. These are wickets with good bounce and carry and the quality spinners should benefit. And blanket statements do not work. Indian pitches aid spinners. Correct. Kumble 350 @ 25. Harbhajan 258 @ 28. Why then Warne 34 @ 43 and Murali 40 @ 45. The same bowlers in England. Kumble 36 @ 41. Harbhajan 14 @ 50. Warne 129 @ 22 and Murali 48 @ 19. So of the four, Warne and Murali have both done very well in England. Finally, at home, Warne 313 @ 27 and Murali 493 @ 20. So the home advantage is very very significant for all bowlers. All three have done poorly in Australia. I agree that Murali's weak spots have been Australia and India. Warne had two weak spots, in India and West Indies (17 @ 40). So nothing much to choose from. Ananth: ]]

  • shrikanthk on September 30, 2012, 3:42 GMT

    i agree with you about Hayden. i didn't see his early games against WI, but have heard he was quite poor. in his pomp, i assessed from his style that fast bowling might sort him out, and wasn't surprised by his ordinary showing against Eng 2005

    Great fast bowling sorts out most batsmen. Hayden no exception. The great thing about Hayden was that on his day on a good wicket he could place you in a position from which you can force a win on Day 1 of a test match! And he did that on a fairly consistent basis. You can't say the same thing about steady, admittedly more solid openers like Simpson or Langer or Taylor.

  • Vikram on November 16, 2012, 13:41 GMT

    Very interesting article Ananth. Here's an attempt at an alternate team. There are some personal preferences (like yours) especially when there was a close call. After all, as the manager, I should be comfortable with the players in the team, right? Same rules as yours, players selected for being specialists, split of 7-1-7. However, I have chosen 3 openers with one of them (Sehwag) having the ability to cover the middle order if needed. Openers: Hayden, Gooch, Sehwag Other batsmen: Dravid, Harvey, Headley, Crowe Wicketkeeper: Sangakkara Fast bowlers: Imran, Waqar, Donald, Barnes, Pollock Spinners: Saqlain, Chandrashekhar [[ Why do you people insist on doing the very thing I do not want you to do. That too, sensible readers like you. Your criteria is "excluding my XV". To me that seems quite silly. How do I take this. Is it really the best world XV. Pl remember we are not playing a match. I am trying to get the readers' responses into an Excel sheet and work out a wonderful Readers' XV. Are you helping me in doing that. Ananth: ]]

  • Gerry_the_Merry on September 30, 2012, 17:16 GMT

    Between 1995 and end of career, Murali and Warne, including only away matches in England+WI+South Africa+NZ+Pak (but nor Z, BD), and home matches, or matches against India, where batsmen have been extraordinary against spin, Murali and Warne have following stats - Matches, Wkts, Avg, Strike Rate

    Murali - 33/200/22.5/54 Warne - 38/190/24/50

    Murali definitely has delivered the numbers. I only have a problem with his action. [[ Yes, I guessed so. Once the "action" comes in we should stop further discussions. At least in this blogspace. Ananth: ]]

  • Waspsting on September 30, 2012, 12:51 GMT

    "May I know where this "outclassing" took place." (Murali & Warne)

    there was a very amusing comment in an interview with Stuart MacGill along the same lines of "Warne completely outbowling Murali on a tour of SL" (paraphrased).

    He also added that to him, you measure how players play relative to each other in the same conditions.

    My reading of this comment supports my theory of players giving opinion that indirectly reflects well on themselves. (MacGill has a better record than Warne in matches they played together - and I bet he knows it!)

    ----

    On an unrelated note going back to a debate we had on this blog about the relative merits of economy rate vs strike rate in ODI cricket - note John Buchanan's position (according to Macgill) - a guy who presumably had the biggest say in how those all conquering sides approached the ODI game

    http://www.espncricinfo.com/magazine/content/story/565775.html

  • Waspsting on September 30, 2012, 12:30 GMT

    was also full of praise. Pace and spin seem to have come alike to him (assuming "hype" bias' would work against rather for Headley, the high regard he's held in by Eng and Aus players is especially telling)

    - noted bad wicket player, by any standard

    Compare him to the legendary Wally Hammond. they played 3 series' together -

    33 in Eng, Hammond 24.66, Headley 55.40 34/35 in WI, Hammond 25, Headley 97 39 in Eng, Hammond 55.8, Headley 66.8

    Didn't play much Headley, but his career record, at every level, is near flawless. I rate him VERY HIGH up there, amongst the top 10 batsman certainly.

  • Waspsting on September 30, 2012, 12:19 GMT

    "Coming back to the team on his day Lara was a better match-winner than Sachin"

    @Harsh - what is the specific thinking behind this?

    re: Headley, I agree with you, for a # of reasons, despite as Ananth points out, the "too few games" point against him.

    - he played over a long period of time, which is categorically different than averaging 60 for 3 consecutive years ("hot streak"). That he didn't play much... it is what it is.

    -successful in all conditions. Eng, Aus and WI

    - the 10 100s to 5 50s in tests. The ability to convert 50s to 100s, i think is the hallmark of a really classy player (in 1st class, the record is a very a good 44 - 100 too)

    - ave 70 for 10,000 runs at 1st class. very impressive

    - impressed all about him (i assume for WI player in that time, that'd be particularly difficult and if anything, subjective bias would work AGAINST him - much like it does for Pak players even now)

    - Hutton called him "one of the best players of fast bowling I saw", while Grimmett...

  • Waspsting on September 30, 2012, 11:55 GMT

    re: Simpson - I was thinking of as a strong contender for a place as an opener for Aus 11 - not a place in the middle order competing against guys like Bradman, Chappell, Ponting etc.

    I'd go with Ponsford and Hayden - but Simpson's worthy of a more thought than he usually gets.

    As for his bowling, i was thinking of him as a useful bowler, not an all-rounder. As a bonus bowler, he's VERY USEFUL.

    He dismissed Gary Sobers 4 times in test cricket as well as Barrington, Cowdrey and Graeme Pollock twice.

    not bad for a guy in the team for his batting.

    Not bad for a batsman!

  • Nitin Gautam on September 30, 2012, 9:53 GMT

    Lastly like all great fast bowlers Mcgraw, donald, akram, & to some extent steyn too had exceptional record everywhere & it didnt matter for them if they are bowling at dustbowls of India or low slow pitches of SL or fast hard boucny Australian tracks of seaming eng & NZ.they excelled everywhere similarly I dont have any doubt had Murali been an Australian he would have been more accepted with equally gud record..

  • Nitin Gautam on September 30, 2012, 9:48 GMT

    @Shrikanth While I agree that a 10wk haul @ galle against BD is not same as 10 for at perts/m'bourne etc however Murli has pretty much proved himself everywhere vs everyone..Just for a comparison leaving out home country (aus for warne & SL for Murli) & waterloo countries for both (India for warne & Aus for Murli) although murli in Aus is much worse that warne in India, here are the stats. Warne-67tests, 355 wickets @23.93 Murli-55 tests, 255 wickets@22.84 Thing to notice here is warne lost 313@27 while murli lost whopping 493@19..low count of murli's wickets elsewhere is due to less tests & considering his more than 5wk/test he would hv matched warne's tally in equal no.of tests. If we remove WI for warne & India for murli (the next worst place for both respectively) their record says Warne-338 wickets@19.5 & Murli 255@20 very slight diff & really nothing to chose from..both are such an exceptional bowlers that leaving out any1 will raise few eyebrows.though personally I prefer warne

  • shrikanthk on September 30, 2012, 4:44 GMT

    also ppl use BD, ZIm to downgrade the wickets captured by him to glorify warne & I dont patronize this too..every test wicket & every test run & every test win is as important as any other..

    Nitin: I don't like the slightly dogmatic tone in this comment. I won't downgrade Murali for any reason. But surely a 20 wicket haul against Bdesh on pitches like Galle isn't quite the same as a 20 wicket haul over 3 tests held at Perth, Melbourne and Sydney.

    Let's face it. Had Warne been a Sri Lankan, I am sure he'd have an even better record. His success over a very long period on unfavourable wickets for spin bowling is without a parallel in cricket history. Australia has always been a graveyard for spinners. To average 25-26 in that continent is truly remarkable. [[ And if Murali had been born in Australia he would have found better acceptance, better support on the field and possibly 100 wickets less. And why should Australia be portrayed as a graveyard for spinners. These are not dead batting wickets. These are wickets with good bounce and carry and the quality spinners should benefit. And blanket statements do not work. Indian pitches aid spinners. Correct. Kumble 350 @ 25. Harbhajan 258 @ 28. Why then Warne 34 @ 43 and Murali 40 @ 45. The same bowlers in England. Kumble 36 @ 41. Harbhajan 14 @ 50. Warne 129 @ 22 and Murali 48 @ 19. So of the four, Warne and Murali have both done very well in England. Finally, at home, Warne 313 @ 27 and Murali 493 @ 20. So the home advantage is very very significant for all bowlers. All three have done poorly in Australia. I agree that Murali's weak spots have been Australia and India. Warne had two weak spots, in India and West Indies (17 @ 40). So nothing much to choose from. Ananth: ]]

  • shrikanthk on September 30, 2012, 3:42 GMT

    i agree with you about Hayden. i didn't see his early games against WI, but have heard he was quite poor. in his pomp, i assessed from his style that fast bowling might sort him out, and wasn't surprised by his ordinary showing against Eng 2005

    Great fast bowling sorts out most batsmen. Hayden no exception. The great thing about Hayden was that on his day on a good wicket he could place you in a position from which you can force a win on Day 1 of a test match! And he did that on a fairly consistent basis. You can't say the same thing about steady, admittedly more solid openers like Simpson or Langer or Taylor.

  • shrikanthk on September 30, 2012, 3:30 GMT

    Simpson is a surprise... i think he's somewhat underrated overall - a good contender for an all time Aus 11 spot

    Nice thought. To my mind, he is the preeminent Australian cricketer of the 60s. A giant figure. A very solid batsman with a huge appetite for big scores. However I am not sure how good a bowler he was. A test average fo 40+ and a FC average of 38+. He's not even in Wally Hammond's league as an allrounder.

    So if one has to pick him he should be picked as a top 6 batsman. I can't bring myself to do that. There are atleast a dozen Aus batsmen I'd rate atleast slightly ahead of Simpson : Bradman, G Chappell, Doug Walters, Ricky Ponting, Neil Harvey, Victor Trumper, Charlie MaCartney, Steve Waugh, Stan McCabe, Allan Border, Michael Hussey, Matthew Hayden.

    I can't rate Simspon ahead of any of these names. [[ I would only say that Simpson is a candidate for a spot in the post-war Australian team. Ananth: ]]

  • Gerry_the_Merry on September 29, 2012, 16:06 GMT

    On His comeback, in Sri Lanka, Warne outclassed Murali completely, taking 5 wickets in all 4 innings. After that it was hard to argue against Warne. [[ I am sorry, Gerry, this is one of the most reality-challenged comments I have ever received. If you mean, on comeback in SlK (Test # 1685), the following are the facts. In the next two Tests between Slk and Aus, Warne captured 20 wickets. You have convenientlky forgot to mention that Murali also captured 20 wickets. Damn the results. What more could Murali have done other than 10 wickets per Test and a 28-ball 43. At the beginning of Test 1685, Warne was ahead of Murali in career wickets. For the rest of their careers, the summary is given below. Warne: 37 Tests 212 wkts at 25.2 & Wpt of 5.7 Murali: 47 Tests 309 wkts at 22.6 & Wpt of 6.6. May I know where this "outclassing" took place. There is nothing wrong in your preferring Warne. However pl do not offer such outlandish arguments. Ananth: ]]

  • Nitin Gautam on September 29, 2012, 14:25 GMT

    @Som Fully agree on everything you said about Murli...no one can question his statistical superiority over any bowler. sadly if choice has to be made between players like murli & warne some amount of personal preference come into picture. Murli is surely the don of bowling & as wapsting said arguments can be made & have been made to prove who is better than whom but Anyways in a playing XI I would have 1 spinner only & that will be decided by the venue as i mentioned in my previous post. Regarding action controversy, when ICC has cleared him NO ONE should bring this to malign the genius Murli. also ppl use BD, ZIm to downgrade the wickets captured by him to glorify warne & I dont patronize this too..every test wicket & every test run & every test win is as important as any other.. Having murli in XI & warne in reserves as exactly same for me..no bias here at all in those stalwarts..lucky ot witness themboth

  • Waspsting on September 29, 2012, 11:34 GMT

    @Som - i do put Murali in my world 11, and if i had to choose 1 spinner, would pick him.

    I believe the choice of Warne over him, usually, has more to do with 'hype' than anything else. Ask people who choose so and they say vague things about "charisma", "the game coming alive", "swagger" etc. - nonsense basically.

    In crickino all time 11, Warne was UNANIMOUSLY chosen as the spinner.

    That said, stats aside, i don't dismiss the Warne>Murali opinion off hand (i.e. - there are VALID arguments to that effect, though i personally think the counter-arguments are much stronger)

    things like away records - I believe Warne comfortably edges Murali there.

    On the flip side, Murali has better average, strike rate and economy rate in Test, 1st class, non-test 1st class, ODI and 20-20 cricket!

    Rationally i think its hard to go past that.

    And as Ananth says, the chucking thing, rightly or wrongly, will subjectively always be held against Murali

  • Harsh Thakor on September 29, 2012, 7:41 GMT

    I find it has become customary to prefer Marshall to Lillee but I still rate Lille ,arguably,the best paceman ever ,with his classical variety.In Ananth Narayana's analysis of 2009 Lillee tops one list and is only 2nd to Hadlee ,ahead of Imran,Mcgrath and Marshall,with better match performnaces.In Richard Sydenham's book of 100 all time cricket 11's Lille is voted by 53 players in their 11's against Marshall's 35.

    I also ascertain that Wasim Akram,to me the most naturally talented paceman of all would be a better choice in a world 11 than Imran or Hadlee with his left arm variety and superior trajectory.No pace bowler could do what Wasim could do with a cricket ball ,who was the best ever in the art of reverse swing.

    Coming back to the team on his day Lara was a better match-winner than Sachin .However he lacked his rival's consistency.In that light I feel George Headley is the strongest candidate to join the Don.Imran and Hadlee were better suited to the 2nd 11. [[ Too few Tests to go by. Then what about Graeme Pollock who probably faced equally good bowling. Ananth: ]]

  • Som on September 28, 2012, 17:55 GMT

    @NG - Well applying the logic of horses for courses, we could choose alternative batsmen for any World XI. I am sure among the top 10 list of batsmen some would have done better than others under various conditions, geographies. But the purpose of the World XI is to choose someone who rides over these inequities. And in that case, even if Murali is a few mm better than Warne (which is highly debatable, plethora of analyses from Ananth has shown he is shoulders ahead of any kind of bowler in the last 2 decades, and Warne does not even come in the top 3), he would slot in.

    Is it then a matter of personal opinion more than what the data says, or am I missing something? Actually I would ask this to anyone who would not select Murali in any all time XI of any era.

    Is the issue of 'action'(long put to rest), something that rears up somewhere? [[ Som, the nail on the head time. Murali is way ahead, on numbers. It seems to be a matter of personal choice. It is also a fact that Australia, and to a lesser extent, India, have never seen Murali at his best and this may be the reason. Of course a few people would still hve the "javelin throw" block. Ananth: ]]

  • Nitin Gautam on September 28, 2012, 10:57 GMT

    @Wapsting Re. Lara...while he may not be at his sublime best against super express bowlers (though I dont agree to this as he handled all the bowlers with extreme ease when he was in his own elements) he is still damn good player to warrant a place in any team for 1992-2012 era.

    His high backlift play raised certain doubts about his timing against bowlers like waqar & donald yet the same style brought him many runs at very gud pace. Considering he is positioned 5th & Kallis 6th...that would make a gud pair to bat (although notional)

  • Nitin Gautam on September 28, 2012, 10:51 GMT

    @Gerry all the players you suggested are very gud players although not regular openers so not in contention. Lax & Dravid have opened in tests but while dravid had some success, Lax was poor. dont know for sure but perhaps KP & Inzamam never opened in tests. Just thought that for a team to compete against best of the bests, I cant have makeshift openers @Som I agree on Murali but really you cant have murli & warne apart by more than few millimetres..as Anantha suggested I will have warne in Aus & WI & murli in Asia while in Eng & SA..someone has to sit out..no diff among them really for me. @Wapsting Thanks for so many worthy options but I guess Hayden & michael vaugn can be a gud pair of openers..Mark taylor is really very gud option.. By all this discussion one thing is clear..rest of the 9 & 5 in reserves proves that this era is terribly short of classical test openers like gavaskar, sutcliff, greenidge, Gooch etc..

  • Waspsting on September 28, 2012, 10:31 GMT

    @Ntin

    i agree with you about Hayden. i didn't see his early games against WI, but have heard he was quite poor. in his pomp, i assessed from his style that fast bowling might sort him out, and wasn't surprised by his ordinary showing against Eng 2005 - a good attack, but not a patch on the bowlers you've named.

    Still, as Boll says, Atherton seems a strange choice to me. For one, he came a long way short against the top quicks he faced - 2 Ws, Ambrose and Walsh. McGrath also sorted him out. Only Donald did he do well against among the top paceman.

    Given your objections to Hayden, I'd think Atherton would also fall short?

    Against the top pacers - I think i'd prefer several to Atherton - Mark Taylor, Slater, Graeme Smith, even Alec Stewart - what do you think?

    (we can save the next point for the next article, but there's one other player who i think would REALLY have struggled in the 70s-80s fast bowling jurassic age. Brian Lara)

  • Som on September 28, 2012, 8:45 GMT

    @Nitin - Good team but Murali in reserves is interesting. Statistically or otherwise, he has been the Don of the last 20 yrs and therefore perhaps the very first one to be included in any World XI of the last 2 decades. And any criteria for balance and variety would only come into the equation to rally around his presence. Thoughts?

  • Gerry_the_Merry on September 28, 2012, 5:37 GMT

    Nitin Gautam, why not take Pietersen, Inzamam, Dravid, Laxman into consideration for opener's spot?

  • Nitin Gautam on September 28, 2012, 4:13 GMT

    @Love goel Trescothick was a very gud player agreed attacking & easy on the eyes however considering how prone he was to getting into psychological doldrums & withdrawing playing I wouldnt consider him against those bowlers of 80s... @Boll Just realized I missed a classical opener of 90s & '00s who in one of previous anantha's articles proved to be a real tough nut & scored more than 75% of his total runs against a very gud bowling..Michale Vaughn...So I guess hayden & Vaughn can be considered as openers & that will make it left, right combo too...in fact Anwar is also a very gud option to replace hayden if he turns out in his pre 1998 form..thus removing Sehwag & Atherton

  • Boll on September 27, 2012, 14:06 GMT

    @Nitin, yep Athers had a couple of wonderful performances - 1 century in 66 innings vs Australia (ave 29, Sr 37) pretty much puts him out of the reckoning for mine though. Everyone`s going to struggle against the bowlers you mention, Athers may have averaged in the teens.

  • Love Goel on September 27, 2012, 13:53 GMT

    @Ntin - that's a fine team

    But is it so easy to dismiss players like Justin Langer? How will Marcus Trescothick be compared to Atherton.Or how does the South African Gary Kirsten rank amongst openers

    Surprisingly all 3 are lefties. May be an extra point for that I will go for Kirsten over Atherton

  • Nitin Gautam on September 27, 2012, 13:31 GMT

    @wapsting

    Certainly Hayden was the best opener post 1999 that Australia had..attcking, dominating & thoroughly intimidating but how he performed against WI quicks made me doubt on his abilities against holding, imran, garner, hadlee, Lille etc..

    Although I have to agree Hayden was a fear inducer in most of the bowlers on 1st day of any test match..May be cook or hayden can be taken as openers

  • Nitin Gautam on September 27, 2012, 13:25 GMT

    @Gerry I completely agree on Atherton but as I mentioned I didn't find any opener from 92-12 who can skilfully tackle the frightening, colossal bowlers of past. I have my reservations on sahwag also since he has not fare much better still I had to chose so I chose Atherton for his dual with donald (sad to miss him in my team of this era..certainly deserving )& Sehwag for his impact. Of all openers you mentioned, IMO only cook can be termed as classical opener with great record albeit in highly batsman centric times while Heyden was terrible in 90s against top quality WI bowling, smith, gayle, strauss etc would not be more than journeyman against those legendary bowlers. & Gambhir???? no way to even consider...have to say Bowling is still so rich in my time which easily competes with best of all time..batsman have a relatively small pool to chose from..

  • Waspsting on September 27, 2012, 12:14 GMT

    general stuff i noticed in the analysis

    Walcott and Weekes' adjusted average goes up quite a bit - but i think that was a pretty high scoring period(?) particularly in relation to Hutton's going down - i perceive that he was probably comfortably the biggest scorer of his period (?)

    - Sutcliffe, Hammond go down... because of Bradman outscoring everyone at the time(?) (even the batsmen suffer because of him, the bowlers who bowled to him obviously have their figures ruined)

    -Simpson is a surprise. He took aeons to get his 1st 100, but was superb afterwards - even when making a comeback in his 40s. Given his slip catching and useful legbreaks, i think he's somewhat underrated overall - a good contender for an all time Aus 11 spot.

    -Hill looks very impressive. Streets ahead of his contemporaries - almost Hobbs like.

    (will skip on commenting on the bowlers - given that we seem to have found flaws with its structure) [[ Simpson is underrated, there is no doubt. Bradman, Hammond, part of Hobbs, part of Hutton, certainly the between wars period was rich in batting. Hill was fantastic. His 188 is one of the most underrated innings. When I had it in my Wisden-100's top-10, I had to educate many people on that classic. Ananth: ]]

  • Waspsting on September 27, 2012, 11:47 GMT

    @Ntin - that's a fine team. the only surprise to me was the openers.

    Why not Matt Hayden - far and away the outstanding opener of the period? [[ The highest average amongst modern openers. Ananth: ]]

  • Boll on September 27, 2012, 11:35 GMT

    @Nitin Gautam. Yep, great team...apart from Athers!!certainly not doubting his guts and resilience, but against the best (Aus/WI) of his era he averaged 30 in 60 tests. Cook, Strauss, Hayden, Langer, G.Smith, Gayle, Gambhir...?

    Lillee et al. vs Atherton - licking their lips.

  • milpand on September 27, 2012, 9:21 GMT

    Shiv is also a very violent batsmen. He scored the 4th fastest test 100. Greek allrounder in ethics, physics, politics, logic and marine biology in his spare moment or three commented, "One swallow does not make a summer, nor does one day; and so too one burst over a short period does not make a batsman violent." [[ Since I was off my responding mood, I passed the comment as it was. Gerry should realize that violence is not a trait normally identified with batsmen: we normally think of violence when we think of "Sholay", "The gangs of Wasseypur", "Pulp Fiction", "Kill Bill" et al. On your comment what about the two hundreds of Gavaskar: the Delhi one against West indies in 1983 and his penultimate ODI innings. Ananth: ]]

  • Nitin Gautam on September 27, 2012, 5:58 GMT

    If I have to select a world XI (team of 16) than I wont go for players whom I have not seen yet I guess players in last 20 years have been such a force that they can challenge any team in history. Players like Botham,Imran,Gooch,Hadlee were at the fag end of their craeer & I never saw them at their very best so I wont risk including them.so here is my team. 1.MIke Atherton 2.Sehwag 3.Ponting 4.Sachin 5.Lara 6.Kallis 7.Gilchrist 8.Wasim Akram 9.Ambrose 10.Warne 11.Glenn Mcgraw

    Reserves:-Pollock, Dravid, Murli,Waqar,Saeed Anwar.

    However hard I tried, I could not find openers from this era who can fight with greatest of bowlers in history (holding, garner, Lille, Imran,hadlee etc)however Mike Atherton with his steely resolve pairing with Sahwag are the best choices according to me. From all the players that I have seen at their best (92 onwards), I guess bowling can compete with any pair of bowlers in world & WK an all time best option & allrounder 2nd only to GS.This is a very gud team

  • Gerry_the_Merry on September 26, 2012, 4:09 GMT

    Imran Khan was also a fairly violent batsman, at times even more violent than Richards, Kapil Dev and Botham. He once smashed 25* in 7 balls in a 1985 one-day match in Sharjah coming in in the second last over, against West Indies, including 22 runs off Joel Garner, with two sixes flying out of the stadium.

  • Ananth on September 26, 2012, 1:21 GMT

    Since the comments have gone towards another direction, let me summarize the decade analysis article and the follow-ups. It is agreed that 10 years, while a nice period to hang one's hat on, is rather arbitrary, especially for the bowlers. It is also essential to look at a shorter period for the bowlers whose careers are shorter. Hence I will re-do this analysis with varying periods for batsmen and bowlers. The snippets I provided on the worst period for players generated enough interest for me to consider including that analysis also in the follow-up. I would do a best period, worst period and the variances. However as far as I am concerned, the career is the defining point. While the period analysis will show the player at his best and worst, the Career analysis will define his value to the team and his place in the overall schema. Over a fairly long period what the player delivered in comparison to what his peers, within and without the team achieved, must form the ultimate measurement of the players's value and worth to the team. So I will do that analysis also. My next article is the Lara tribute and this will be followed by a very interesting one on Bowler pairs. Ananth

  • Randy Jackson (fake) on September 25, 2012, 13:06 GMT

    "Be who you are and say what you feel, because those who mind don't matter and those who matter don't mind." - Theodor Seuss Geisel [[ Thank you, my dear friend, for the clear words and the payback of the commenter's name misuse. Ananth: ]]

  • Waspsting on September 25, 2012, 12:02 GMT

    I wanted to hear what everyone thinks about chemistry of the team (including Ananth).

    For expressing chemistry, I'll use a 3 digit code - #of batsmen-#of pacemen-#of spinners (collapsing allrounders into their strongest suit - e.g. Sobers and Kallis would fall under batsman, Imran, Miller would be called pace bowlers)

    The most common chemistry for teams is 6-3-1, as Harsh has selected.

    Dr.talha and I have gone in with 5-3-2

    Gerry has 5-4-1

    ----

    Obviously, a 5 bowler line up is more attacking than a 4, and in choosing a World 11, i think its positive to go in for it.

    how to divide the 5 into pace-spin?

    3-2 seems best to me, because the marginal utility of a 2nd spinner is probably greater than the marginal utility of a 4th pacemen (particularly if Sobers is around to bowl too)

    thought? [[ Not now, WS. I have to complete my next article and do some work on the T20 WC. As I have already expressed, let us do it later. Ananth: ]]

  • Waspsting on September 25, 2012, 11:54 GMT

    i have 5 reserves in mind to make up a squad.

    Hobbs, Headley, Knott, Waqar Younis, and O'Reilly. opener, batsman, keeper, paceman, spinner.

    in conditions highly favoring pace bowling (sideways movement) and/or particularly unsuited to spin - drop one of the spinners, stick in the extra batsman (Headley)

    given the 5 bats, 3 pace, 2 spin combo i prefer - certain balances are important for the 5 reserves.

    MUSTS -an opening bat -a keeper -a paceman

    the other two are little more open. 3 spinners in the squad is unnecessary (though i've done so) - you'll almost never have need to play 3 spinners in the starting 11.

  • Dr. talha on September 25, 2012, 7:28 GMT

    @Gerry the Merry. As ananth said, 11 is not sufficient. While making this eleven believe me i wished cricket was game played by 15 players. And when i started finalizing the bowlers it changed to 20!! Ambrose plus two of the greatest batsmen of the modern era, sachin & Lara, dropped from my 11.. How can we drop Akram who has bowled some of the most magnificant deliveries in the history of the game (from Lamb & lewis in 92 WC to dravid's dismissal in chennai 99). Will i ever want Waqar to be in the opposition. What if the bowl starts to reverse? The guy whom u can trust to win you matches, even if the opposition needed only 40 to 50 runs with 7 or 8 wickets remaining. Remember his 35 wickets in 6 tests against the mighty WI(early 90's. It was the best display of pace & swing i have ever witnessed. He got rid of Greenidge, Haynes, Lara, Hooper, Richardson etc.(few of the best players of fast bowling) multiple times.Never seen greenidge struggle, the way he did in front of Waqar. [[ You are better off reserving this effort for the main article. Ananth: ]]

  • Nitin Gautam on September 25, 2012, 5:33 GMT

    ohh ok ok

    I just realized that you have mentioned this in your comment to him. I Failed to notice that.

    However I really wish whoever wants to know these material degrees, learn to appreciate what value & insights you bring to all cricket enthusiasts & I have not even started to highlight the mammoth efforts you must have put in all these articles so far. [[ Nothing will be gained. These are the people with closed minds. Ananth: ]]

  • Dr. talha on September 25, 2012, 5:24 GMT

    @Ananth. You hav already had more than 200 comments on this article. I believe when u will come up with an article on all-time world X1, be prepared for more than 400. and if you dont have Sachin in that, be ready for 1000!! Still remember the time, when i watched indian matches just because i wanted to see sachin bat. That was the greatness of the man. [[ Who I have in the XV is my prerogative. However I will never do anything against my own common, cricketing and analytical sense. Ananth: ]]

  • Nitin Gautam (Really) on September 25, 2012, 5:18 GMT

    Posted by: Nitin Gautam (Not really) at September 25, 2012 4:32 AM What is your degree in Ananth? Sometimes I feel you are really dumb.

    Seems I am missing some fun here :) [[ The "Not really" was added by me. The fun is in my imaginary degree and the tracking of sender. Anyhow your mailid is private and while one can borrow your "public" name they cannot borrow your mailid and IP address. Regards Ananth: ]]

  • Gerry_the_Merry on September 25, 2012, 5:17 GMT

    Dr.Talha - you have left out Curtly Ambrose...!!!! Beware - he will be in the opposing XI.

    The second XI could have a bowling combination of Holding / Lillee / Ambrose / McGrath / Akram. Sounds equally ferocious, not one bit less.

    But there is just too much talent to cram into a single World XI. Imagine, if after 50 years we were to do this, would we leave our current choices out merely because new batsmen and bowlers took stage?

  • Dr. talha on September 25, 2012, 4:45 GMT

    As i have mentioned earlier i will also have both Murali & Warne in my side. So from number 6 to number 11 my side will be the same as Wapsting. The top order will also be the same. But the only difference may be Viv in place of Sachin. Tough decision, as i am a huge fan of sachin, but one of the two, has to be my 12th man. Lets wait till Ananth's article..Final X1:

    Gavaskar Hutton Bradman Viv/Sachin Sobers Gilchrist Imran Hadlee Marshal Warne Murali

  • Nitin Gautam (Not really) on September 25, 2012, 4:32 GMT

    What is your degree in Ananth? Sometimes I feel you are really dumb. [[ There is a branch called "Dumbology". I have graduated in that very exclusive branch from 'Dumb'arton university, Scotland. Nitin (the real one) I am indeed sorry that this nincompoop from Richmond, US, has used your name. He does not realize the mutual regard and respect we have for each other. I deliberately retained your name, just to expose him. Anyhow you are safe, staying about 10000 kms from him. Incidentally he is the same Randy Jackson whose inane comment was received a few days back. Great fun, though. Ananth: ]]

  • Ramarao on September 25, 2012, 3:01 GMT

    Good to see support for Imran, who is a natural leader genuine allrounder who can win a match with both bat and Bowl. Inspite of all this, he may well fall short of Greatest of them all Garified Sobers. Its very sad to see some people trying to degrade Tendulkar's performances, the value he added to batting and his talent. Other players who played in his era might have comparable records with him for a period of time, but overall there is no one who comes close when you take only test or only ODI or overall ODI and Test record. Lara comes very close in tests and Ponting comes second overall to Tendulkar in the period he played. The people involved in Cricinfo all time X1 or people of caliber of Richards, Sobers, Lara, weekes Warne and Bradman did not appraise Tendulkar for nothing.

    Ananth, can you come up with the analysis of Best batsmen overall in ODI and test toghther over the period of Sachin career(23 years) [[ Not really worth it especially as this period excludes the career of the only player who could challenge SRT in the combined stakes. Nothing gained in my saying that SRT scored millions of runs in the two formats together. Ananth: ]]

  • Ananth on September 25, 2012, 2:22 GMT

    It amuses me to see World XIs being selected. I agree it is a pleasant exercise and everyone has a right always to do so. However it is my firm feeling that selecting a eleven is not sufficient. Properly done we should select a group of fourteen or fifteen, with a bowling group to choose from. If I have to simulate a 5-Test series across the world I would insist on it. Would you play Warne everywhere or Murali everywhere. I would probably play Warne in Australia and South Africa and Murali in South Africa and sub-continent. Would you play Imran Khan everywhere. I would rather keep him for sub-continent and Hadlee for some surfaces. I may play Ambrose in Australia but Marshall in England. Maybe I may select McGrath everywhere, or not. Some time in the future I would do an article on these lines. Do a selection of the best fourteen or fifteen using an analytical look at all countries and let the readers come with their own selections. Maybe poll the readers' responses and come out with a final selection. But I have no problems with comments on these lines. Ananth

  • dinesh on September 24, 2012, 16:33 GMT

    "These are only suggestions. My core group of must-be-selected players will be revealed in a few days. Until then you have to wait. Ananth"

    One comment from you and there are already 7 World XI's. You sure know how to keep people on their toes as Selecing a World XI is every Sports lovers Dream irrespective of the Sport.

    I remeber You saying Selecting a WorldXi is Long and a Huge Subjective Process with almost nothing to gain.

    And i dont want to Quash Others excitement but i feel A WorldXI form Ananth Should come only after you have finished almost every kind of Article.#WinkstoHimself [[ To my knowledge I have never selected a World XI and am unlikely to do so in the bear future. One reason why I am staying away from any comments now. Ananth: ]]

  • Som on September 24, 2012, 14:55 GMT

    (...cont)

    So in my mind, the batting spots are still open and debatable: Contenders - SRT, Viv, Chappell, Border, Barrington, Weekes, Walcott, Kallis, Lara, Ponting, Dravid, Miandad. Even if SRT just edges in (after default inclusion of DGB, Sobers), we still have 1 or 2 slots open. And one-on-one Kallis will perhaps edge out Lara, Ponting, Miandad and Dravid. So essentially we have Viv, Chappell, Border, Barrington, Weekes, Walcott and Kallis to choose from. Even when forming a second XI, this shortlist perhaps has a surplus of contenders. How are we to judge Barrington, Weekes and Walcott in this list vis-a-vis others, whom I know a bit about. Thoughts? Or are there others whom I am missing out. Or is there reasons why some whom I pushed down the order, needs to be brought back into the equation.

  • Gerry_the_Merry on September 24, 2012, 14:44 GMT

    Gavaskar Hutton Bradman Richards Sobers Gilchrist Imran Marshall Warne Holding Ambrose

  • Som on September 24, 2012, 14:43 GMT

    Excellent teams Wasps, Harsh. In the last 3 decades, I have never watched a batsman who had the promise and potential as Lara. But what Lara achieved is not the very best. While Viv, SRT and Chappell did better translate, Gavaskar, Border, Kallis, Ponting and to a lesser extent Miandad and Dravid matched him. So Lara's inclusion in an all time XI is not so undisputed. In my mind browsing through all the analyses and stats, I have never quite gotten it quite clearly as to which 2 among Hobbs, Hutton and Gavaskar should be picked as openers. Having watched Gavaskar and SRT, I still have my doubts whether SRT was definitively the best test batsman India produced. But this analysis in particular and many others makes a strong case for SRT's inclusion in the middle order of an all time XI. I find this ironical (vis-a-vis SMG). And regarding the two W's and Barrington - I do not know why I will ignore them over anyone other than DGB and Sobers. (cont...)

  • Waspsting on September 24, 2012, 13:51 GMT

    bit more about my 11.

    The KEY ELEMENT in mine is chemistry of 5 bats, 5 bowls (considering Sobers a batsmen and Imran, Hadlee bowlers - for simplicity's sake).

    With that Gilchrist's inclusion over say Knott becomes an easier decision (with a 6-4 combo, would probably choose Knott).

    I particularly wanted a Pakistani bowler - meaning a guy who could use the old ball, and was expert at removing the tail, specialized in swing (not cut), and in-movement (given that one intuitively grasps that most bowling contenders are out-movers)

    Again, the 5-5 comb makes Imran's claim to that spot over Wasim and Waqar stronger than for a 6-4.

    Hadlee's selection also helped by the 5-5 combo

    Note however, purely as bowlers, Imran and Hadlee are as strong a contender as anyone.

    Hobbs strong contender for openers spot.

    Tendulkar's spot could go to anyone of 12 other guys.

    Marshall spot could go to number of players.

    Warne esp. threatened by O'Reilly, Grimmett. Murali less so - for chemistry's sake

  • Harsh Thakor on September 24, 2012, 12:44 GMT

    The only definite selections in an all-time list would be Bradman and Sobers.Bar these 2 we could have two evenly matched teams of the 1st and 2nd 11's.

    team 1. 1.Barry Richards 2.Gavaskar 3.Bradman 4.Sachin Tendulkar 5.George Headley 6.Gary Sobers 7.Adam Gilchrist 8.Shane Warne 9.Malcolm Marshall 10.Wasim Akram 11.Dennis Lillee

    Team 2 1.Hobbs 2.Hutton 3.Bradman 4.Viv Richards 5.Lara 6.Sobers 7.Imran 8.Knott 9.Mgrath 10.Barnes 11.Murlitharan.

    With Bradman and Sobers in both teams are virtually,equal in batting and bowling.Imran,Murli,Mcgrath and Barnes could well match Lillee,Marshall,Akram and Warne. Similarly the likes of Hobbs ,Hutton,Viv Richards and Lara could match with Barry Richards,Gavaskar,Tendulkar and Headley.

  • Waspsting on September 24, 2012, 12:39 GMT

    you'd lose the edge of twin spin attack (how often when the wickets turning and Warne was the danger man have we not thought how much faster the wickets would fall if MacGill could have been fitted into the team?) - but gain in batting solidity.

    I think the batting's solid enough - for the reasons given above. And the value of a twin spin attack - a rare luxury, but a very useful one - is more valuable to the team than more batting.

    Note the left-right sandwiches for batsmen throughout the order.

    Note the combo of solid openers, sensibly attacking middle order and balanced-on-the-side-of-attack lower order (Gilchrist - butcher, Imran anchor and the others brisk)

    Note variations of bowling. Marshall express blaster, Hadlee nagging - both primarily out-movers. Imran the in-mover, with new ball and old. (plus left arm stuff from Sobers) Warne turning away, Murali turning in (Sobers doing all his slow stuff)

    Fielding/catching not bad, could be better.

    Thoughts, gentlemen?

  • Harsh Thakor on September 24, 2012, 12:30 GMT

    I can never forget the outstanding performances of Viv Richards Greg Chappell and Dennis Lillee in Pcaker cricket but for which they would have far beter stats Lilee would have had over 450 scalps adding the Packer era while Viv Richards and Greg Chappell would have made 28 centuries.Had Brian Lara played for a strong West Indian team he might well have been the best West Indian batsman of all.He championed the cause for a relatively weak side and at his best he was arguably 2nd to only Bradman.

    Imran is the best match -winning cricketer after Sobers deservingly ranking amongst the top 8 cricketers of all time.However,he was mainly a fast bowler who could bat.Stats have never done justice to Gary who could be also classed as an all-time great left arm bowler who in his peak captured 194 wickets in about 47 tests at 29 runs,per wicket.In his peak period I feel Ian Botham was the best all-rounder after Sobers who could swing a match with both ball and bat like the 1981 Ashes.

  • Waspsting on September 24, 2012, 12:29 GMT

    I have no clear favorite re: Sobers and Imran - but wish to stress that there isn't much between them (contrary to the common almost cliche that Sobers was streets ahead of all other all rounders)

    As a few people have said, I would have them both in my World 11 (Hadlee too, as it turns out)

    1. Gavaskar 2. Hutton 3. Bradman 4. Sobers 5. Tendulkar 6. Gilchrist 7. Imran Khan 8. Hadlee 9. Marshall 10. Warne 11. Murali

    having 5 batsmen (not 6) leaves batting potentially vulnerable. But a combination of Bradman being worth 2 bats, Gilchrist being worth more than we'd expect, Imran & Hadlee's batting and Marshall & Warne's wagging leaves the batting as solid as desired.

    centurions til 8, 50s people till 10.

    In exchange for the potential batting weakness, i get the sternght of 5 bowlers - and the ideal balance of 3 pace, 2 spin (all best of the best). with Sobers to back both pace and spin.

    The alternative is to drop Warne or Murali and include another batsman (cont)

  • Harsh Thakor on September 24, 2012, 12:20 GMT

    On mere statistics this would be the all time 11. 1.Hobbs 2.Hutton 3.Bradman 4.Weekes 5.Tendulkar 6.Sobers 7.Imran 8.Gilchrist 9.Warne 10.Barnes 11.Muriltharan.

    My moral xi on true merit.(considering only players discussed) 1.Hobbs 2.Gavaskar 3.Bradman 4.Viv Richards 5.Tendulkar 6.Sobers 7Alan Knott 8.Warne 9.Marshall 10.Lillee 11.Mcgrath

    Ananth,to me no one all-rounder can really be compared with Gary Sobers,be it Imran or Kallis.No all rounder has made an impact so consistently with both bat and ball like Sobers,be it the 1966 series in England of for the Rest Of the World.Arguably he is the greatest cricketer of all who who could turn the complexion of a game or series more than Bradman,Imran Tendulkar etc.Imran was the best all-rounder from 1981-1988.However he was never silmuntaneously a champion with the bat and the ball,like even Botham at his best.Lillee and Viv Richards would have come far ahead had it not been for the advent of Packer cricket. [[ I will publish all these Best XI lists without any comment from me. Ananth: ]]

  • unni on September 24, 2012, 11:27 GMT

    Regarding the batsmen statistics I have a question regarding Kallis's stats. You mentioned that he had an average of around 6 from 1998 to 2008. But the statsguru query doesn't give this.

    http://stats.espncricinfo.com/ci/engine/stats/index.html?class=1;filter=advanced;orderby=runs;player_involve=2232;spanmax2=31+Dec+2007;spanmin2=1+Jan+1998;spanval2=span;template=results;type=batting

    Can you please verify? [[ It is a typo. It should be between 2001 and 2011. Since corrected. Even then you should not take 1 Jan and 31 Dec. You should take the exact Tests. Pl refer to the Excel sheet. Figures for Kallis. End of 1562 (2001) 2952 in 72 End of 1988 (2011) 11947 in 208 For 10 years 8995 in 136. Avge 66.13. The Test in the Excel sheet (2006) is the exact 10 year end Test. The last Test within the period for Kallis is 1988. But the period-based query in Statsguru should come out correctly. Ananth: ]]

  • Som on September 24, 2012, 9:46 GMT

    And Ananth, wouldn't it be impossible to leave out number 9 for Marshall and 10 for Murali with what we have seen here and in other analyses. And most definitely 3 is the great man and one of the openers perhaps has to be Hobbs (?). If so, 7 choices are clear, ain't it ! [[ These are only suggestions. My core group of must-be-selected players will be revealed in a few days. Until then you have to wait. Ananth: ]]

  • Dr. talha on September 24, 2012, 5:49 GMT

    I think sobers and Imran both were good enough to make their way in the world X1, only on the basis of their batting and bowling respectively. As an all-rounder i would certainly pick Imran. I have always believed you should be a better bowler than a batsman in order to be called a genuine allrounder. Even Sanath Jayasuriya has over 300 ODI wickets, that does not mean that he would be selected in the all-time ODI X1, when we will be selecting an all-rounder. We are not noticing that Sobers had a bowling strike rate of 91.9, which is very very ordinary. A lot of batsmen in the history of the game may have fewer wickets than sobers, but had a lot better strike rates. Kallis has a SR of 69. But tell me how many of the great bowlers had a batting record similar to Imran?[[ If the fifth bowler has a BpW value of 90, it really does not matter. Similarly the no.7 batsman could be 30. However if you get a 37 batsman why complain. And if we have Gilchrist at no.7, we will have a no.8 batting at 37. Surfiet of riches indeed. My feeling is that Sobers walks in as the no.6 batsman. So his bowling is a bonus. And Imran has the best credentials as a captain-all-rounder. So it looks like I have fixed my 6/7/8. Ananth: ]]

  • Gerry_the_Merry on September 24, 2012, 3:39 GMT

    Contd...

    A look at the top 5 in Sobers' group would not inspire a World XI selection apart from Sobers himself. In other words, this period did not see outstanding and productive talent in a high scoring period, diluting the quality of Sobers' lead a bit (purely subjective). Imran's lead is in a period where we would be in a fix over whom to leave out in a World XI. Imran's group is far far classier than Sobers' top 5.

    So Imran has equal lead over peers in same period in bowling as Sobers in batting, Far superior competition to beat (perhaps the best bowling period in history) and finally better opposition abroad.

    Sobers and Imran were both captains, titanic figures, and very good all-rounders. But here Imran does not fade in comparison - with a century against a top class Windies attack, and an accomplished performance in England as a batsman in 1982, he has a few batting feathers in his cap.

    My vote - Imran.

  • Gerry_the_Merry on September 24, 2012, 3:33 GMT

    On Sobers v/s Imran, and who was better, a few observations:

    Sobers' best batting period, as per the article was 1958-68, with an average of 73. So if he did a mere 65 in 1962-68, perhaps it was not his best period at all.

    I looked at a period of 10 years for Sobers (as a batsman) beginning 1/1/1958 and 7 years for Imran (as a bowler) beginning 1/1/1980.

    Top 5 batsmen were - Sobers - 4842 runs @ 72, Barrington 6296 @ 61; RG Pollock 1739 @ 58, Hanif Mohd 2915 @ 54, Kanhai 3944 @ 51.

    Top 5 bowlers were - Imran - 184 @ 15.92, Hadlee 227 @ 18.91, Marshall 228 @ 20.35, Garner 198 @ 20.83 and Holding 184 @ 22.84.

    Of Sobers' runs, >50% came abroad @ 65 and rest at home @ 82. Of Imran's wickets, 118 came at home @ 16, and 66 abroad @ 18.7. So far so good.

    But Imran's hauls abroad included a series in England 21 @ 18.6, India 14 @ 18.9 (2 tests out of 6), and 16 @ 19.5 in Aus, all challenging.

    Of Sobers' runs abroad, there was one stunner against England - 722 @ 103, but rest ordinary.

  • Dinesh on September 23, 2012, 18:15 GMT

    Ananth: I personally Prefer 9 and 7.Though not much of a difference. But as WaspSting Said I trust Your Judgement. I think as you probably Start doing the Article it would be more clearer and you will probably See some thing among 10,8 and 9,7.

    But My Judegement on this is as good as Anyone else's as i dont have Many examples to back my Preference as i only Derive my own Analysis only after you put forward yours.

    @Waspsting:I agree with all your points.One of those Players whose metamorphosis was unbelievable.I think his Career is a lesson to All others that its never too Late to learn new things and Transform Your Career to altogether a different level.

    I would add another player here who threatened to Transform but his body and Mind were more fragile and he Turned in a Could have been Player. Andrew Flintoff was that.From 03-05 he Averaged Almost more than 40 with the bat and in 05-06 he averaged 24.9 in Bowling picking 111 wickets.A player who could have been Great but.... [[ Probably the expectations of the "next Botham" did not sit well. And the responsibilities of captaincy thrust too soon. First 30 Tests 55 wkts at 45. And then a transformation. Ananth: ]]

  • Som on September 23, 2012, 16:04 GMT

    Variance as a function of career length would be a useful measure too, because for players with short career length due to a great number of years of overlap between the best and worst periods, variance will have a different order as compared to those with much longer career lengths.

    9/7 years may have some merit as they are almost 66.6% of the average career length, and for players with most mutually exclusive best and worst phases, the overlap is just 50%.

    Actually there are many ways to think about this and there may not be a very precise choice - 9/7, 10/8 both may be equally justifiable.

    Also another way to arrive at this number is to find out the average number of matches played by the shortlisted players and then make a call on how many games qualify as a good tenure for such evaluations which incorporates multiple form/out of form, ageing, multi-country exposure scenarios. And this may as well end up being 9/7 or 10/8 or something close.

    Can't wait ! Thanks Ananth ! [[ Som, the strength of this analysis is its simplicity. You gys know the period is 10 years and could visualize things in your mind as well as do your own analysis. If it changes to 8/6 or 9/7 for batsmen/bowlers, you guys would still be comfortable. However I do not want to go beyond that. It will become too complex. Ananth: ]]

  • dale on September 23, 2012, 11:22 GMT

    Re:Sobers (62-68)He scored 94 runs per test @65.5 and took 3.8 wickets per test @27.2 Imran (82-88)He scored 43.5 runs per test @46.2 and took 4.9 wickets per test @16.8 With Sobers you are getting 50 more runs per test and with Imran you get 1 more wicket per test for 21 less runs conceded.

    As early as in '76 Imran bowled very fast but was yet to add or perfect the scientific element to his bowling ie reverse swing.During the WSC series he was the third fastest bowler on show after Thomson and Holding. As a captain Imran was as vital to Pakistan as Worrell was to West Indies.

  • Waspsting on September 23, 2012, 10:48 GMT

    like Adam Hollioke.

    But these guys all "transformed" into all time greats - they reached the highest standard and stayed there for a prolonged period (10+years).

    Pre-transformation, their numbers are ordinary - thus making their segment-wise career analysis highly variable, as you noted. Imran more than anybody.

    these guys virtually became different players.

    a subjective indicator of this might be that when Sobers was flaying attacks in the 60s or Imran running through sides in the 80s - the opposition captain telling his men "come on, guys - this guy used to be a #9 batsman" or "this was a nothing bowler once" would draw nothing other than laughs!

    its hard to even imagine Adam Hollioke making a comeback, and averaging 50 with the bat and 18 with the ball for the next decade.

    That basically, is what Imran Khan did

  • Waspsting on September 23, 2012, 10:34 GMT

    "The bowlers have shorter careers and warrant a lower period fixing. 10 years for batsmen and 8 years for bowlers, respectively, represent around 75% of the average career-length and seem to be perfect periods....Is there a case for 9 and 7?"

    I'd trust your judgement on the matter.

    a problematic point though might be in the lenghts of career, eps. the "prime years" of fast bowlers compared to spin bowlers.

    I assume a top spinner can keep up a high standard for longer than a fast bowler(?)

    An arbitrarily chosen figure for years for bowlers in general will hurt/help one type of bowler or the other(?)

    ---

    @Dinesh - Imran was what we might call a "transformation" cricker. Such players are rare, and he might be the most extreme example of one.

    Sobers started off as a spinner batting at 9. Something like Pat Symcox.

    Steve Waugh started off as a useful bat, useful bowler all-rounder. like say Chris Cairns.

    Imran started off as a bits and piece player (cont)

  • Ananth on September 23, 2012, 3:57 GMT

    Som, WaspS, Dinesh, Arch, Dale, I have done an analysis of qualifying batsmen and bowlers. Made sure that the war years are excluded. Made sure that the pre-1900 bowlers are excluded. Made sure that Warne/Vaas type of late order run-accumulators are excluded from Batting. Considered only batsmen with 3000+ runs and bowlers with 100+ wickets. After taking all these safe-guards, I have 158 batsmen and 145 bowlers. These represent probably 90-95+% of the relevant players in the two disciplines. The average career length for batsmen is 13.01 and for bowlers is 11.0. So all your points are valid. The bowlers have shorter careers and warrant a lower period fixing. 10 years for batsmen and 8 years for bowlers, respectively, represent around 75% of the average career-length and seem to be perfect periods. As and when I re-do this work, with best and worst periods, I will use these revised numbers. Is there a case for 9 and 7? Ananth

  • Gerry_the_Merry on September 23, 2012, 3:34 GMT

    "This is probably where Marshall and McGrath and Pollock edge him to a lower position. "

    Dinesh, this may be a fair comment numerically. But in the context of the evolution of bowlers and their tradition, it leaves out many important factors.

    Marshall, McGrath and Pollock, the three that you have mentioned, came into the teams in 1981 (technically 1979, but in a full strength team, only in 1981), 1992 and 1995 respectively. In each case, there was a top class fast bowler or bowlers to pick up the threads from.

    In Imran's case, in 1971, Pakistan was not a fast bowling country at all, and Imran lacked direction. By 1976-77, he had raised his pace and skill quite a bit, and saw success in West Indies and Australia, got picked by Packer, climbed the pace/skill ladder higher, and never looked back.

    A worst decade analysis cannot pay heed to these factors, and the fact that often, bowlers flounder a while before coming into the mainstream. Hence interesting, but not sufficient to rank. [[ But can open our eyes to additional insights. Ranking is anyway mostly subjective. You will never take Lohmann in your all-time Test team despite his being 50% better than the next. So subjective considerations come in. Ananth: ]]

  • dale on September 22, 2012, 21:34 GMT

    The 10 year period is perfect for the measurement of sustained excellence.This represents a good percentage of most players' careers. Re bowlers: The match starts out with runs being an infinite commodity while wickets are finite (40) but as the match progresses, runs become more finite and more measureable against the wickets.eg 200 runs to win and 10 wickets to fall. Then it is a matter of which discipline prevails over the other or there may be a draw instead of an outright result. The bowler often must be "assisted" by the fielder accepting the chances (catches) especially close to the wicket.Maybe more attention should be paid to the fielders because the tendency is to ignore this facet of cricket except when we discuss wicketkeepers.We know how valuable Gilchrist was to the Australian bowlers but how many remember the brilliance of Solkar to the Indian spinners ? If I am not mistaken most wickets fall under the "caught by" route. [[ dale, Pl see my common response. Ananth: ]]

  • arch on September 22, 2012, 20:38 GMT

    @Wapsting Regarding Sobers from 62-68 and Imran from 82-83. Looking at the figures I would pick Sobers just for his batting and Imran just for his bowling. Sobers would be the second best batsman in the side after Bradman, and Imran would either be the best or second best bowler, after Barnes, in the side. There is nothing to choose. There second suits are equally strong.

  • arch on September 22, 2012, 20:33 GMT

    I think Ananth, you will have to do a career length analyses and that will further illuminate your excellent work. As far as the difference between Imran's best and worse decades is concerned, those of us who have followed his career know that he was a middling medium pacer at the start of his career and only after watching Lillee bowl did he decide to bowl fast, and bowl fast he did, changing his action and his speed. Maybe that is why he does not come up more often in best bowler conversations since he never burst into the limelight as a force of nature, and then later on he extended his bowling career long past his peak. Many people who remember Marshall, McGrath and others remember people bowling at top speed for 12 years while too many people remember Imran either as a gangly youngster or as an old man captaining the two Ws. But I would argue that there is an even greater story here, the story that hard work, determination and a dint of genius, can allow you to reach the stars.

  • Som on September 22, 2012, 13:49 GMT

    Ananth - Having different periods for batsmen and bowlers is a good idea. But this brings into question the practice of measuring based on years as opposed to innings/matches. Does higher frequency over a short period result in better overall records because one can take care of good form or peak abilities or a longer spaced out career allow freshness and lesser fatigue. Would Cook or Mcgrath be as effective if they were born prior to 90's or in teams who have less number of games per year.

    While normalizing based on average career span, across eras, I guess this becomes paramount. There is much to credit the consistent high level returns of someone like SRT (you get the drift...) [[ If you fix, say 40 Tests, this could take between 4 and 15 years. So that is not a solution. If the period is too long all the problems will crop up. The player who plays for a long time to achieve this number of Tests is going to have to go through multiple stages of form loss/gain and it is not fair to these players. And the one who completes these Tests in 4 years has a much shorter period and advantage. I think years must remain as the unit of comparison. Ananth: ]]

  • Dinesh on September 22, 2012, 12:44 GMT

    Coming to Imran: The toast of the best decades Article.Let me punch a hole in his invincibility(pun intended) and on my own belief of eh being Solar systems ahead.

    From 1980-86 he took the Best bowler on the Planet -184 wickets in 35 matches at an average of 15.xx But what about the other 14 years? 53 Matches 178 wickets at an Average of 30. So in 6 years he took almost the same number of wickets as he took his other 14 years and he played 18 fewer matches. This is probably where Marshall and McGrath and Pollock edge him to a lower position.

    I am real happy that i gave a kick to the Shin and You in the heat of the Moment Came up with this. Looking forward to the whole article or atleast an excel. [[ No, this deserves a full length follow-up after I do some work on the career lengths. And the idea that we could get varying period sizes for batsmen and bowlers has just now percolated. Ananth: ]]

  • Dinesh on September 22, 2012, 12:36 GMT

    Ananth: Just by looking at the Stats The worst decade stats is probably the best follow up we can have for the Best decade as this is when we can differentiate people who were almost always irrespective of their worst years were very Good AS Som pointed out just look at Marshal He was truly Outstanding right through his carreer. And let me Point out the fallacy as well with both the Articles.

    Marshall and Mcgrath and Pollock had 12 years of Top flight Cricket and Imran had 22 years of Course some were lost to Injuries. So however we tweak the Stats there will be an Overlap for Marshall,Mcgrath and Pollock. I know you have given a valid reason for not bringing this 10 down to 5 years. but i fel 6-7 years is a reasonable number as All top flight Bowlers have 11-12 years of Carrer so we can have a division(i know there are lot of disadvantages as well). Contd [[ Dinesh, you are on the dot. Pl see my response to Waspsting who seems to have the same idea, a few minutes before you. Let me do an analysis on career lengths. Ananth: ]]

  • Waspsting on September 22, 2012, 12:19 GMT

    @Ananth - looking at your update on bowlers has got me thinking.

    since bowlers - especially pace men - have shorter careers - and primes - than batsman, maybe 10 years isn't the best figure to assess bowlers with?

    I think 6 years for bowlers - esp pacemen - would give a clearer picture.

    For the great batsmen, lets say we're assuming a 15 year career at least. (and we've chosen to use assessments based on 10 years)

    For a bowler, i'd assume a 10 year career at least (and the corresponding basis would be 2/3rds of that - or about 6 years)

    Long enough that great figures can't be put down to a "hot streak", but short enough to see some discrepancies between bowler at best and worst.

    What do you think? [[ Excellent point. But this will require some spadework to be done on the career lengths of players. I must get this right. You seem to be intuitively correct: Dinesh, sitting a few thousand kms from you seems to have the same idea) but let us get it right. Ananth: ]]

  • Waspsting on September 22, 2012, 12:11 GMT

    Shri - its good to have you join us.

    re: Morris - my point is conclusions based on 3 years leave scope for interpretations. "hot streak"? or "genuine indicator of how good a player is?"

    For 10 years, this isn't so (The point isn't about Morris per se - he's an example)

    ---

    "To my mind, Sobers from 1962 to 1968 is the greatest cricketer ever. Nobody comes close."

    You really think its that clear cut? (particularly if we can agree that bowling determines match outcomes more so than batting)

    figures for Sobers in that period and Imran for 82-88

    Sobers (62-68) 23 matches, 2162 runs @ 65.51, 89 wickets @ 27.29 Imran (82-88) 34 matches, 1479 runs @ 46.21, 167 wickets @ 16.89

  • Som on September 22, 2012, 10:57 GMT

    Ananth - I couldn't find a way to do that with one query in Statsguru. It needs some complex manipulations in excel after downloading disparate data through Statsguru. [[ One reason why I had mentioned that it was not possible through Statsguru. I take the deision to exclude the specific analysis at the lowest match level. Also this is a very specialized requirement. Ananth: ]] The worst decade is an eyeopener. Where longevity of career does not translate into aggregation (at similar rates, like SRT's), the variance between the best and the worst phases separates the best from the second best. And this is where Marshall beats Imran hands down. Wonder what of Lillee's career was so phenomenal that he tends to be chosen by many as a bowler superior to Ambrose, McGrath and Hadlee. Also wonder why Pollock who was one of the best ODI bowler during his career (and is one of the best ever), and a top class test bowler is so under rated. [[ I have given you guys only a sampling of the worst decade numbers. Already it has produced an excellent insight from you. Your comments on Marshall are perfect. The point is, if you select Marshall, it does not matter which decade he brings with him. You are assured of a great bowler. If Imran brought with him first decade, things would not look rosy. Tht means this analysis is not going away. At a later date what I should do is to determine both best and worst decade values and look at the variation. Then it is possible that the middle value is the one to be looked at seriously. And the almost continuous overlooking of Pollock bugs me. No one talks of him. But he is almost always there in the top-10 or so. Ananth: ]]

  • Gerry_the_Merry on September 22, 2012, 7:00 GMT

    Ananth, Imran at 29.62 is fine, but I remember he hardly played between 1971 and 1975. So perhaps just 2-3 tests in his first 5 years. So perhaps his career really is 85 tests between 1976 and 1992. [[ Yes, Imran captured only 128 wickets in these 10 Tests, but the rules are laid down as 10 years. Ananth: ]] By contrast, in the main article, Ponting, Kallis, Dravid etc. played between 160 and 175 innings in 10 years.

  • shrikanthk on September 22, 2012, 2:31 GMT

    Regarding Imran and Sobers: As a lot of people point out, Imran was seldom equally dominant with both bat and ball simultaneously. Moreover Sobers exhibited greater versatility with bat, ball and on the field.

    To my mind, Sobers from 1962 to 1968 is the greatest cricketer ever. Nobody comes close.

  • shrikanthk on September 22, 2012, 2:26 GMT

    The worst period for hobbs was from 1920- 1921 period where he averaged 50

    Interesting. Always fascinated with Hobbs. If you look at his first class career, the story is totally different. His best decade in FC was 1920-1930 I think, with his worst decade being 1905-1914.

    However in tests, it is exactly the opposite as Ananth points out!

    Goes to show that as he aged Hobbs matured like good wine and became extremely prolific against First-class attacks. However age did blunt his reflexes which meant that his effectiveness in international cricket lessened.

  • shrikanthk on September 22, 2012, 2:18 GMT

    e.g - Bradman praising Morris to the moon based on a 3 year, 15 match period when he scored heavily.

    Fair enough. But let's not forget that Morris lost 3-4 of his best years to the War. Also the second extenuating factor is that the pitches of the early 50s were notoriously bowler friendly. For eg: The 1950-51 Ashes series was probably played on some of the worst wickets used in an Ashes series. Not easy conditions especially for an opener.

  • Ananth on September 22, 2012, 1:47 GMT

    I have done the work on the worst decades for bowlers and have summarized the results below. A few surprises.

    Barnes S.F    17.92
    Marshall      20.31
    Ambrose       20.81
    Garner J      20.96
    McGrath G.D   21.53
    Davidson      21.67
    Pollock S.M   22.22
    Donald A.A    22.35
    Holding M.A   23.17
    Muralitharan  23.27
    These are the top-10 of the worst decades values. Ah! what happened to the best bowler, Imran Khan. Well he had a very ordinary start to his career and his figures for the first 10 years of his career, 1971-1981, puts him way down as given below. Very interesting indeed.
    Thomson J.R   29.54
    Imran Khan    29.62 (45th)
    McKenzie      29.82.
    
    The variations for Imran are upwards of 50%. For Marshall around 5%. Before you guys ask me, let me explain about Barnes. Barnes has only one workable decade which turns to be both his best and worst. Ananth

  • Som on September 21, 2012, 18:18 GMT

    Ananth - I see my mistake and your point. Thanks much. [[ Som, can you do something for me. See whether you can modify the query to see whether the criterion can be changed to "career wkts gte 25". I am not that much familiar with Statsguru. Thanks Ananth: ]]

  • arch on September 21, 2012, 17:47 GMT

    What about the worst decades for the top ten bowlers? [[ Will do and post by evening. Ananth: ]]

  • Som on September 21, 2012, 16:11 GMT

    Ananth - Not clear why we would require anything more complex than just adding all the runs that each bowler who qualifies has given while taking the wickets, and then divide those runs by the wickets taken. I did not use any sumproduct to arrive at my data. Runs are additive across innings, matches and players and so are wickets. I do not see any double counting when applying my approach. [[ Did I ever say there was double counting. I only said there was a different methodology. Pl re-read my comment. Anyhow there seems to be a problem in your query. The 25 wickets limit is not for the 10-year period. It is for the career haul of the bowler concerned. This is the only way the true part-timer is kept out. So the criteria should be a minimum of 25 career wickets. I am not sure how this can get into the query. According to your method the following players have been kept out of this specific period computation. Waqar Younis, Devon Malcolm, A Fraser, Bishop, Aaqib Javed et al. You could correct me if I am wrong. Ananth: ]] For eg, between 1990-2000, if bowler A gave 500 runs in 11 innings of 7 matches and took 28 wickets and if bowler B gave 600 runs in 12 innings of 6 matches and took 26 wickets, the combined average for these bowlers in that phase would be (500+600)/(28+26) = 20.37. And if these were the only 2 bowlers on planet at that time, the 'peer avg' would be 20.37 and the 'peer sans A' for A would be 600/26= 23.08. Am I missing something?

  • Waspsting on September 21, 2012, 15:25 GMT

    I'd conclude that 1976 was the meat of Viv's career, even of his best decade - and its absence takes away the meat from his worst decade.

    But he's up there with the best of them (in his best decade) even without the year (say he averaged 60 instead of 90 that year) - i'm happy to give him his due for best decade.

    (leaving aside that there's no justification in removing 1 year out of a total of 10 - unlike in the case of a career breakdown)

    But consistency - that doesn't seem to be Viv's forte, relative to his great contemporaries.

    At his best (as measured over a considerable time period), a cut above them thus somewhat justifying his generally being ranked above them.

    Not as consistent though, apparently.

  • Waspsting on September 21, 2012, 15:18 GMT

    table has three figures - # of calender years played, # of years average over 50 and # of years average under 40

    G. Chappell 14 10 1 Miandad 18 9 6 Border 17 7 5 Gavaskar 17 7 6 V. Richards 18 8 8

    For best decade , first i've been crude and just done 11 years (eg 76-86). The first figure is # of years ave of 60, second is years ave over 50 and final figure is years ave. under 30

    V. Richards 4-2-4 G. Chappell 2-5-1 A. Border 5-1-2 J. Miandad 4-2-4 S. Gavaskar 3-2-4

    What to make of consistency? Border and Chappell stand out, but Richards is right there with Miandad and Gavaskar. Gavaskar's lack of consistency might be understandable given he's an opener, Miandad's is harder to explain.

    Border looks the best of the bunch to me. 5 years ave over 60 and only 2 under 40

    Chappell only 2 over 60, but routinely over 50 (most over 50s of the lot), and fewest under 40s.

    Richards, Miandad and Gavaskar all in the same boat about.

    Yet Viv's best decade is well ahead of the others..

  • Waspsting on September 21, 2012, 15:07 GMT

    Richards: 42.20 (???) 1981-1991

    From perspective of Richards career - since his career average is 50 and his best decade about 60, it follows logically he'd have a pretty bad worst decade.

    From peer analysis perspective - since he's outstripped the others (Chappell, Miandad, Border, Gavaskar) on best decade, but his career ave. is about the same as them, it again follows that he'd have a pretty bad worst decade.

    A few articles ago, I noted that Viv's ave. drops from 50 to 45 with the removal of 1 year (out of 17), and also felt that removing 1 year out of a 15+ year career wasn't unreasonalbe.

    The year in question was so meaty that it seems to seriously effect even a 10 year analysis (though i don't think removing 1/10 of time frame is justified).

    Anyway, I was looking at the stats i'd dug up for that article - and did a crude replica for best decades

  • dinesh on September 21, 2012, 14:24 GMT

    Ananth:

    How can Jack Hobbs worst decade be better than his Overall Average. (64.64 - 56.94). Is this another Typo error? or am i missing anything.

    The worst period for hobbs was from 1920- 1921 period where he averaged 50. he dint play any matches for the next 3 years. and i tried all permuations and combinations but could not go Below this average. His worst Average even after including this time frame to Other years did not go Below 54. Or did i wrongly calculate and put in a Wrong Querry in Statsguru. [[ Hobbs' average between 1908 and 1921 was 56.04 which was adjusted upwards to 64.64. Start of career (Test 97) to Test no 152 when he had scored 2970 at 56.04. Ananth: ]]

  • Waspsting on September 21, 2012, 14:19 GMT

    "Lock??? when Laker was the real star in every way"

    supports your claim that the backwards done rankings are fishy. Laker peaked in his '56 series with 897 points, Lock reached 912 in '58 whilst taking 34 wickets @ average of 7 against NZ at home.

    Given Laker was averaging 10 in the same series (albeit with only half the wickets), given it was home conditions, and given NZ then make Bangladesh now look like Steve Waugh's Aus - hard to see how such a performance can take a guy to 6th highest ranking of All time.

    (fine bowler Lock - and good figures in the period leading up to and at his peak rating - but 6th highest of all time???)

    Maybe the rankings work like tennis - the better you do, the harder it is to keep up the points and easier it is to lose them?

    Still don't see how Lock could top Laker - let alone reach 6th highest rank ever.

    I'd noted earlier that for career [[ And ICC Rankings methodology is almost always closed book. Ananth: ]]

  • Som on September 21, 2012, 12:38 GMT

    Ananth - Not sure why your results do not match with mine, even though we have the same method. Thoughts? [[ I have not done seen the Statsguru query at all. Let me see. Maybe it is the slight difference because of methodology used. Som I am assuming that you have taken these 77 bowlers to an Excel sheet and totalled with and without Imran. Then that method is different to mine which adds the wickets and runs at the lowest innings level. We have had this before. Your method is the sumproduct values of 77 bowlers. Mine is the sumproduct of about 4500 values (361 tests x 4 inns x 4 bowlers x 75% est). I would dare to say that mine is the more accurate method since the summation is done at the lowest level. Ananth: ]]

  • Gerry_the_Merry on September 21, 2012, 12:25 GMT

    Ananth, not arbitrary at all! Tendulkar played Ambrose only in 1997. From then to retirement, it is 33/120 as mentioned by Nitin Gautam. I have not bothered to double check, trust his numbers, and from that derived 65/285.

    You can see from Sarosh's comments that is a cosy little party I have spoilt... [[ Maybe there is a party. But I have maintained a balance and ensured that this is not blown out of proportion. The Career level Peer analysis, in my opinion, is far more important in the overall context of things. Ananth: ]]

  • Ananth on September 21, 2012, 12:18 GMT

    I have given below the revised top-10 table of bowlers. This has been done using the gte 25 wkts bowlers only. Imran Khan 17.88 Barnes S.F 17.92 Muralitharan 18.62 Hadlee R.J 18.64 Marshall 18.75 McGrath G.D 19.56 Laker J.C 19.95 Ambrose 19.96 Garner J 20.20 Wasim Akram 20.78 No great changes. Ananth

  • Ananth on September 21, 2012, 11:52 GMT

    Dinesh: you were on the dot and I was on the semidot. The worst decades are as follows. Hobbs: 64.64 Bradman: 90.23 Sobers: 55.25 Gavaskar: 49.29 Richards: 42.20 (???) 1981-1991 Tendulkar: 52.44 Lara: 49.69 Thanks for the kick on the shin. Ananth

  • Waspsting on September 21, 2012, 11:26 GMT

    was looking at ICC highest ever ratings, to see how it compares to this analysis.

    For both batsman and bowlers, the top 10 has 6 common amongst the two lists.

    The most converging part is the top 5 of the bowlers. Here it is -

    decades analysis

    1 Imran 2 Barnes 3 Murali 4 Hadlee 5 McGrath 6 Ambrose

    ICC highest (excluding Lohamann)

    1 Barnes 2 Imran 3 Murali 4 McGrath 5 Ambrose & Lock

    (Hadlee's joint 10 on the ICC list. bit of a surprise to see Lock reach an all time high of 5th position on that list - he's 19th on decades list) [[ Lock??? when Laker was the real star in every way. Thanks for this. I had not looked at it. Ananth: ]]

  • Waspsting on September 21, 2012, 11:14 GMT

    - a slight flaw would be that we're comparing all players for each's best 10 years. I'd rather use standard decades - each player might have some of their best years or not included in them. As it stands, its awkward to compare say - Richards 76-86 to Border 83-93 (Ananth's got this base covered with peer analysis stuff he's planning) ---- An obvious question this analysis begs is, "well, what about the rest of their career?"

    Possibly a "worst decade" analysis is a good complement because you wouldn't expect too much of a discrepancy between worst and best over such a long period(and some overlap of years even). [[ Yes I agree. I jumped the gun and shot my foot. I am now certain that the gap will be much less. Ananth: ]] @Ananth - a suggestion, if i may? in a peer analysis, you might consider a column for peer comparison with just team-mates for given time period (as opposed to world cricket), because that's the basis that a guy makes it into his team.

    e.g Tendulkar now or Richards at retirement... not top class in the world, but good enough for a place in their side. [[ For the peer analysis I will; ask for suggestions because I think it is going to be a path-breaker, kick-started by this one. Your current suggestion is a tough one. Nitin or someone else asked for a country level comparison and I had explained the data problems. Maybe as a separate component of the peer analysis. Ananth: ]]

  • Waspsting on September 21, 2012, 10:59 GMT

    re: this analysis generally (most points i make have been made already, by Ananth and Sarosh in particular) -

    I see the strenght in its being "sufficiently long". We can say of a players golden stats for 3 years that its "just a hot streak", but its unlikely anyone could have a "hot streak" of 10 years.

    e.g - Bradman praising Morris to the moon based on a 3 year, 15 match period when he scored heavily.

    Sarosh's point about Richards is a solid example of the value of the analysis. if he outperformed the rest of the world - many fine batsman around at the time then too - for 10 years to the extent indicated, it at least gives a solid reason for how highly he's rated

    - We need a basis for what we mean when we refer to a player. When we speak of "Sobers" we're not thinking of the debutant who batted at 9 (Marshall's early career is another example of this).\

    10 years seems a good minimum (if arbitrarily chosen figure) for establishing a players quality (cont)

  • Sarosh on September 21, 2012, 10:31 GMT

    Nitin, You are perhaps relatively new to this blog, which is why you probably responded to G_t_m. The rest of us don't bother. His logic is simple - When SRT was playing well, all the bowlers he faced were either sub-par or not at their peaks. Or otherwise it's not that SRT crashed past 2003; it is that overnight suddenly all bowlers started bowling very well- Simple! [[ No, Nitin has been long enough here and he is one who normally respons in a very reasonable and fair manner. Ananth: ]] Ananth, This gives me an idea for an analysis. We know that "averages" in matches "including" so and so bowler are very vague. How about an "innings success ratio" vs. top bowlers. For eg. Suppose we consider Richards against Lillee. With the traditional use of stats Richards “average” against Lille takes a hit even if another bowler gets him for a Duck. Not sensible. Suppose Richards played 20 inn. against lillee; And he averaged say 50 or more in 8 inn. In 5 inn. he got out to Lillee .In the remaining 15 inn. he got out to a another bowler. So, in 20 inn. against Lillee he "lost out" 5 times. Got the better of him 8 times. Things were “Neutral” (got out to a 3rd bowler) 15 times. His “hit” to “flop” ratio should be 8/5 = 1.6. This will eliminate the problem of Huge scores against a particular bowler against a string of low scores giving a false impression , the problem of N.Os , the inaccuracy of averages in matches “including” X bowler etc. [[ Sarosh Let me look into it. I have cut and pasted into my to-do bucket. Ananth: ]]

  • Gerry_the_Merry on September 21, 2012, 9:38 GMT

    Ananth, 33 tests, 120 wickets is no way comparable to 65 tests, 285 wickets (total 98 tests 405 wickets).

    A bowler who takes 8/45, 7/25, 6/24 in the space of 4 years is very different from 33/120.

    Averages tell only part of the story. [[ Gerry You are doing the exact thing you are accusing others of. You have split Ambrose's career in an arbitrary manner, derided 120 wkts in 33 Tests by a bowler who is one of the 3/4 star bowlers, not accepting averages and so on. It is natural that Ambrose's skills would have waned in the third segment of his career. But that should not be used as a stick to lower the quality of bowling around that time. I suggest that there be no response to this line of thinking. Next someone would take a part of McGrath career in which he is down and then pull Lara down using that segment. Not Cricket as I know it. Ananth: ]]

  • Som on September 21, 2012, 9:16 GMT

    Ananth - Here it is: http://stats.espncricinfo.com/ci/engine/stats/index.html?class=1;filter=advanced;orderby=wickets;qualmin1=25;qualval1=wickets;size=100;spanmax1=21+nov+1989;spanmin1=21+Nov+1979;spanval1=span;template=results;type=bowling

    I then took all this data into an excel and calculated the 'bowling average' as 'total runs'/'total wickets'. And included/excluded Imran's figures from the equation as required for 'peer' and 'peer sans'.

    Why I asked about the last test is because in case of Mcgrath the test which bookends his best 10 yr phase actually starts 1 day after 10 yrs (on 21st Nov 2005, when his 10 yr started with a test whose day 1 was 21st Nov 1995). So there is both a potential overstretching the phase and also incorporating other peer's data who may have been playing a test which either started the same day or did not end till that day. But this might be very minor in terms of effect.

    In case of SF Barnes and others with less than a decade this becomes tricky.

  • Nitin Gautam on September 21, 2012, 8:54 GMT

    @gerry Nitin Now I doubt if you even watched that series. Ambrose frequently did not even take the new ball.

    I agree he had not taken new ball sometime but how does that matter. Donald, akram, Mcgraw sometimes were not been given new ball, does that make them any lesser, even steyn now (recently eng series) was not given new ball regularly. This could have been a strategy by captain. Re. comparison with holding..what should be remembered is that he operated with same menacing accuracy albeit with slightly lesser pace still managed to outdo himself if compared with pre 1997(strictly considering his avg/wicket). & no..had it been Richards/Lara..I definitely would have written. this analysis was done to find out who performed to what extent if best decade in taken & SRT came 5th period..If you go through my previous posts, I have asked Anantha to have avg BQI faced by each batsman in their best decade..that would have put things in more perspective.but probably thts too difficult to do.

  • Som on September 21, 2012, 8:14 GMT

    Ananth / Pavan - Thanks for the clarification. The method I am using is exactly as you mentioned, but my data comes from Statsguru. And in Imran's case, instead of 30.89 (sans peer avg), it comes to 30.34 for GTE 25 bowlers who bowled during the phase Imran had his best decade. Is there a reason why Statsguru may not have all the right data?

    Also Ananth, is the data for a player's best decade include the end test(and the corresponding time frame), or exclude that test. [[ Should not really matter. Without looking at the program code, I would say probably includes. Ananth: ]] Thanks ! [[ Through Statsguru?. Can you please send the URL for this query. Ananth: ]]

  • Jon on September 21, 2012, 7:46 GMT

    Thanks Ananth, this is a really fascinating analysis. It's something I've often wondered about considering conditions were so different in different periods. One interesting thing is it doesn't take account of the expaning amount of cricket played, and the expanding number of opponents. If I remember right Bradman played the vast majority of his tests against England in either England or Australia, whereas modern cricketers have to adjust to a variety of different conditions and opponents. I can't suggest a decent statistical way of adjusting for this though.

    I also wonder if there's a way to do this type of analysis for all-rounders and for wicketkeepers?

  • Dinesh on September 21, 2012, 7:28 GMT

    Ananth: Re: "If I had done a "worst decade" for all batsmen, Lara's might very well have been 49 while SRT/GS's might have been 42."

    I just went through the Stats of SRT for his worst decade. It wasnt 42 as you said but 52 if we consider Statsguru with the normal process of calculating Avg. What i did with those Stats was take all innings above 30 as a whole innings and out of his 17 Notout's 12 were above 30. So his average then Came down to 47. The reason SRT worst decade AVG cant go below this is however we use to incorporate his 2003-2008 Span then we have to add 5 years after this or before this Span which were very very productive. So His worst Decade was from Sep 2002- Sep 2012 and i took this period because this will exclude his 2002 England Series and include the last New Zealand Series.

    But as you said No players Stats can be pulled down coz he dint play XYZ.In that way Lara dint play Ambrose or Ponting dint play at all WI Quicks/Waqar/Wasim.And No way BCL/RTP are inferior. [[ You are correct, Dinesh. I was probably way off mark. I did some rough calculations and said "if SRt was 7 above his average in his best decade he would be 7 off the average on his worst decade". Even then I meant 48 and wrote 42. And the decades are not mutually exclusive. And, to boot, Tendulkar's worst years were 1989-1992 when he averaged only 35. So there is no way any decade of Tendulkar can be below 50. Anyhow a decade average of 42 is way, way too low to contemplate for players with 55 averages. For players with sub-50 averages, that too with long careers, probably yes. I will anyhow do the worst decade just to get the numbers right. Ananth: ]]

  • Gerry_the_Merry on September 21, 2012, 7:24 GMT

    Nitin Now I doubt if you even watched that series. Ambrose frequently did not even take the new ball. 33 tests and 120 wickets says it all. Yes he was a formidable bowler, but in his 1997-2000 avatar, would I put him alongside the 1980 Holding? No way.

    I watched almost every series of Ambrose starting from England in 1988. if you get footage, compare him in 1988-1995 and 1997-2000. In the former period, he was in my opinion, the greatest fast bowler of them all, of all time, frequently single-handedly dismantling high quality batting line ups in away tests.

    My objection is on reaching a breezy conclusion that since xyz reached his best numbers in a 10 year span where a/b/c played (and they were great bowlers), ergo xyz must have been a great batsman. [[ Conclusions are reached by raeders. I am only saying that SRT had a decade of 63 average. Sobers had a decade of 72 average.That is all. Ananth: ]] I am sure my reference to Tendulkar was the only reason for your response. Had I said Richards, my remark would not have triggered any counter comment. [[ Maybe not from Nitin but from somebody else. The point is not to pull down valid numbers using subjective conclusions. Re Ambrose, Career midpoint 1995, 218 at 20.87. Second half 187 at 21.12. Around 1% variation. Picked up from my Test midpoint analysis. So where did you get the notion that he went downhill after midpoint. Ananth: ]]

  • Nitin Gautam on September 21, 2012, 6:40 GMT

    There is no need to decry greatness: be it Bradman's, Tendulkar's, Lara's or Kallis'.

    ABSOLUTELY agree. This was an analysis that showed the performance over an extended period of time (decade) & it should be taken as such. eroding greatness comes from nitpicking & finding faults instead of accepting solid facts. someone here said GS in his best decade thrived on home conditions & most of great bowlers didnt have their best decade coinciding with GS's. Does that conclude GS was not as gud as say any player below him. CERTAINLY not. In his best 10 years he is 2nd best ever batsman behind only incomparable Don. absolutely no need to decry greatness by pointing pitch, bowlers or anything else.

  • fahad on September 21, 2012, 6:03 GMT

    Wonderful Ananth ! Many many surprises...the biggest for me is that the best Pakistani spinner turns out to be Iqbal Qasim! Well I just realized he has the best career average as well but who would have thought, given how underrated he is even in Pakistan. In fact he is ranked third from the subcontinent after murali and bedi. [[ Even based on raw data 170 wickets at 28.11 for a spinner represents top quality. And, unlike Indian pitches, the Pakistani pitches were not the spin-friendly ones because of the proliferation of [pace bowlers over the past 40 years. Iqbal Qasim is under-rated and deserves recognition. He (and Tauseef) were the ones who stood between Gavaskar and victory in the later's farewell Test. Ananth: ]]

  • Nitin Gautam on September 21, 2012, 5:17 GMT

    @Gerry the merry In the specific case of Tendulkar, 1) he did not play Curtly Ambrose until almost at the end of Ambrose's career, when he was an economical and intelligent bowler, but not a terror.

    Dont understand the basis of yours this comment. SRT played Amrbose on his own turf in 1997 & since then his overall record until his retirement in 2000 speaks 33 match 120 wickets @ 19.99..even lower than his career average.however slower he was in comparison to his peak, it was way too difficult to make runs against him & yet it was 1997 only. & if speed is the only thing you count for a bowler to be menacing, how would you compare shoib, lee with mcgrawth. Is it so difficult to accept that SRT avgd 5th best in a decade where out of top 30 bowlers, 10 have their best performances. As Anantha right said he played an extraordinary set of bowlers overall & produced very gud results. Its a fact..I dont think there is need to split hairs on this. [[ There is need for a sense of fairplay here. It did not bother me that Lara was way down on the list. It showed that he was more consistent than many other batsman. If I had done a "worst decade" for all batsmen, Lara's might very well have been 49 while SRT/GS's might have been 42. But again no need to draw any unnecessary inference from these. There is no need to decry greatness: be it Bradman's, Tendulkar's, Lara's or Kallis'. Ananth: ]]

  • Pavan Kumar on September 21, 2012, 4:45 GMT

    Ananth,

    For calculation of Imran's figures in the example, numerator was 32.2 and denominator is 32 which makes for upward adjustment rather than the dopwnward adjustment shown in table. Calculation is 18.37*32/32.2=18.26. So making calculations for bowlers with career wickets lt 25, Imran's figures would now be 18.37*30.09/30.89=17.89 [[ Yes, this was a calculation I had done using the calculator and not the program based work. The 32 is equivalent to 30.09 and has to be in the numerator to reward tough bowler periods and penalize bowling-centric periods. So the revised adjusted average will be 17.89. Mainly because the period average had a greater variance to the all-time single figure. Many thanks, Pavan. Som, please note. Ananth: ]]

  • Uday on September 21, 2012, 4:41 GMT

    @Wasp/Ananth/Boll - Imran often doesnt get his due because played in one of the most competitive eras for fast bowling, alongside Lillee, Marshall, Holding, Hadlee etc., and for all rounders, alongside Botham, Hadlee and Kapil, and therefore he tends to be lumped in with all of them together when great players are discussed. Moreover, towards the end of his career he was outshone by other pakistani fast bowlers, and this perhaps adds to him not being as highly regarded as he should. Sobers, on the other hand, was head and shoulders above all his contemporaries for most of his career.

    Imran is also very well known for his leadership role, and i think this tends to deflect attention away from his greatness as a player (not that it should of course!)

  • Sarosh on September 21, 2012, 4:36 GMT

    Boll, Again ,what you are effectively stressing is the isolation at the top factor. An era with Borg,Mcenroe,Connors sharing the spoils and also the No.1 ranking will raise more debates about who was the best as compared to the Sampras era where it may be said he had to contend with less all-time greats. Of course, this may be attributed to Sampras's dominance itself. But you get the drift...

    A sportsman who seems to be a cut above his contemporaries always seems to be better than players of another era. After all, peer comparison is the only valid comparison there is. All else is conjecture. [[ Agreed, Sarosh. I am going to do that peer compariosn across the career of a player: the only period which is significant, fixed and irrefutable. Ananth: ]]

  • arch on September 21, 2012, 4:29 GMT

    I would love simulations of the first and second XI's - you can do several runs - one with career averages and one with peak averages. I think you might have different results. [[ Yes, the peak averages can certainly be used in these fantasy-based scenarios. Lot of work and has to be planned right. Ananth: ]] If I remember right, you did a series analyses a long while back in which you graded all-round performances for players in a series, but the all rounders had upper and lower cutoffs for their batting and bowling which I think disproportionately affected lower order bats. I think if you were to do that analyses without upper and lower cutoffs you would get figures on how much each player contributed to the team. I hope I am being clear here, this analyses was a long time ago but I remember it was very very good but felt a bit unfinished. @delmesiter Comparing Botham's bowling to Imran's in the 92 world cup is a bit unfair as Imran was both older and injured, playing on painkillers. Imran was the best bowler in the 1987 World Cup though. I think Botham was injured for that one, I think.

  • Sarosh on September 21, 2012, 4:28 GMT

    Incidentally,when accounting for N.Os ,Sobers benefits the most. From avg. 73 to 61 ! With 12 N.Os out of 78 inn. Kallis 66 to 54. Ponting 63 to 53 . Dravid 60 to 52. Tendulkar 62 to 56 . Lara 54 to 52 and Richards 58 to 55. Guess some sort of in number between Avg. and RPI would be closer to the "Truth". [[ Sarosh I have an in-between measure. I have used in some of my Ratings work, where I have the license to try out something. an alternate method. I consider a not out as a genuine not out only if it is below a certain tally of runs. In otehr words give the benefit of not out to the batsman only if the score is, say below 20. 13 not out is a genuine not out. 45 not out is treated as an innings. It has worked very well. It is in between Avge and RpI. Lara's 6 not outs are 400*, 153*, 80*, 48*, 14* and 13*. Allow only the last two innings as Not outs. In the other four he has had the chance to play a longish innings. It will work very well. Ananth: ]]

  • Gerry_the_Merry on September 21, 2012, 4:20 GMT

    Hate to spoil the party, but want to make a general comment. Some previous articles have been very sharp in isolating pitch quality / bowling quality etc. to determine the true value of a batsman's average.

    THis article focusses on a specific aspect - how good were a batsman's best years. To take this article to infer that Tendulkar is the best is as usual, an optimistic and erroneous conclusion.

    It is generally true that in such inferences, data points are selectively chosen, or blatantly twisted to support breezy conclusions.

    In the specific case of Tendulkar, 1) he did not play Curtly Ambrose until almost at the end of Ambrose's career, when he was an economical and intelligent bowler, but not a terror, nor backed by a pack of torrid quicks. 2) His figures against Wasim/Waqar/Donald/Pollock/McGrath, the best of those times, are decent, but not exceptional.

    How did he generate 63.xx then? And to say the his best came coincided with best pace attacks...Same for other batsmen. [[ Gerry, let me say this. Tendulkar played a very good set of bowlers during 1992-2002 and averaged 63. That is fact. As a 10-year period goes, his performance is right there, in the top-10. It would be churlish and nit-picking to dilute this effort. Ananth: ]]

  • Sarosh on September 21, 2012, 4:17 GMT

    Waspsting, You may have a point about bowler opinions by batsmen and vice versa. For eg. I read an article as to how Nadal keeps prasing Federer to the skies as the greatest player ever and then proceeds to continually whup him! So- the thinking naturally goes, That if Nadal keeps beating the best player of all time then...

  • Sarosh on September 21, 2012, 4:14 GMT

    Ananth, disagree on several of your points: “I am also aware that this analysis, as it stands now, has found broad acceptance because it shows Tendulkar in excellent light. If the 8 or 12 year analysis pushes him to, say, 12th in the table, it will not have the same level of acceptance.” I have taken a rough look at SRTs best 8, 10, 12 yrs spans. From 30 Nov 1990 – 30 Nov 2002 (12 yrs), 30 Nov 1992-2000 (8yrs) – SRT averages 60+. He will top his direct contemporaries in all 3 cases. Finally, that is what matters. Players prior to him or after him may put up better numbers. “that will be THE analysis and will not have any arbitraryness thrown in. In almost all cases the player decides on his career period. It is his choice if he plays for 9 years, 16 years or 24 years.”

    Again, absolutely wrong. Do we pick say our India all time XI‘s based on SRT, Dravid or VVSs performances over the last year or so? Or even Dravid’s very poor 4 years before his last England tour? No way. We pick Dravid or VVS based on their performances at their “peaks”. Provided, of course, this is of sufficient length. Dravid 1996-2006. VVS in the 2000s. The overall career average which includes injuries, age decline etc does NOT tell us about a player’s “True” caliber. This sort of analysis does. Federer may go on for another 9 years ala Connors and suffer steadily worse head-to-heads vs. all top players. This should NOT and will not affect his legacy because of his prior achievements which are “locked in”. In cricket, by contrast, continued non performance can and does drive down a player’s average.

    The idea for this sort of analysis was based on the hammering legends such as Richards and SRT were getting. People were wondering how they always command a place in the All time World XIs based on career averages of 50 something, when several batsmen including their contemporaries averaged more. This sort of analysis helps to explain why. A career average does not. Richards wanted to play on. He was literally forced out. He would have surely dented his stats more and more had he stuck on. Does anyone think that people who saw him for his decade long glory peak use his “final” career average to rank him? Lara avg. 40 in his last year. He may have gone on if up to him and may have ended up with a sub 50 career avg. Would this, or should this ,affect his standing in the game? What if Hussey had retired while averaging 80 or 90+? Would a well timed retirement have guaranteed him a place among the greats? A somewhat similar thing seems to be the case with Pollock - entirely different circumstances of course. A player may decide on his retirement, but this certainly does not mean that he will choose the correct date as far as flattering his stats is concerned. That information is easily deduced only in hindsight. The career analysis you plan is great- but this analysis is far more informative in telling us why certain players strike such a chord in spectators and seasoned observers alike who have seen them at their peaks for sustained periods of time. If a player could sustain a certain level of performance for around a decade – then clearly it is circumstances such as age, injury or ill-timed stretching out of careers which are the culprits and not his intrinsic cricketing ability. [[ No problems with all the points you have raised. After all you opened the door on a 10-year period and I took it to completion, probably putting in 95% of the work involved. I know the value of this analysis. This analysis is very good with regard to determining a player's peak period. And possibly for determining all-time XIs and so on. And comparing players on the player's peak basis. However I maintain that the true measure of a players's value is an analysis of his complete career. It does not matter a tinker's damn if Richards and Lara were forced out and Tendulkar is/Gavaskar was allowed to choose his exit point. No one chooses a retirement date to preserve a 50.xx average. He would always think that he would play better that 50.xx figure. If Tendulkar seriously does not think he can average 50+ in the next 10 Tests, he should not play. It is the career which is the single solid edifice which defines everyone's value to the team. No one can question this broad-based statement. If a player played for 20 years and had a 10-year period averaging 65 and a 10-year period averaging 35, you could select him in an all-time planet XI citing his 65. But he will be at par in terms of his overall career contribution with some one who averaged 50 and 50 in his two 10-year periods. The peak value of 50 would not get the later player into the all-time planet third eleven. These are hypothetical examples. But it is true that a high peak will necessarily be compensated by a low trough. Federer's career cannot be looked in isolation on the glorious 2006-2008 period: the poorer 2009-12 period will also have to come in. In other words he can be truly evaluated only on the 2003-201x period. If Hussey retired when he had scored 2500 runs at 70, only a fool would place him amongst the greats. The cut-offs take care of that. I am so strict on this that most of my analysis exclude Headley and Pollock. I even raise the bar to 4000 runs if I think it is warranted. And let me say this: if a player scored 4000 at 65, he better be a serious candidate for the second spot. Ananth: ]]

  • Som on September 20, 2012, 19:46 GMT

    Ananth - For the 11 bowlers on the top of the list, I calculated the averages (peer and sans peer) by taking all bowlers who qualify (gt25) and their contribution during the 10 yr period. So in case of Imran, those values came to 29.88(peer) and 30.34(sans peers). I used them instead of 31.74 and 32.2. Now in the AdjAv calculation, I used the 'sans peer' data and the 'overall avg', but did not use the 29.88 value. Was the 'peer avg' just for illustration or was it used in the normalization. Thanks ! [[ I pride myself on the quality of my communication. Somewhere there I seemed to have slipped up. Let me take only Imran's case. The last two figures are very relevant. These are the Peer bowling averages between Test # 861 and Test# 1127. The 31.74 is the Bowling aveage for ALL bowlers and 32.2 is the Bowling average for allbowlers sans Imran. But this is based on consideration of all bowlers, including those 1 for 281 type of bowlers. The calculation is 18.37*32.0/32.2 = 18.26. So far so good. Now the numbers 31.74 and 32.2 have to be recalculated based on consideration of only gte 25 wicket bowlers. The only place this can be done is within my database and I have not done this until now. This has to be done match-by-match and analysis-by-analysis. Exclude the analysis for bowlers with career wickets lt 25. No other way to do this and Statsguru cannot be used. For Imran Khan I have calculated this manually. The figures are 30.46 (with imran) and 30.89 (without Imran). So the recalculation is 18.37*30.09/30.89=17.89 (Many thanks, Pavan). I don't think I can be clearer. Ananth: ]]

  • delmeister on September 20, 2012, 18:34 GMT

    Very quick comments before coming back to this later...photo with Imran is 1982 series in England, Lamb the non-striker. Very, very interesting comments indeed by Wasp on 'who was yr toughest opponent'. Finally, before people continue to run Botham down on his overall bowling figutes, please remember this. The Perth test of 1986/7 Ashes was the last he bowled a ball off his full run. He tore an intercostal muscle in that test and was a military medium pacer for the rest of his test career, which ended 1992, whereas even the struggling Kapil Dev was medium fast when he was chasing wickets record, and everybody takes this into account.But I will also concede that Botham wasted his talents (largely) between 1982-86, exactly the period of Imran's allround peak.Mind you,botham's medium pace in 1992 world cup was far superior to Imran's.Interesting thought! Wonderful analysis again tho Ananth

  • dale on September 20, 2012, 18:17 GMT

    Continued: Imran did not perform at an high rate of efficiency in both disciplines at the same time i.e in the same match.He is well below Sobers, Miller ,Botham and Cairns in this measurement. Both magnificent players who deserve all the accolades bestowed on them - this is all subject to our whims and fancies but as illustrated by this analysis Imran can easily be selected as one of the very best fast bowlers while Sobers is one of the very best batsmen. Sobers dominated cricket for a very long time - batting, fielding and good enough as a bowler to finish his career with 235 wickets for seventh spot on the all time list.His bowling should be viewed in context to his batting- batting for long periods of time plus long spells of bowling plus fielding brilliantly in close catching positions. The hype is also that he is by far one of the most popular cricketers of all time.

  • dale on September 20, 2012, 17:33 GMT

    Sobers would not make an all time xi as one of the top four bowlers and Imran should not make an all time xi as one of the top six batsmen. Sobers did as much bowling (23 overs per inning) for WI as Imran did for Pakistan but they performed different roles. Imran a magnificent strike bowler ,Sobers an accurate and versatile stock bowler. Sobers was the premier batsman for the WI despite the presence of other outstanding batsmen.But he outperformed them all in addition to the heavy workload as a main bowler. Imran, did not feature prominently in Pakistan's fortunes in his role as a batsman. Almost 78% of his runs were scored from the #7 thru #10 batting positions. He scored less than 1,000 runs and averaged 36 in Pakistan victories. This is not a reflection on his abilities as a batsman but rather an acceptance that his batting was not by any measure nearly as important as his bowling was to Pakistan.

  • cricket-india on September 20, 2012, 17:05 GMT

    ananth, how about an all-time best XI from you for each test country, based on your statistical analysis? let's see how differnt it would be from cricinfo's, and let the debates start:-) [[ It would take too long since I normally do only two articles a month. Ananth: ]]

  • Dinesh on September 20, 2012, 13:12 GMT

    Ananth: Many a comments have come stating that He hasnt faced a certain bowler, so he isnt great.

    I think we can have an analysis here so that we can come clean on it and people wont say a damn word again on a Batsman's greatness. People always say that Viv dint face his own bowlers so he isnt that great compared to gavaskar. I think if we can remove the Average of tbe West Indian bowlers from the analysis when calculatiing the Adjusted Average for the Period when Viv played and See how Viv fared Along with Others when he faced teh same Set of bowlers. In this way we can do for other batsmen as well when calculating the Avg for SRT we can remove the bowlers of India from the analysis and see how he along with others Fared when facing the rest of the Batsman. I dont know how successful this kind of article will be and how huge and tough it will be calculate.but can you give it a try or tweak accordingly to come out with a decent article. Just a request and not demanding. [[ Now I have the data to get the period figures between any two Tests by the simple expedient of determining and storing the 1877-to-Test figures for each Test. I then find the diffrences and determine averages for a period between any two Tests.. I cannot really include the country level figures for each Test. It is just not worth the effort. Too much work, too little gain. Anyhow I have an excellent measure for each batsman which is the Average Bowling Quality faced by a batsman based on the weighted CTD figures. What more is needed. If Richards really faced average bowling, this number will reveal it. Incidentally he faced pretty good quality bowling. Given below are the figures. Richards 32.66 Chappell G 32.29 Dravid 34.12 Border 32.79 Gooch 30.54 Ananth: ]]

  • Boll on September 20, 2012, 12:53 GMT

    @Wasp/Ananth. I think what people tend to forget about Sobers is that he was quite clearly the pre-eminent world batsman for the better part of 15 years. Based on the (retrospective) ICC rankings, in the 16 year period between 1958 and 1973, he finished the year ranked No.1 12 times, twice he was ranked 2nd, twice he was ranked 3rd - an astonishing period of dominance.

    If we compare this to modern greats, I believe Lara, Tendulkar, Ponting and Sanga have each achieved a No.1 end-of-year rank on 3 occasions.

    To add to this, between 1962 and 1973 he was only once ranked outside the Top 10 bowlers in the world.

    While I would agree that reversing the batting/bowling stats may give us more of a matchwinner, I`m not sure it`s all hype. [[ As a player, probably Yes. As a batsman, probably No. Ananth: ]]

  • Som on September 20, 2012, 12:41 GMT

    Ananth - They are actually much lower than before (as in the other approach in your sheet). Pls find below with percentage difference alongside. Am I missing something?

    Players NewPrdPeerAvg(>25) Peer Avg Bow % Diff Imran Khan 29.88 31.74 5.9% Barnes S.F 24.48 26.73 8.4% Muralitharan 30.57 33.02 7.4% Hadlee R.J 29.68 31.66 6.3% Marshall 30.17 32.29 6.6% McGrath G.D 30.23 32.69 7.5% Ambrose 28.84 31.33 7.9% O'Reilly 27.59 33.01 16.4% Laker J.C 26.46 28.99 8.7% Garner J 29.39 31.35 6.3% Wasim Akram 29.65 32.12 7.7% [[ Yes, Som, you are missing one thing. The last two columns in the two average tables are as yet not known on the basis of taking only gte 25 wickets. The figures 31.74 and 32.20 for Imran cannot be used. These values are inclusive of all 1590 bowlers. Ananth: ]]

  • Waspsting on September 20, 2012, 11:54 GMT

    I dare say Sobers' bowling stats aren't on the same platform as Imran's 50 average over 10 years... but Sobers was at the forefront of bowling, while Imran wasn't of batting (?)

    Sobers was opening the bowling and bowling spin on days 4 and 5 vs Imran batting at 6-7, shepparding the tail and remaining not out often... its a difficult comparison, but I'm ok giving Sobers an edge on this.

    Imran was an atrocious fielder, too. if your including captaincy though, he gets points on that.

    That said, I'd always value a bowling all rounder ahead of a batting one. [[ Yes, I agree. More of a match-winner probably. Ananth: ]] Simple logic tells you that a bowler has a greater weight than a batsman in test cricket. So a great bowler/good batsman (Imran) is more valuable than a great batsman/good bowler (Sobers).

    that noted, Sobers gets hype while Imran doesn't. Assessments like Ian Chappell's that Sobers was as far ahead of the next all rounder as Bradman was the next batsman.. are over-done, to put it mildly [[ Ian is prone to "hype" often. Sobers, Weekes, Walcott, Worrell, kanhai is not a list with those sort of daylights. Ananth: ]]

  • Som on September 20, 2012, 11:54 GMT

    Ananth - Here it is:

    Players NewPrdPeerAvg Imran Khan 29.88 Barnes S.F 24.48 Muralitharan 30.57 Hadlee R.J 29.68 Marshall 30.17 McGrath G.D 30.23 Ambrose 28.84 O'Reilly 27.59 Laker J.C 26.46 Garner J 29.39 Wasim Akram 29.65

    Does the 'Period Peer Avg' of 28.84 during Ambrose's best years (90-00') indicate among other things, that the batsmen's approach to test cricket and scoring got defensive? Is that indicative of a phase where many great batsmen of the 70's-80's retired? Could SRT's dominance during much of this era be attributed to just superlative performances or both superlative performances and an non-defensive approach towards run making in tests? [[ No, Som these are not correct. The Peer values would be completely different if the same "only gt 25w" approach is used. They would all be lower. Ananth: ]]

  • Waspsting on September 20, 2012, 11:47 GMT

    naming the guy whose got the better of you for the most part - that makes the opinion giver look like he's failed.

    There are exceptions who answer along the lines of what the scoreboards indicate, but generally, this is what i've found.

    What do you guys think?

    ---------

    re: all rounders - Miller I agree is up there with Imran, and Mike Procter had the potential to join them in that class.

    Hadlee's batting wasn't in those guys league, and Botham, Kapil fall short on bowling. Pollock isn't far short of the Imran/Miller league.

    as for Sobers and Imran... that's a tough one to assess.

    For a not very brief period, Sobers was one of the best bowlers (top 3-6) in the world, while he was almost undisputed top batsman in the world.

    Imran, though statistically at the top, probably isn't as "undisputed" the top bowler over the likes of Hadlee and Marshall as Sobers was for the batsman. [[ WS, please read my detailed response to Unni. What this article conveys is only that "during his best period" Imran was right there at the top. But this is a selecrive analysis. Only when I complete the complete Career analysis will we see who was on top across their career which is what really matters. Ananth: ]] On Imran's batting vs Sobers bowling...

    (cont)

  • Boll on September 20, 2012, 11:40 GMT

    @Vivek, I presume you`ve been focusing on the `runs per decade` chart if you`re mentioning the similarity between the records of Strauss and Pietersen.

    It`s an interesting aside, but the clear bias towards top-order modern cricketers who played a lot of tests limits its validity as any real comparison of batting performance.

    If we look at their overall test batting records, Strauss has been a very solid opening bat, KP has been a superstar - probably all-time England No.4 (not to mention his exploits in the shorter forms of the game).

    In terms of overall contribution, there`s not much between them. Strauss` captaincy and general management probably gives him the edge - a well-liked and trusted leader who took his team to the pinnacle and probably left at the right time. KP has been a divisive figure, but whatever we may think of him, his ability and performances with the bat have far exceeded Strauss and place him up there with some of the greats.

  • Waspsting on September 20, 2012, 11:34 GMT

    a theory of sorts I have is that players tend to name someone they've had reasonable success against (and exclude those who've usually gotten the better of them) when asked "whose the best you've faced or bowled to?"

    Gavaskar named Roberts. Roberts dismissed him 4 times - Holding dismissed him 10. Marshall and Imran dismissed him more than 4 times (not sure of exact figure), and often for low scores.

    Lara names Wasim. He had his 50-50 tussles with Wasim, as one would expect when two greats meet, but I'd say Lara came a long way short of 50-50 against McGrath.

    McGrath names Lara, who dismissed more often than one should the "best batsman I've bowled to", and often for low scores. Furthermore, even when Lara scored against Aus, he usually just saw off McGrath and attacked the others

    The subjective opinions seem to be a sort of self-celebration - "this is the best I've faced (and look how well i've done against the best, hence, i must be great!)"

    as opposed to... (cont)

  • Ananth on September 20, 2012, 11:32 GMT

    At the end of over 19.3, India's score was 149 for 4. Afghanistan's score was 136 all out. The difference of 13 runs is less than the 14 no balls delivered by Afghanistan. Not to speak of the 5 catches dropped. Or the Lbw which would have been reversed with DRS. Says that the match was much closer than it seemed and India are ill-equipped to handle the bigger teams. Unfortunately the two previous WC wins, adopting the 7-batsmen theory, seems to be blinding the Indian captain from facing realities. If Yuvraj's performance yesterday is pointed out he only made up for the poor bowling of Zaheer. But where is the fifth bowler? or for that matter the fourth bowler? Ananth

  • Waspsting on September 20, 2012, 11:27 GMT

    going off on a slight tangent regarding subjective picks, w/ Sarosh

    for batsmen, i perceive that there's an aesthetic bias used in subjective 'picks' - attacking players are usually ranked higher than equally effective but more defensive players. While i don't necessarily agree with this - I understand it as a standard.

    For bowlers though, there doesn't seem to be anything.

    Inspiring action? Lillee had that, but so did Waqar - while Imran's was a thing of beauty.

    Attacking spirit (as opposed to 'niggling')? Waqar was out and out attack - he's under-valued, but Garner (relative to Lillee) and Ambrose/McGrath (relative to Wasim) are also in the same boat

    Visible sideways movement? Imran and Waqar should rate higher than they usually are.

    I can only conclude that subjective bowling assessments are highly whimsical at best - there is no standard, let alone a solid statistical one - that supports common 'expert' opinions (cont)

  • Som on September 20, 2012, 11:00 GMT

    Ananth - Unni's suggestion and the concept of only including GTE 25 wickets intrigued me. I ran the data for the first 11 on the list. Considered 30.1 as 'Overall Avg' and for each players best decade, for the 'peer other avg' calculation included only the GTE 25 bowlers. This is what I got:

    Players PrevAdjAvg NewValOthBowAvg NewAdjAvg Imran Khan 18.26 30.34 18.23 Barnes S.F 18.57 26.35 18.8 Muralitharan 18.6 31.12 18.89 Hadlee R.J 19.07 30.17 19.12 Marshall 19.09 30.71 19.21 McGrath G.D 19.53 30.65 19.86 Ambrose 20.04 29.19 20.47 O'Reilly 20.04 28.70 22.45 Laker J.C 20.2 26.81 20.84 Garner J 20.66 29.71 20.74 Wasim Akram 20.9 29.97 21.29

    As you can see O'Reilly drops off the top 10 list and Garner goes above Laker. Since I did this for only the top 11 from the previous list, I am sure, when done for the entire list, further reshuffling of rankings will occur. Thanks ! [[ Som The period peer values will also change. That has not been done. Ananth: ]]

  • unni on September 20, 2012, 7:51 GMT

    Good. I agree that (as I mentioned in the last comment), mostly this effect would get cancelled out in this analysis. However, I feel that it could make an impact when you use it for individual peer analysis.

    I would still be interested (just for knowing) the distribution of the part-timer usage across various top bowler's peak (10 yrs). When you have time, please find this out.

    Another thing which I would like to know about this analysis is the stability of the results. I think some sentences I read above triggered this. How does player positions change if we repeat this for a range of years. i.e, how the tables looks like if we do it for 8/9/10/11/12 years? If this happened to be relatively stable (which I think, it will), then this deserves to be the one of the best objective measurement. [[ I feel we should not go over-board with this analysis. I am certain that if I change the basis to 8, 9, 11, or 12 years, the results will change quite seriously. No need to do those analysis to prove my statement. If Tendulkar's best years were 1992-2002, how the periods 1992-2000 or 1994-2002 turn out will depend solely on how Tendulkar performed during the periods 1990-1992 and 2000-2002. There is no way this can even be anticipated. It will vary from batsman to batsman and vary within a batsman across the years. And there is no way to determine which period is best. What is the benchmark. I am also aware that this analysis, as it stands now, has found broad acceptance because it shows Tendulkar in excellent light. If the 8 or 12 year analysis pushes him to, say, 12th in the table, it will not have the same level of acceptance. Readers, please do not misunderstand me. Many of your commendations are genuine. The analysis is very sound. What is necessary to understand that the 10-year period, sound it seems to be, is arbitrary. I myself think it is an excellent lapsed period to hang one's hat on. That is only an extension of our comfort level with a nice number like 10. Finally let us not forget the half-empty or half-full syndrome. Tendulkar's non-best-decade period's average is 49.29 (16 years). Lara's non-best-decade's average is 50.70 (6 years). Note how much of a difference is there in the difference between the two sets of values. Ultimately there is only one solution. That would be the ultimate career analysis adjusted for peer values. The entire career (best and worst periods, on-form and off-form, injuries-ridden and fit-as-a-fiddle, facing great, good and poor bowling, flat decks, green tops and dust bowls - all evening out in the end) has to be considered at a stretch. The complete career figures are to be adjusted for that specific career-book-end period, with all peer calculations and recent tweaks we have discussed. That will be THE analysis and will not have any arbitraryness thrown in. In almost all cases the player decides on his career period. It is his choice if he plays for 9 years, 16 years or 24 years. I will do it soon. Ananth: ]]

  • Sameer Kumar on September 20, 2012, 7:01 GMT

    Great analysis on the best decade stats. Still boggles my mind on how you deal with such scale of statistics on a weekly basis! I have come across your blog recently,(with the VVS article) and I was wondering if you ever did in past an analysis of Test Batsmen taking into account both the batting average of others in their careers (like you did here for the 10-year period) along with quality of opposition bowling?? What I mean to say is that someone like SRT and even Gavaskar were probably better batsmen than Viv Richards, Ponting and Lara (or may I even suggest Bradman here) because they faced Ambrose, Walsh, McGrath, Holding, Marshal, Garner etc etc. Is their some statistical formula you can use to factor it in?? If you have not done it earlier, please consider doing it, if possible. And just to make it a bit more tricky, think its possible to factor in the number of varied grounds/surfaces they batted on as degree of difficulty? Thanks... In advance [[ I get the impression you want to set out with a premise "Xyz is better than Abc" and ask me to find ways to prove that. It does not work that way. Just because Gavaskar faced Marshall, does it make him a better batsman than Bradman. What about the bowlers faced by Bradman. Anyhow did not Lara not face McGrath. And great bowlers did not just come from one country. Anyone would say that it was easier to face the West indian quicks in West indies than the New Zealand swing bowlers in helpful conditions down under or face the Indian spinners on the dust-bowls of India. Please go through the list of 150 odd article I have written and peruse some of those. Ananth: ]]

  • Vivek Shantharam on September 20, 2012, 6:30 GMT

    Excellent. Two things come to my right away.Lara doesn't figure in averages list due to his inconsistency in consistent performances during a year and Tendulkar due to his longevity comes below him in runs per year list. One another thing, the world fails to realize is the similarity in Strauss/KP's record. Nothing much to separate them and their contribution. Hayden's "second innings" started in 2001, otherwise he d have been more higher too. His would be similar to KP/Strauss as well, ie quick runs in lesser years.

  • Ananth on September 20, 2012, 4:46 GMT

    Unni: I just did the overall average for all 2055 Tests excluding the 1000 odd sub-25 career wicket trundlers and the effect is very significant. The overall bowling average has dropped from 31.9 to 30.1. The drop is 6%. That means these 1000+ bowlers had, as expected, high bowling costs associated with them. I will use this in the complete Peer analysis. Thanks Ananth

  • Ali Shah on September 20, 2012, 3:57 GMT

    This is probably one of your top 3 works if not the top 3. I can't remember when I read so many comments on your blog. Hats off to you

  • Sarosh on September 20, 2012, 3:54 GMT

    Wasptsing, Quite right about the bowlers in the World XIs. The stats are not as strong in their favour as it is for the batsmen. Even if we take say 8 yrs as Ananth suggests, the stats stay much the same for batsmen. As we reduce the time period to just a few years some players start putting up Bradmanesque stats – but once we start getting to 3 /4 years – the time periods render such studies meaningless. Longer than a decade and it seems difficult to sustain top performance for various reasons. Even if we use RPIs (an even more faultier measure for Test matches than averages), Sobers, Richards and SRT are a cut above the pack on a sustained basis. Bradman, as usual, is a cut above any pack.

    As regards bowlers, I feel perhaps the “isolation” factor kicks in. For eg.in Lillie’s peak decade 1972 -1982, the closest we have is Holding (1976 – 1986) and Willis (1974-1984). Nothing much to choose from between Lillee and Holding. As mentioned, for batsmen, some of them “cluster”. For eg. Ponting, Kallis and Dravid are direct contemporaries and debuted at almost the same time. Since they have similar stats there seems to be not much to choose between them. Similarly, great players of an era seem to cluster around similar stats. This contrasts with players who were alone at the pinnacle (so to speak) over their best decades.

    Perhaps, this explains some of it. But you are right – the bowler selection for the world XIs is still much more subjective than the batsmen selections.

  • Jas Singh on September 19, 2012, 23:27 GMT

    amazing analysis! Good Insight!

  • unni on September 19, 2012, 18:57 GMT

    So, let me try to work this out.

    1. Obviously, employing part-timers wouldn't affect 'Avge' column. 2. Employing part-timers would somewhat bloat 'Peer Avg' column. 3. Overall average (sea-level) already has the effect of all part-timers. 4. Now we are talking about a ratio ( which is Sea-level / Peer-Avg ), so in effect, the part-timer effect should have been cancelled in this factor. 5. Now we are multiplyig this with the Avge which is independent of the effect. So, it looks like the step 4 could be harm-less as long as there is not much variation in the part-timer usage across different eras. i.e, without hard-data it would be difficult to analyse the effect of this, purely in a logical manner. And the needed data is : How many part-timers were in action (or how many runs gave by them and how many wickets were taken) in each player's decade? And you can have a reasonable condition for who is a part-timer. gt 50 wickets could be a good one. [[ Unni, I think I have got the perfect solution. 50 wkts was too high a cut-off. Between 45-50 wkts, we have Jasu Patel, Chauhan, Toshack, Abid Ali, Chappell, Winston Davis, Meckiff, Peebles et al. Almost all are serious bowlers. So I lowered the mark to 25 wickets. Exactly at 25 wkts, I have Flower/Compton/Kuruvilla/Sohail/Bosanquet et al. I have checked the 24 to 1 group. Very few can even be remotely considered to be a serious bowler. So the stats for 25 wkts are Total wickets: 60700+ Gt/Eq 25 wkts: 53000+ (nearly 90%) No of bowlers: 570 out of 1590. The bowlers excluded are these 1000 or so and numbers like 1 for 294, 1 for 287, 2 for 351, 3 for 378, 10 for 622 and so on. So your intuition is correct and the conclusion seems to be very sound. But I will implement it for the full Peer analysis only. Fundamental changes have to be done. Additional data elements have to be created and populated. And so on. Thanks a lot for some great insights. Ananth: ]]

  • unni on September 19, 2012, 17:07 GMT

    While re-reading your article, this just struck me. >> The peer adjustment has to be crisp and clear and has to make a telling impact. While the bowling comparisons have to be across the board, the batting peer comparisons have to be restricted to the Top-7 batsmen.

    Regarding the bowler part : I did some analysis earlier and was trying to consider all bowlers' statistics. Then I had problems because, the sides used a lot of part-time bowlers and this was affecting the combined figures lot of harm. (In that case I was trying to analyse one team's performance). So, I have a feeling that if you don't stick to maximum 4 best bowlers in each match, you might end up in problems. Unfortunately I'm unable to pin-point what that problem could be now as I'm in a hurry. [[ Unni I get your point. I think this deserves a serious look before I do the comprehensive peer analysis. I am also aware that the 32 figure includes all sorts of people: not jut part-timers, but even the famous five, Hooper, Boje, Shastri, Rafique and Giles, the only bowlers with bowling averages exceeding 40. But they all took 100+ wickets in Tests. So they can never be ignored. We are evaluating the batsmen, an elite group of 3000+ run players, against a tough benchmark: the top-7 batting average. However the bowlers, the same elite group, is evaluated against ALL wickets. This article is only the starting point for the complete evaluation of players using peer-adjusted values. So we have to get that correct before I do that particular analysis. 4 best bowlers in an innings is fraught with pitfalls. 4 best based on their ctd figures or 4 best in that innings. What about considering only bowlers who have captured 50 career Test wickets and above. Ananth: ]]

  • Waspsting on September 19, 2012, 14:43 GMT

    "... a more objective back up to the subjective “All time World XIs”...Cursory looks at career averages have people scratching their heads as to why seasoned observers inevitably produce the same picks over and over. ... the likes of Sobers, Richards and SRT were clearly a cut above a pack in their day. Not many sportsmen perform consistently better than the “No.2”s for periods as long as a decade."

    What of these same "seasoned observers" typically -

    -rating Warne over Murali? -Lillee into world 11s? -Akram the best of his time (ahead of Ambrose, McGrath)? -excluding Imran from the top?

    you've selected a handful of examples in making your point and ignored just as many that run completely counter to it.

    The analysis does NOTHING in justifying 'typical' subjective rankings of players - into world 11s or more broadly, relative to their contemporaries.

  • Dr. talha on September 19, 2012, 10:02 GMT

    I fail to understand why people include either warne or murali in the World X1, why not both. If u have a world class off-spinner and a similar leggy both should be there in the playing eleven. It would be a deadly combination. I will have gilchrist at 6 and my bowling attack would be: Gilchrist(WK) Imran Hadlee Marshall Warne Murali

  • Nitin on September 19, 2012, 9:01 GMT

    Excellent Article Ananth.. Just one observation from the top 30 bowlers around 10 have their best decade between 1992-2002 (Ambrose, Wasim, Waqar, 1/2 Walsh, Pollock, Donald, Warne, 1/2 Murali, 1/2 Mcgrath, 1/2 Shoaib ) Whereas same figure is 7 for 1980-1990 with Imran, Hadlee, Marshall, Garner, Holding, 1/2 Walsh, 1/2 Willis. Bowling in these two periods was of best quality. In 1992-2002 Tendulkar has his best decade. Lara and Dravid had their best half decade during the same period. Richards, Border and Miandad have their best decade in 1980-90 but they didn't face bowlers from their own team i.e. Marshall, Garner, Holding in Richards case and Imran in Miandad case.

  • Shadab Raza on September 19, 2012, 6:42 GMT

    Good job.

  • Tanny on September 19, 2012, 6:06 GMT

    Someone mentioned Imran boosting his performance by preparing crumblers in the 1980s. If he did, that was just one series, 1986. In 1981 the pitches were flat, and plus he wasn't the captain, anyway, to have been able to prepare any pitches.

    Some also say Imran took a lucky break by not playing two years during the timeframe considered? This is like saying Dennis Lillee took a break in the early seventies. They were both at their peak when injury hit and took some great potential performances away from them. The problem with Imran was he was already 30, and coming back at that age was much more difficult than Lillee coming back in his mid twenties. Same with Waqar, his injuries were early on, so he could come back for a bit. Even though both Lillee and Waqar definitely were robbed of some more Tests at full pace. Same obviously for Shane Bond. Marshall, McGrath, and Hadlee, on the other hand, did not have any career-threatening injuries. [[ Injuries and absences would happen at different times in a player's career. I think the 10-year period is long enough to take care of all these events. Imran has captured 250 wickets during the said 10-year eriod and that is a lot of wickets. If there had been a bowler who played 50+ Tests in 10 years and captured 250 wickets at 18.26, he would be challenging SF Barnes for the best bowler of all time. So I suggest do not bother. Ananth: ]]

  • Sarosh on September 19, 2012, 5:35 GMT

    Uday, Spot on. It is not merely about facing most of the best bowlers of all time. SRT also produced the 6 best normalized average of all time while doing so. Also, what makes players stand out is not only their avg. over a decade , but also whether OTHER players averaged a similar amount. This is where Bradman (again as usual), Sobers, Richards and SRT stand out. For eg. The “W”s played in a similar time frame. So, also were Kallis,Ponting,Dravid etc contemporaries. So, this case of somewhat “isolation” at the top reveals something.

  • Sarosh on September 19, 2012, 5:28 GMT

    Ananth, Thank you for following up and the great effort you have put in. It clearly resolves some earlier issues and also throws up some amazing surprises – none more so than Imran Khan. We notice there is not even a remotely negative squeak about Richards, SRT, Akram etc now (The usual suspects have gone into hiding with tail tucked firmly between legs). These players were Titans of their time and they should be recognized as such. A certain amount of heckling and criticism from keyboard warriors is always on – but beyond a point the criticism crosses the line to insult.

    Re.SRT and Bang/Zim. A check between 30 Nov 2002 and 30 Nov 2002 shows an overall avg. of 62.55. Sans Ban, the avg. actually goes UP to around 63 (Zim around that time was actually a pretty decent bowling side. Certainly better than India). Even sans Zim though, the avg. comes to around 61. [[ Sarosh Bng/Zim come into relevance only for complete careers. Even then I think these things even out. Ananth: ]] The true beauty of this analysis, in my opinion, is that it finally provides a more objective back up to the subjective “All time World XIs”, which almost always include Bradman, Sobers, Richards and SRT. Cursory looks at career averages have people scratching their heads as to why seasoned observers inevitably produce the same picks over and over. Bradman aside (as usual), the likes of Sobers, Richards and SRT were clearly a cut above a pack in their day. Not many sportsmen perform consistently better than the “No.2”s for periods as long as a decade. [[ We should not forget that if the period, for some reason or other, taken as, say, 8 years, the tables might look quite different. However I feel the top batsmen of these table might not change much. Similarly if RpI was used instead of average, players like Lara would benefit. However average is the most widely accepted measure. Ananth: ]]

  • Dr. talha on September 19, 2012, 5:15 GMT

    @Gerry the Merry. "Imran made crumblers" Brother in 1980-81. Imran was not the captain, so how come he had the authority of making the pitches? I may agree with you about 1986-87 series but in that series, it was not only Imran who got 18 wickets in 3 tests, but even Qadir did. Marshal, Gray & Walsh bowled brilliantly, but then again, the WI bowlers bowled well everywhere in the world from 1975-1995. Even in the absence of Ambrose, Walsh & kenny bejamin demolished india (Azhar's formidable unit)in Mohali in 1994. Imran always came up with something special when he played the mighty windies. Even in 1976-77 in WI, he got 25 wicket in 5 tests.

  • Dr. talha on September 19, 2012, 4:56 GMT

    From 1985 to 1995, when Imran again became the captain after his injury, Pak played 21 test series. They won 13 and lost only 2. And they won almost all the major ODI tournaments durining this period. Which is quite remarkable! Then in 1995 the match fixing controversy came in, and started the downslide of Pak cricket. Though they did manage a few memorable wins under Akram (WSC in Aus, 3-0 whitewash of WI, 1999 Indian tour & reaching 1999 WC final). But the consisitency was lacking. Ponting took over a very formidable S.Waugh side, which had already started to dominate world cricket. And then after all the greats retired, he coulndt keep up the winning streak. even his batting average has dropped from 58 to 52, and looks to go down futhur. I believe Border, taylor and waugh did all the hard work, which led to the Aus dominanace.

  • Dr. talha on September 19, 2012, 4:43 GMT

    @knowwho.I m a great admirer of Botham but i believe his brilliance lasted for a very short span. Thats the reason why his bowling & batting record are no where close to that of Imran. Secondly botham failed as a captain and against Pak & WI (2 best teams of the 80's) he performed poorly. Imran during his time had the best percentage of Man of the match & Series award, and still must be in the top 3. Why i rate Imran highly as a captain is because, i remember during the 70's even if Pak team drew a match against Eng, aus or WI, Pak players considered that to be a win. this was the mindset of the team. Imran lifted such a side. And after 2 years of his captaincy he had the courage to remove the senior players (which included his first cousin, Majid) and bring in youngsters. He made a team which challenged the very best and the influence of the great Khan lasted even 3 years post retirement. (cont..`)

  • Harsh Thakor on September 19, 2012, 4:31 GMT

    Sadly Viv Richards was interspersed by the Packer years but for which he would have been the batsman of the 1970's.Greg Chappell would also have come very close.Viv also reached his peak from the 1976 period and an important part of his peak period stretched to the 1980's.In the 1980's to me ,Gordon Greenidge was Viv's greatest challenger with Javed Miandad.Alan Border was an epitome of consistency like Sunil Gavasakar.In pure stats I feel Gavaskar dominated the 1970's more than a batsman dominated any decade after the war.Similarly Dennis Lillee but for Packer years ,would have dominated the 1970's more than any pace bowler ever in a decade.

    I am pleased with the likes of Bradman,Sobers and Hobbs at the top-3 of the best batsman ever with the likes of Tendulkar.Weekes place is deserving and Tendulkar's but sadly Lara hardly figures high in the adjusted averages inspite of incredible performances in the 1990's,overshadowing Tendulkar in his peak. [[ There is no doubt that Tendulkar before 2002 was sensational. Lara was reasonably consistent but despite all his huge innings, offset these with a string of small scores. I get the feeling 10 years is a fairly long period. And the low number of not outs: 400*, 153*, 80* and three low score* s tells the tale. In fact I just found out. Lara's % of not out innings, viz., 2.6%, is the lowest of the top-100 run-scorers. This one sentence conveys a lot of information. Ananth: ]]

  • Harsh Thakor on September 19, 2012, 4:19 GMT

    A great effort Ananth.In bowling it is a tribute to Imran Khan,statistically in his peak period arguably the best paceman of them all who took 236 wickets in just 40 tests he bowled in.His strike rate was better than Hadlee or Marshall from 1980-1988.However arguably he benefited from his break period from 1983 to 1986 ,missing test series in Australia in 1983-84 and home series v England,New Zealand etc.To me morally in the light of statistics Richard Hadlee was the bowler of the 1980s who took 330 wickets in 60 test at an average of around 19 runs and as trike rate of around 47 balls per wicket,in his peak period from 1978-1988.He bowled ,for a relatively weak attack.It is better than Malcolm Marshall and Imran Khan at their best .

    George Headley has no mention who averaged over 60 runs and infact in the 1930's Bradman was compared to George Headley,arguably the Don's all-time greatest rival. [[ Harsh Headley did not make the 3000 run cut-off. But I have edetermined his adjusted average and shown this to an earlier comment of .Arijit. The response is listed below. [[ Instead of including Headley I could show his adjused average as 64.93 (66.72*35.92/36.91). (Test 187 to 274 - 1930 to 1939). He would be there in the top-6. But then what about Graeme Pollock. ]] Ananth: ]]

  • Gerry_the_Merry on September 19, 2012, 2:49 GMT

    Dr.Talha, before you conclude that the wickets in Pakistan did not favour bowlers when West Indies toured, please look at the scorecards. They make for scarcely believable reading. I remember each of those 7 tests almost ball by ball. Front line batsmen averaged in teens (on both sides). When the same teams played in West Indies, batsmen did better. In any case Imran has clearly said he prepared crumblers, which ended up favouring the Windies quicks also, so there is hardly any scope for argument.

    But with other teams Imran simply bamboozled them with high class bowling, e.g. when India and Australia visited in 1982-83. Imran reach his peak in terms of pace+skill combination. It was not just sub-continental ball tampering. He also played one of the greatest series ever by an all rounder in 1982 in England when he took over as captain. It is a tragedy that at his absolute peak 1977-83 he did not play more tests. Peter Morris has made the same point - played too few tests in peak phase.

  • Lourens Grobbelaar on September 18, 2012, 19:29 GMT

    Thank you for all the effort you put into these analysis Ananth. It surely is a mammoth task. I do not always comment on your articles, but read most of them. I only comment when I feel I am not only commenting for the sake of commenting.

    I was wondering why the bowlers weren't also evaluated and ranked according to Strike Rates, but only averages? Surely for a bowler both are important, especially in a Test Match analysis.

    Greetings, Lourens. [[ Lourens, the Bowling Average is a wonderful measure, comprises as it is of both Strike rates and Bowling accuracy values. It is possible that the strike rate is a more important measure in Tests, say 55-45, but the Bowling average is overall an excellent measure. Anyhow I have analyzed these separately in many of my previous analyses. Ananth: ]]

  • swapnil Ghayal on September 18, 2012, 18:30 GMT

    i am reading your post almost 3 years now. [[ Since you have taken the trouble of sending a comment, I have published the one sentence which is fit to see the light of the day. Thanks, Ananth: ]]

  • petermorris on September 18, 2012, 16:32 GMT

    Batsman best decade analysis: Based on adjusted average congrats Ananth one of your 2 or 3 best ever and thats saying something It proves what I always thought.that sobers was the second greatest batsman ever.its also the only thing that Ian chappell and Geoff boycott fully agree on that sobers was the best bat they ever saw. It is a much better way of judging different eras than for example your best 80 innings period which grossly favoured the likes of Ponting and kallis who could cram in 80 innings in very short time periods while at their peaks.one gripe its obvious that if you play for England you play more far tests than anyone else thats the only reason cook and KP feature in highest average no of runs per year 83 tests in 6 years cook and 88 for Kp in 7 years I mean come on its not a level playing field You talk about Mohammad Yousuf how many tests did he play in those other years why did you not include a no. of tests played column? [[ In the average table I have innings palyed and in the runs table I have Tests played. Ananth: ]]

  • knowwho on September 18, 2012, 14:39 GMT

    @dr.talha

    Thk and believe the entire article is abt test cricket and not one day cricket. AnywAY dont thk any of the tournaments you say would be equal to a WC period. And no way people can come near ponting atleast as of now.

    @ananth Probably decade period is not the right indicator for allrounders( which means one can assume that they perform with both and ball in the same test). may be 5 year period would be the way to for them and weightage should be given to test in which they have performed in both aspects.

  • dinesh on September 18, 2012, 14:30 GMT

    @Boll: Apologies for jusmping into conclusion that Imran was solar system's ahead without looking at other' Stats(Given there were 2-3 of the very best playing at the same time).

    I was so stunned by Imran's stats that i felt no one could Come near. and thanks to you that i got to know of Hadlee's numbers.I have never watched him play but he somehow happens to be my fav bowler. I watched a lot of his videos of youtube and was mighty impressed and was thinking what would have been the Situation of the Batsmen of 80's where they had to face Westindian quciks, Imran, Botham, Hadlee and some others.

    Any batsman who averages over 50 during that period might probably have averaged near 55 now. Just a subjective statement, but i feel the numbers will probably will be close to this.

    This article of yours does a lot of Justice to the batsman who played during 80's.I thank you for that as i never got to watch them(Born 43 days after Tendulkar made his debut) and made me go to youtube.

  • ruudraza on September 18, 2012, 14:13 GMT

    sorry, i missed how u adjusted the avereges but still the players ahead of sachin did'nt play in a very competetive era. The list of top bowlers is almost exclusively filled with players who peaked after the retirement of sobers while a few of them played before sobers peaked. also all west indian batsmen ahead of sachin have much better avereges home than away where it would have been more difficult to adjust. weeks 69 home and 49 away, walcot 69 home and 40 away, sobers 66 home and 50 away whereas Sachin 55 home 54 away. [[ Very difficult to justify these statements. Also subjective. And let us not forget that Bangladesh and Zimbabwe were two of the opponents. This is not to belittle Tendulkar's wonderful average during his peak. Ananth: ]]

  • knowwho on September 18, 2012, 14:10 GMT

    @dr.talha

    while one should acknowledge that imran was a top cricketer arguably the best produced by asia it must not be forgotten that the total number of wins under him was 14 and during his career is 26. It would be better to analyse from this point of view rather than saying he is greatest/best.a After all cricket is a team game.

    compare this with marshall any other cricketer who played around the same number of matches during his period. Say 80-88. I am sure imran during his peak was a lone ranger 79/84 and hence couldnt win more. For that matter richard hadlee didnt he win matches for nz in aus,in nz against win, in 88 he won a test i in india,in england too. So let us accept these facts. I beleive the allrounder should contribute with ball & bat in the same test.By that account i believe the best allrounder should be ian botham. from 78/82 his best years.

  • Som on September 18, 2012, 14:00 GMT

    Ananth - One of the very best. Thanks much.

    I had this wrong impression that that best phase of Lara would be very high in this list. SRT's achievements are phenomenal. This shows that before he turned 30, he reached his peak. This could be an interesting subject to investigate, where players were before they turned 30.

    I sorted the bowling average data based on ascending values of 'Other Bow avg', and it shows that apart from Barnes, the next best era was between 1946-1965. The 80's are not the best, in fact between 2002-12, when Zaks had his best, the quality of bowling was better. So was 80's not the best bowling era then? [[ We also seem to under-rate bowlers like Lindwall, Miller, Davidson, Tyson, Trueman, Tayfield et al. Ananth: ]] Given Weekes' and Walcott's, standing, wonder why they have not been taken far more seriously when discussing the very best.

    Though I always thought Imran was one of the best, but the fact that he would top this analysis came as a surprise (a very welcome !). Now I understand why Wasim is so apologetic in his tone when mentioned as the best bowler from Pak.

  • Boll on September 18, 2012, 13:21 GMT

    @Ananth. Yes, I would agree that PAM (peer adjustment method) and CLAM (career level adjustment method) - copyright on those shocking acronyms by the way - solve a multitude of problems. As you suggest, they give a truer reflection of individual performance. They also remind us (or inform us) of how the balance between batting/bowling has changed over the years.

    As an added bonus, they automatically refute the often exaggerated claims that so-and-so played in an era dominated by bowlers, or that whats-his-name faced nothing but pie-chuckers.

    @ruudraza. Not quite sure how you draw those `clear` conclusions...

    No statistical analysis can tell us everything, but in particular a glance at the top 5/10 on the list of the best-decade (ave) analysis here is a walk along the avenue of the greats. Sure, some are missing and perform better on other measures, but it`s a memorable list just the same.

    Cheers, Dave

  • Nitin Gautam on September 18, 2012, 13:20 GMT

    Sad to see Lara not appearing among the top in avg base best decade for batsman list.He is too good a batsman to not appear in top 5-7 in any list.

    while I fully understand this is mainly bcos of his extremely low no. of not outs (4 as told in article) but asking for removing runs against minnows & removing N.O. is far fetched. after all no batsman would like to get out & his fearless breath taking batting was the reason he is so revered as well as the reason for his so low count of N.O. no reason to appreciate him because of same if not ready to accept its flipside.

    Lara has done enough to be among the greatest & IMO best player to watch (atleast among those who I have seen) when on song at his best. [[ That is the way Lara played and all of us should accept that. He has also shown fair level of consistency since his best decade average is closer to his career average. Ananth: ]]

  • Nitin Gautam on September 18, 2012, 13:03 GMT

    Anantha

    1st of all this is extremely comprehensive article that, as per me, only you could have envisaged & compiled. stats degree or now, your articles are absolutely reader's delight & Im sure just like me, others also wait impatiently for your articles.

    Imran's credentials as one the top 5 was never in doubt but this was surprising but certainly not more surprising than finding warne at such low position. even below akhtar, gibbs, walsh willis etc.

    SRT had his best decade in best possible bowling decade during his time so tht once again proves why he is held in such high regards overall.

    however is it possible to add the avg BQI faced by batsman in their best decade (in both tables i.e. based on avg & max runs), just to have an idea of how gud bowling attack he faced in his best decade.

    Similarly for bowlers, if its possible to add what was the avg of T7 batsman he bowled to in his best decade. probably some more insights will come. [[ Nitin The reason I was able to get the peer values for Test 1 to Test 2 is because I have now the 1877-todate values for each Test. It is a simple matter of doing the difference. However it is tough to get the bowling quality since that is unique for each batsman and has to be determined in flight, so to say, by perusing all the matches. I have the career figures but these do not mean much since the last 10 years have not been that great on the bowling front. Same problem with the bowler information. Ananth: ]]

  • Dr. talha on September 18, 2012, 12:21 GMT

    @Boll. I think u missed the captaincy part. That why i mentioned few of Imran's captaincy achievment. No doubt keith miller was a great all-rounder. But over the years we have seen how great players have under performed when they became captains. Imran was one of the exceptions. That photo is of Allan lamb. I dont think Imran played against Robin smith.

  • Boll on September 18, 2012, 11:39 GMT

    Ah, apologies - just jumped out at me, and I jumped in without thinking. Had a sneaking suspicion I`d missed something... [[ I get the feeling this oversight of yours shows clearly that the Peer adjustment method is one of the best we have all come out with: It is certainly a joint effort. Worrell's 51.75 has been adjusted upwards to 54.77 and Sanga's 59.66 has been adjusted downwards to 54.88. The adjusted values are probably truly reflective of their respective skills. What I also like is the slight upwards adjustment of Tendulkar indicating that his best years coincided with a very good bowling era. This also means that not all modern batsmen are treated in a similar manner. Of course,when the career level adjustments takes place, the peer values would be over much longer periods. Ananth: ]]

  • ruudraza on September 18, 2012, 11:19 GMT

    Great analysis anand, these stats clearly show that Mcgrath Murali and Headlee should have been picked ahead of Akram, Warne and Lille in cricinfo's all time greatest team(Imran had a great avg but did'nt pick as many wickets as these three). wheather at there peak or overall they were the best. As for people doubting Sachin, clearly he was better and more consistent at his peak than his peers . Considering the fact that Sachin peaked in the 90s whereas dravid, ponting and kallis in 2000s when the quality of bowling had declined significantly people should stop comparing him with anyone and accept he is the greatest. [[ It is not correct to say that people should stop comparing one player with the others when the concerned player is behind others in quite a few measures. The bowling strength has been taken care of in the adjustment. So the final average cannot be ignored. There are five other batsmen ahead of Tendulkar. And I suggest do not bring in ODIs into this discussion. Ananth: ]]

  • Boll on September 18, 2012, 11:15 GMT

    @Ananth. I may have missed something re. Sanga, but is it possible to have a best-decade average of 54., when his career average is 56 or so? Or have I misread some of the fine(r) print? [[ You suddenly made me feel that I have missed out something. Excellent common sense question. Sanga's actual best-decade average is 59.66 which has been adjusted down to 54.88 (59.66 * 35.92/39.05). 39.05 is the Peer T7 average between Tests 1537 and 1991. Ananth: ]]

  • sumanth on September 18, 2012, 11:03 GMT

    Brilliant Analysis Ananth.... May we have similar kind of analysis for ODI? [both bowling and batting]

  • Boll on September 18, 2012, 10:47 GMT

    @Ananth, yep it`s a subjective call. If I did have to pick an all-rounder and a `keeper though I`d go with Sobers and Gilchrist. (Knott, I believe was the `keeper that many purists were backing). Imran and Hadlee would certainly come into consideration purely on the basis of their bowling.

    @Dr Talha. I see that you`ve cunningly raised the batting bar to omit the great Keith Miller (bat ave 36.97, bowl ave 22.97) and also an excellent record against the best team of his time (Eng) bat ave (34), bowl ave (22). I appreciate your spirit, but think that `no argument that Imran was the greatest` may be stretching things a little.

    Any ideas on the photo? A.Lamb? R.Smith? vs England I presume, year?

  • Santosh Sequeira on September 18, 2012, 10:05 GMT

    Great Work Ananth, i think i remember in one of your earlier articles you had highlighted that Imran in his last 50 test matches (going by his averages) was as good a batsmen as Viv Richards and as good a bowler as Malcolm Marshal, now that's some monster we are talking about over period of 50 tests.

    It is commendable that after putting in such efforts you are being so very receptive to adverse comments by some people who lack the basic reasoning to fathom the purpose of these articles. Actually we should pity them.

    All i have to say to those writing disparaging remarks is that if they want to be critical, please let it be from a cricketing point of view. If qualifications do matter then stop watching sports as it is the Achilles' heel for most sportsmen (Tendulkar, Afridi et al.).

    regards

    Santosh [[ Thanks for the nice words. I have always accepted brickbats, if aimed from a sound platform, but criticism for the sake of criticism bugs me. But I handle these now better than earlier when I would have reacted. Ananth: ]]

  • Dr. talha on September 18, 2012, 9:47 GMT

    There is no argument that Imran was the greatest. Perhaps there is only ONE cricketer in the history of the game that fulfills this criteria. 1- Bowling average of under 23 2- Batting avearge of over 37 3- Brilliant against the number one team of his era (WI) 4- Captained his side to a lot of memorable wins, in both tests & ODIs. Imran had the habit of taking challenge against the very best. Series wins in Eng and India in 1987(two very formidable sides, with no neutral umpires & 3rd umpire). Would have beaten WI in 1988 as well,if the umpire had the knowledge about leg spin bowling. He was the only captain along with ponting to have won 4 ODI tournaments with 6 or more test playing nations. 1986 Australasia cup (beating Kapil's men in the final) 1989 Nehru Cup (beating Vivs team in the final) 1990 Australasia cup (beating Allan borders Aus, world champs at that time) 1992 WC (Grahm gooch' eng)

  • Dr. talha on September 18, 2012, 9:07 GMT

    @Gerry the Merry. It was not the wickets in pak that helped the bowlers, but it was the great bowling and fieldiing skills of that windies team, that their bowlers suceeded in Pak. Remember their were a number of great bowlers who toured Pak in that era and performed horribly, and most of them blamed the wickets for thier poor performance. The WI bowlers had the skills to bowl in Asian conditions. Look at their performance on the 1983-84 indian tour. No wonder they dominated world cricket for so long.

  • Boll on September 18, 2012, 8:49 GMT

    @Dinesh (again) - must admit I was a little sceptical about the stats you mentioned re.Imran 1980-1986, but you`re absolutely right - 184 wickets at 15.9, and an average of under 20 in every country he played in. Astonishing figures.

    I`m not sure he was solar systems ahead of everyone else though... Hadlee`s 227 wickets in that period at 18.9, including 130 in 20 away tests at 18 is just as incredible. Hadlee in Aus (52 wickets in 6 tests at 14) still gives a whole generation of Australian fans nightmares.

    Tales of his exploits on and off the field at The University of Sydney, my alma mater, remain legendary. Great to see him coming out on top of an analysis such as this - not to mention 3rd highest ICC bowling ranking of all-time. The early 80s was a period of great fast bowling, and Imran was as good as any of them

  • Boll on September 18, 2012, 8:30 GMT

    @Dinesh, I think one of the disappointing aspects of the whole Cricinfo XI selection (for national teams as well as World XIs) was the requirement to select 1 all-rounder, and that players such as Sobers, Imran, Hadlee were able to be selected only in that category. I`m not sure that a national team necessarily needs one - the 2 great modern sides (WI and Aus) did without for most of the time - and less sure that a World XI needs one. I can also certainly see I.Chappell`s point that the `keeping position (which went to Gilchrist by a wide margin) should have gone to the best wicketkeeper, without reference to his batting skills. [[ That is a subjective evaluation. Unless otherwise the analysts are able to prove with catches floored data, I find it difficult to agree that Healy (or Strudwick or Taylor) is better than Gilchrist. Ananth: ]] Your suggestion that a World XI should have both Sobers and Imran (or perhaps Hadlee) has a lot of merit, simply because they are in the very top echelon in one discipline alone.

    Apologies for getting slightly off topic there.

  • Boll on September 18, 2012, 8:19 GMT

    Very interesting work as per usual Ananth - your lack of a degree in statistics notwithstanding!? I particularly enjoyed your laconic response to the question...`NO`...lovely.

    Thanks for taking on board my suggestion for the last article (not `a demand` as someone put it). I`m sure I`m not the only person who appreciates your willingness to take suggestions on board, and undertake the extra work necessary to investigate them. You spoil us.

    Anyway, I just wanted to respond to a couple of Dinesh`s comments. Firstly I completely agree with you re.Sobers that he warrants a place in a World XI on his batting alone. I`m too young to have seen him play, but enough men of my father`s age still say that he`s the best they saw/played with/played against (not to mention his statistics) to convince me that he`s clearly amongst the best 5 batsmen of all time, probably top 3.

  • unni on September 18, 2012, 7:37 GMT

    @Anshu : I agree with Ananth's comments. Maybe I would try to picturize this without taking a hard stand on this. Suppose we are trying to measure the 'greatness' of mountains. There are two possibilities. 1. For each mountain, find the average height of the surrounding terrain, find out how high is this average compared to the sea-level. And offset the hight of certain mountains with this factor. This is to say that, "OK, I know that in Himalayan region, generally the mountains are high; so I'm not that surprised if I find a 7000m one". This is what happens in the current method. 2. Your recent method : It is something like instead of comparing the terrain height with the sea-level, compare it with average height of the country in which it is situated. I would like to go with a sea-level as this evens things more up. So, in my opinion, no harm in having a large sample size for the denominator. [[ Very nice comparisons. The all-time average is a known value as is the sea-level. So the comparisons are with known all-inclusive figures. Doing it otherwise may mean that Mt.Aconcagua may seem taller than Mt.Everst because that one rises above flatland. Lovely exchanges. Thanks to you both, Anshu and Unni. Ananth: ]]

  • Anshu N Jain on September 18, 2012, 6:43 GMT

    Ananth, I see the merit in your comment. The suggestion does not work for comparing players' performances (time-bound, test numbers-bound etc.) across different ages/eras. This, though, could be a good absolute peer comparison gauge for any duration of a particular player's career.

    On my first suggestion, I still hold that having to bear the burden of history is agreeable to bearing the burden of posterity for this analysis. [[ Anshu This is not really Peer analysis work. I have only used the Peer numbers to adjust the averages for the concerned 10 year period. The real Peer analysis is when we compare a player's numbers with those of his peers and that will be done soon. At that time I will look at all these suggestions. Ananth: ]]

  • Ananth on September 18, 2012, 3:27 GMT

    Arijit, you are absolutely correct. The impact of Wides and Noballs is far more than what I had imagined. I did a study and found that I have over 6300 bowler analysis lines which contained wides and noballs data. Almost all these are for the post-1985 period. I myself never anticipated such numbers. This must be the result of the wonderful contributory effort of scorecard collection done by a few readers recently. But I now feel that this cannot be just done as an addendum to an existing article. You certainly have stoked the fire and I think this warrants a separate article. The impact must be quite significant. My apologies for unserstating this. Ananth

  • dale on September 17, 2012, 19:30 GMT

    I commend you for exhibiting such class and dignity in answering the few childish and insulting posts which are published ever so often.The work you do is greatly appreciated by any genuine cricket fan. [[ Dale, thanks for the kind words. I do not know why people are not able to appreciate the work. They are more concerned about whether I have a statistical degree or am a statistician. Couple of years back a reader, who was a good contributor until then, said he was quitting because I was not a statistician!!! As if it mattered that I had only a B.E. degree and not a B.S. degree. And Batting average = Runs / (Innings-Not outs). One does not need to know about Kurtosis or Newton-Raphson equation to do this. Ananth: ]] I have always felt that the batting average should be augmented by the RPI or RPT and likewise the bowling average should be considered along with the WPI or WPT. I am asking if there is some way you could manage to incorporate these factors in your analysis because I think both sets of measurements can co-exist. If not, I understand. [[ The Batting average and RpI(RpT) are two contrasting ways of looking at the Batting performance. Anything we do combining these two would be artificial and part-double counting. Probably the only way out is if there is a broad acceptance of what I do which is RpAI or Runs per Adjusted innings. Consider Not outs as Not outs only if these are below a certain score. In other words Not outs are Not outs only if the score is below 20 or whatever. Bowling average is perfect. It is a combination of BpW and RpB. But WpT is a separate entity. I would consider both if situations warrant inclusion. Ananth: ]]

  • Anshu N Jain on September 17, 2012, 18:32 GMT

    Unni: good thought there; the analysis should indeed have the essence of time-less-ness, but i feel this is not achieved by burdening the past with the future. It only complicates matters further when we want to do this over a specified time period.

    Ananth/Unni: let me throw in another suggestion:

    Use the T-7 avg in the SPECIFIC tests over the time period in which the player played, and the T-7 avg in ALL tests over the same time period (instead of the avgs used in the present adjustment formula). Same for bowling avgs.

    This will be a true peer comparison and adjustment, and does not carry the burden of either the past or the future - nearly timeless in my view.

    Also both numeratr and denominatr sample sizes will be comparable (ratio of 5-6). Using the all-time avgs, this problem is magnified as the numeratr is based on 2000+ tests (min ratio of 20 for Ponting - ~100 tests)

    And we can do this over any time period of a player's career, including the entire career.

    Your thoughts? [[ Anshu Your first suggestion made more sense than this one. Let us take Kallis. His unadjusted avge is 66. If in the 444 Tests book-ending his peak decade, the T1-avge for all is 39.18 and the (to be newly calculated) specific Tests T7 is 39, his average will remain at 66. This certainly does not reflect that he batted in a relatively batting-centric period. It is only by using an across-all tests figure will we be able to nail down the ease of batting or bowling. And what if the specific Tests figure is 38. Kallis will even go up. And Hobbs will get nothing even though he batted during a difficult period. Ananth: ]]

  • unni on September 17, 2012, 16:36 GMT

    Anshu/Ananth : I would like to disagree here. Current methodology somehow gives the analysis a quality of time-less-ness since the entire match spectrum is used uniformly to everybody. So, even the players in the past are affected by the deviations which have happened in the future. This exactly offsets the effect on the stats of the later players due to the deviations that were already happened in the history of the game contributed by the past players. If you take the upto-the-test averages, this symetry is lost. So, the past players are assessed with a smaller sample space while the later players are assessed with a bigger sample space. In my opinion, this would tilt the balance one or the other way without any benefit. Or I fail to see the aesthetic quality which it brings. [[ Unni, I love your turn of the phrase. I am glad that I left it open and waited for comments. Let us leave it like that and wait for more comments. I felt your last comment's impact when I worked out the 1877-todate values for Test no 134. The number of wickets was around 4300 while for the last Test it was over 60000. Ananth: ]]

  • arch on September 17, 2012, 14:03 GMT

    I always thought he was one of the greatest bowlers but even I am surprised - I would have expected him to be in the top 3, but as it is, it makes sense. Lots of people have pointed out his injury prevented him from playing at his peak, particularly in Australia where he would have done well, so it is a bit redundant pointing it out here. He is, just simply, a great. I am actually more intrigued by Bill O'Reilly - he really ranks high. Actually, looking at the top bowlers and the top batsmen here, it seems that if we just go by their peak decades, the Cricinfo All Time second XI has the better bowling attack while the Cricinfo all-time first XI has the better batting. What a juicy contest! [[ I should then do a 5-Test simulation between these two teams. Ananth: ]] And to round it, looks like Richie Benaud got his all time XI just right, with both Imran and Sobers.

  • Syed on September 17, 2012, 13:41 GMT

    Hi Anantha,

    Thanks for the analysis. It's very informative. I always had a feeling that statistically the number of not-outs play a significant role in career's average - against one nation or cumulative. If we don't consider the number of not-outs and disregard the minnows (Bangladesh, Zimbabwe, Kenya, et al), we can really get a good snapshot of Lara's career. In addition, considering his stats against the greatest of bowlers in his era. I believe he stands out as the premier batsmen of his generation. As Glenn Mc Grath and Murali stated at their retirement speeches that he was the best they have bowled against.

    Thanks!

  • Anshu N Jain on September 17, 2012, 11:47 GMT

    I feel that there will be complementarity of impact on bowling and batting averages. Just as "difficult" runs coincided with "cheaper" wickets (only in terms of bowling average) in the past, over time "easier" runs will coincide with "dearer" wickets, with cumulative effect. The same token should apply to both. Not sure I have been able to put this across in words though (my writing skills to blame)!

    Another request: Could you also post the adjusted career averages for batsmen and bowlers (with appropriate cut-offs for tests/runs/wickets etc.)? If it can be worked out from the excel you have already posted, thanks! [[ Anshu, let us keep the career values and adjustments to a later article. I am anyhow going to re-visit the Peer comparisons article later. In fact I am not going to any work also now in a hurry. I do not want to rush this. I have some urgent T20 WC related work for the next few days and if I rush, I may make mistakes. Once the -to-date values come in we have to be clear about what-to-date. Best decade beginning, end, mid-point or if career, career-start, end or mid-point. But your idea is excellent and will be used in the Peer analysis. Ananth: ]]

  • Anshu N Jain on September 17, 2012, 10:09 GMT

    Good one Ananth!

    In the peer adjustment, instead of using the "all-time Top-7 Batting average", could we use the "Top-7 Batting average upto that test"? [[ This is a very valuable insight. The golden days for batting are in the future and probably right in evaluating the batsman numbers with the values upto that Test. But look at the impact it will make on bowling, especially for the period until 1920. The 1877-to-date bowling average for Test no 274 is 28.06. All the dearer wickets are in the future bringing the value upto 32. So Barnes will lose out a little: probably correctly so. No problems with the data. Available straightaway. Let me see whether I can re-post for this analysis itself. Probably only the downloadable Excel sheet. Ananth: ]] It doesnt, intuitively to me, seem to make sense with burdening the numbers in the past with those in the future; vice-versa makes perfect sense however, because the game has "evolved" in some sense - pitches have changed (flattened?), batting equipment has improved, rules are geared towards batsmen etc. [[ I get the feeling that the implications will be more for players upto around 1970 or so. After that the numbers would not vary by more than 2/3%. No I am not really sure about who will benefit. Let me do it and see. Ananth: ]] Havent gone through the tables in detail yet. Looking forward to doing so soon...

  • Srinivasa Moorty on September 17, 2012, 8:45 GMT

    Ananth,

    Great Analysis and great article. Having said that, I always feel that there is not enough justice done to the actual match winners who can turn the match single handedly and have repeated the stint time and again, Batsmen like Micheal Bevan, Laxman, Javed Miandad, Inzamam, Amla,etc and bowlers like Ambrose, Waqar, Wasim, Abdul Quadir, Kumble etc. [[ Bevan ??? Ananth: ]] Can we expect some article on these greats who when they take field their teams relax, stating "The match is in safe hands". We have seen one great aricle from you on these lines on Laxman alone when he retired, I am hoping to see something more generic comparison or otherwise covering all the greats. [[ Your criteria is unfortunately way too subjective. Ananth: ]]

  • Ranga on September 17, 2012, 8:11 GMT

    Hi Ananth - The more diverse the topics you choose, the more intertwined they become, to form a perfect sequence of articles! When I saw Imran in this list, the immediate impulse was to run through your blogs on career slices, to reaffirm that his best years with bat & ball (for a fast bowler past 32) were coinciding! I think these lists are turning out to bring increasingly different sets of people.

    Good to see Steyn & Swann leading the post 2000 debutants. Swann in particular, should be commended for getting most of his wickets in seamer friendly / flat batting conditions. Sad to see Umar Gul & Fidel Edwards averaging 17 wkts per year, despite having 3.5+wkts/test. Shows how Pak & WIN have reduced their number of tests (and may be these players' fitnesses as well). This fall for great teams known to produce best of fast bowlers is something that the fans of test cricket should be worried of. [[ Pakistan we can understand. Their home series are all in disarray. How long can they continue to play in UAE. Ananth: ]]

  • Dinesh on September 17, 2012, 7:37 GMT

    Just went through Imran Khan's yearly Stats. I dont have much idea of the Past i.e 70's and 80's. But if you remember anything would you mind telling in 1979 if any UFO Landed in Pakistan and gave Imran some special powers. This is because end of 79 he averaged 33. In 1980 he averaged 21.Then it came down to 19 in 81 and then to a Mind Boggling 13 in 82. 13 runs per wicket and he picked 62 in just 9 matches at almost 7 wickets per match.Just Incredible.

    From 1980 - 86 he was probably Solar system's ahead of other Bowlers.184 wickets in 34 matches at an average of 15.xx . Forget Marshall,Forget Holding,Roberts or lille. Imran Khan was the king of the bowlers during this period. I am not considering 87 and 88 as he played 13 matches and took just 50 wickets. If we look at the stats i feel with due Respect to Sobers, his place in the All time XI as an all rounder isnt guaranteed.I would pick sobers as a batsman rather than as an all rounder.Imran as 4th bowler and allrouner. Hats off

  • Dinesh on September 17, 2012, 7:20 GMT

    Ananth: You said 99-00 was partly Donald's era. I would say it was full Donald's era. by the End of 2000 Donald was finished. He played just 8 matches after 2000 and took jus 22 wickets. Same was the case with Waqar and Wasim as well. Waqar took just 60 wickets post 2000 at average of almost 30. Wasim took 31 wickets after 1999 in 13 matches. So all three were almost finished by the end of 1999.

    Waqar and donald were on top of world Cricket charts for the whole 10 years they played. Tendulkar's 10 top years coincided with the Waqar, Donald, Ambrose best years and Part of Warne, Mcgrath, Murali. Lara's best years coincided with Part of Akram,Waqar,Donald and whole of mcgrath, Warne, and Murali's best years. [[ So there is not much to choose from. Only interesting fact is that SRT faced all top bowlers while BCL and others did not have to face their own bowlers. Ananth: ]]

  • srini on September 17, 2012, 5:03 GMT

    I knew Imran was awesome but I didn't expect him to top the table. But there are a few caveats. From 79-83 IK was spectacular. He bowled heavily and batted but since missing 84 cos of a stress fracture he gradually decreased his bowling. Like I mentioned in an earlier article (forgot which one), he bowled almost 12 overs less before and after his injuries and even his WpT were down (.3 WpT). He certainly was more effective afterwards.

    Also, his RpI was 38 but average was close to 52. In this respect he is like the Michael Bevan of test matches ;-). Now, this is not to denigrate IK. He is most certainly Asia's greatest cricketer so far but in my humble opinion he couldn't have had those numbers if he hadn't reduced his bowling load after those injuries.

  • Uday on September 17, 2012, 1:51 GMT

    Without in any way wanting to join the band of Tendulkar-overglorifiers, its incredible that his best decade came in one of the best decades for bowlers. I don't think there is anyone else in the top 12 who has such a complete overlap with the decades for best bowlers (The only others who do in the entire table are Richards, Border and Crowe) Also, consider the fact that firstly, among his contemporaries, he is the only one who had to face ALL of the best bowlers of his decade, and secondly, he was pretty much the best ODI batsman in this time. Truly amazing. [[ Yes I agree that the 1990-2000 period was the era of Younis, Ambrose and partly Warne, McGrath, Murali and Donald. And no Indian bowlers. But this will be probably be the case for many batsmen. Lara's best is between 1995 and 2005, the era of Warne, McGrath and Donald, Younis and Murali. That gives me an idea. I could use this strict 10-years concept to determine the best 10-year periods for spinners and pace bowlers: and the worst. Would be a real eye-opener. Ananth: ]]

  • Ramarao on September 17, 2012, 1:26 GMT

    Your have come up with an interesting one which is worth a wait. Request : As you observe I have seen only three players (sachin, Lara and steve waugh) averaging above 50 in 90 -99 and as many as 25 players averages above 50 in 200s (2000-2009)( its includign perterson who has 49.94) Also can you come up with analysis of Test and ODI Batting statistics during Sachin Tendulakar's Career which include many Great players like Sachin, warne, Lara, Kallis, Murali, Ponting, Sehwag, Hayden, McGrath, Donald, Ambrose, Flintoff etc.

    Can you also show us your interpretation of the Strike Rates in ODIs and Run rates in ODIS from 89 - 2012. Since a lot changes in these 23 years with bad periods for Batting in early 90s to golden period for ODIS from 93 - 2005. Also the century rate is increased in recent times compared to early 90s. [[ Many requests. Will come thru and pick up the interesting ones. Thanks. Ananth: ]]

  • Ramarao on September 17, 2012, 1:11 GMT

    Great Work Ananth.As usual; Thanks for someone who suggested this for you.

    Can you also Come Up ODI analysis for the Top Batsmen. Also on the decadal Century rates in ODIs. As Before Tendulkar Haynes had 17 and it looked as it is eternal. While Tendulkar Thrashed that record while defining a Process to Make a ODI century, Kohli and Amla are in their way to chase it.

    Dont you think the ODI bowling has Become weak in last 5- 7 years?

    Can you also Come up RPI analysis in Tests and ODIS for top ten 15 Run scorers also including Bradman, Weekes, Hammond, Hutton, Hobbs as you always does.

  • xia on September 16, 2012, 21:44 GMT

    Do a stats check on All rounders as well. I believe Imran Khan was the best Bowling all rounder and during his later stage he avg 50+ with bat too and with his abilities as captain Plus how he taught others like Waqar and Wasim that makes him one of the best cricketer ever and I believe in top 5 cricketers of all time for his combined abilities.

  • Randy Jackson on September 16, 2012, 20:18 GMT

    Your analysis has lots of flaws. I am working on my own analysis with the data. Will send you a copy so you can learn and fix the issues. However, good try but a below par product from you. You need to learn lots of statistical theory to be able to perform a proper analysis. Do you even have a degree in Statistics? [[ NO. Ananth: ]]

  • Ather Ashraf on September 16, 2012, 17:49 GMT

    Please do the Allrounder decade wise analysis. Like imran has not only improved his bowling average. He has also improved batting too especially under his captancy. I think his batting average was aroun 52 in that decade and bowling averge around 19. Please do analysis of Allrounders as well

  • hadi on September 16, 2012, 17:24 GMT

    based on this, test team of planet earth

    bradman sobers hobbs walcott weeks tendulkar gilchrist imran barnes headly marshall

    :)

  • Sudarshan P.N. on September 16, 2012, 16:18 GMT

    Ananth, as I see the table again, notice that in the top 20 are batsmen like Boycott, Barrington, Simpson, etc., who are unlikey to figure in the top 20 of most cricket lovers. This analysis does bring out the value of such unsung heroes. Thanks Ananth. [[ Barrington, stodgy and defensive he might have been, has never been off any recent batsmen analysis. Ananth: ]]

  • Shahnawaz Ali on September 16, 2012, 16:09 GMT

    Fully concur with Abhishek Mukherjee please do a similar analysis for all rounders though I can't figure out how you will decide which decade was best for a particular all rounder since you will have two factors (three counting catches taken / fielding stats) to consider. [[ Get both batting and bowling to a common ground, say, runs and then decide. I have yet to work out a satisfactory wkts-to-runs-conversion methodology. Ananth: ]]

  • Gerry_the_Merry on September 16, 2012, 14:25 GMT

    Ananth, in my previous post, there is a mistake. Richards did not score the only century in those 7 tests in Pakistan. There was another century. Remarkably, it was scored by none other than Imran Khan.

  • arijit on September 16, 2012, 13:36 GMT

    Anantha, I think you are underestimating the impact of no-balls and wides on bowlers’ figures. It should be closer to 1 – a substantial margin --- than 0.1. I did a random check on the scorecards of six Tests played by WI between 1983 and 1986 and calculated the no-ball and wide tally (given in brackets) against the total runs conceded to the opponents. I picked WI because they used few spinners, so the results wouldn’t be skewed. The results: 1st Test 83-84 vs India: 371 (13); 5th test: 331 (10); 1st test 1984 vs England: 426 (22); 3rd test: 429 (20); 1st Test vs England 1985-86: 311 (8); 3rd test: 388 (24). Total: 2256 (97), which is one per 23.26. In a previous post, I think you showed how the career averages of Garner, Marshall and Ambrose were close together at 20.94, 20.97 and 20.99. If we could subtract the no-balls and wides, perhaps the averages of Marshall and Ambrose would fall below 20. Regards [[ Arijit Give me some time. i will do an analysis. Ananth: ]]

  • Dinesh on September 16, 2012, 13:33 GMT

    Now coming to Ordered on Run Column where even in this their Best decade were overlapping. Taking the 2001-2006 decade this was injuries started to creep into Sachin and he for some time became mortal and also when Lara Began his rise Back to the Top.

    Both Again played the same number of matches. 55-57. But SRT scored 1800 fewer runs than Lara and his RPI was 12 fewer than BCL. Coming to Countries Aus going by the number SRT was better but largely aided by his 241 but even Lara was the Same give his numbers were boosted by his 226. In the remaining 6 innings SRT scored 140 and BCL in the other 9 innings he scored 220 runs. So both were below Par. IN Eng SRT ar was better than Lara.In SA,Pak and NZ Lara was better.In ZIM SRT was better. Both played 4 matches in ZIm and scored same number of runs. I would say BCL failed here out 0f his 220 runs 191 were scored in 1 innings and in 3 he dint score any.SO SRT was better here. But in India SRT flopped.His RPI just 43.

  • Dinesh on September 16, 2012, 13:12 GMT

    Thanks for adding the Number of Tests in the excel.

    For the last Two decades the best batsman choice has always been between Sachin and Lara despite however Good Ponting,Dravid,Kallis have been. I just dug a bit deep into their peak batting period in terms of averages. Sachin 1992-2002 and lara 1995-2005.So there was an overlap of 7 years from 95-02 which is when the discussion about who was the best between them started. So i just looked at the RPI from this. Found some similarities. Both Played almost same number of matches.SRT 70 and BCL 71. and at home as well 34 and 35. Lara played 12 innings and SRT 118.But 600 more runs playing 8 less innings and his RPI was a whole 9 runs more than BCL. 54-45.

    IF we come down to Countries.SRT was better than BCL in Aus and England and South Africa.He scored 1541 runs to BCL 1931 Runs in this period.SRT Played 11 fewer Matches than BCL in these countries. So this was the period when SRT overtook lara in the Battle the Premier honours. Contd

  • Gerry_the_Merry on September 16, 2012, 12:51 GMT

    I followed Imran's career from 1977-88 which were his peak years as a bowler (the last great performance was 11 wickets in Jamaica in 1988).

    In the mid '80s, Pakistani pitches were real crumblers, where visiting team bowlers also took wickets by the bucketful, and it is especially so in matches against the West Indies. Imran was at the forefront, but in 7 such tests, there was only one century by Richards for both teams put together.

    Imran, Marshall and Hadlee spanned the 1979-89 period. In this period, wickets, total avg, home wkts, home avg, away wickets, away avg are as follows

    Imran 251, 18.64; Home 118, 14.36; Away 133, 22.43 Marshall 323, 19.91; Home 133, 19.40; Away 190, 20.27 Hadlee 289, 19.28; Home 119, 19.12; Away 170, 18.00

    I know this is not a home/away type of article, but just wanted to make sure that we note the extremely poor pitches in Pak in the '80s.

    During this period, in Pakistan, Tony Gray took 14 @ 16, Sylvester Clarke 14 @ 17, Walsh 11 at 18, Croft 17@18..

  • Farrukh Hanif Awan on September 16, 2012, 12:07 GMT

    Thanks Ananth. That would be great. Politicians should in general terms stay away from all forms of sports. Sports is meant to be healthy competition not war. Sadly in our case, Pakistan vs India is war. Such high emotions. I hope for the greater good the tides are turned for once and Pakistan produce some world class batsman instead of world and vice versa for India. Cheers

  • Ananth on September 16, 2012, 11:41 GMT

    The number of Tests has been posted in the Runs/Wickets worksheets in the Excel file. Ananth

  • Chaitanya on September 16, 2012, 11:23 GMT

    Surprised that you didn't know Imran Khan's bowling average was below 20 in his last ten years and his batting average was 50 in the same period.! What an all rounder in his last decade. [[ Probably I am the living example of the Tamil proverb: "What you know fills the fist, what you do not know fills the world". Ananth: ]]

  • Farrukh Hanif Awan on September 16, 2012, 10:34 GMT

    Dear Ananth,

    I see a lot of Pakistani's and Indian's mostly posting in your blog section. I have one humble request. How about doing a analysis of Pakistan vs Indian series in tests and ODI's. The reason i am asking for that is that it is such a high stake games it brings out the best in players. So just for some analysis can it be done? [[ Farrukh Probably the right time to do it would be just before the India-Pakistan series planned towards the year-end, if ever the politicians allow it to happen. Ananth: ]]

  • Tanny on September 16, 2012, 8:40 GMT

    I'm not surprised either, by Imran being at the top. And one should remember this is the decade he lost two bowling years from Test cricket, right at his peak: the two years after he took 88 wickets in 13 matches in 82-3. He resumed bowling only in '85, at the age of 32, with his pace diminished.

    Also check out the top bowlers 1980-88:

    http://www.espncricinfo.com/magazine/content/story/484478.html

  • dale on September 16, 2012, 7:39 GMT

    A surprising revelation is the prominence of the great spinners who played prior to 1990. Only the incomparable Murali is ranked ahead of O'Reilly,Laker,Lock,Underwood,Gibbs,Verity and Grimmett who incidently is in a virtual tie with Warne. Then we also have to acknowledge Bedi who is comfortably ahead of all Indian bowlers. Chandra was more potent but Bedi was consistently great and was the best of his type in world cricket. I have always wondered why Iqbal Qasim did not play more often for Pakistan. While not as magical as Qadir he was just as efficient and at the very least they could have been known as a set of spin twins a la Laker/Lock or Ramadhin /Valentine. [[ This will be fine-tuned fiurther if and when I can do the adjustment based on separate bowler-type classification, as asked for by Sudarshan. I need to create multiple sets of data fields and would probably do it before the complete peer analysis. That would allow a spinner in a pace-dominant decade to get due recognition. Ananth: ]]

  • Farrukh Hanif Awan on September 16, 2012, 7:38 GMT

    Thanks for the reply. I understand the point you are making but we can as always stick with the real test, that is the TEST cricket. I am a follower of Test cricket as I prefer the game its in natural form where even a player like sehwag and gayle have to sweat it out and not swat the ball out. The trend can be because of the coming t20 worldcup but i hope it remains just a small little blimp in the glorious history that is test cricket. All I can say is maybe cover the guys who don't fall into the ten year category so that we can see emerging players. Anyways thanks for making me look through a million little excel sheets. Always fun [[ I have some time for the T20 version only when it is a truly international contest. As such I think the T20-WC will be a good competition. Any one of 8 teams could win the contest. I will probably rule out only the so called minnows. Ananth: ]]

  • Sudarshan P.N. on September 16, 2012, 7:22 GMT

    Ananth, this is perhaps the best analysis of yours till date. Shorn of purple patches, vital contributions and other short term and/ or subjective assessments, this clearly indicates the normalized performance of a cricketer for a sustained period. I was just wondering whether it would be possible for the adjustment for bowlers to be made on spin vs pace separately. Of course imagination boggles if Ambrose's best 10 years had coincided with that of Marshall and Garner:) [[ Since three readers have made such statements maybe the analysis is good. Anyhow many thanks. I have the numbers as a composite value for bowlers for each Test. So it is easy for me to find the difference between the two concerned Tests to get a period peer value. Would be a major recision to separate this as Pace/Spin. Maybe I will do that before my promised re-visit of the peer analysis for the entire career. Ananth: ]]

  • Faridoon on September 16, 2012, 7:15 GMT

    Hat's off to you for putting together this analysis.

    When I read the title, I got the impression that we will see the best batsmen and bowlers for each calendar decade (1970s, 1980s, 1990s...) This, I might is due entirely to my faulty perception; the title isn't misleading in any way. [[ I think I anticipated this and made the ctitle as correct as possible. "Their best decades" although "10-years" might be a better word than "decades". Does not sound good, though. Ananth: ]] Your numerical take on things always strips away the 'opinions' and 'perceptions' and leaves behind pure, statistical and well-qualified facts that none can refute. Bravo. [[ Many thanks. Ananth: ]]

  • leon on September 16, 2012, 6:52 GMT

    I know you want to give not outs its due respect. But if you don't mind is it possible to reproduce this list without considering not outs? [[ Leon, the most respected and understood batting measure is the Batting verage. I do not want to change that basis. However all required data is available in the Excel sheets. There is nothing to prevent you from doing the innings-based work yourself. Ananth: ]]

  • Farrukh Hanif Awan on September 16, 2012, 5:54 GMT

    Dear Ananth, Big fan. Always a treat to read your analysis. My question being of all the latest bowlers these days. Like for instance, 2010 (since major turners and new blood came in most teams) onwards can you do an analysis or a predictive analysis how they will turn out say in 2020. I know it is a lot of speculative work but I think we will be able to see how they will fare in coming years and compare careers at the end of 2020.

    I remember Imran Khan, the run up, the jump and the ability to calmly appeal and then smile or raise his hands up. Thats all he did. So much composure on a cricketing field. I think Srisanth should be coached by Imran on how to handle himself. He could have been a good bowler for india. Anyways that my point, can we predict how these bowlers or batsman (new age) turn out to be. Like Amla and Kohli, can they come near tendulkars record? Kohli seriously has a shot at it or any bowler who can come close to murli's record. Peace keep posting. ICC needs you. [[ Very difficult to do any prediction on the future. Would there be a reduction on Test matches or increase. Where is the ODI game going. Where is the T20 going. Already we have seen, for the first time, three 3-T20 bilateral series during the past few months. Will this be the trend. ICC needs me like a sore thumb. I need you guys and you will probably like to read my articles. Ananth: ]]

  • Vimalan on September 16, 2012, 5:43 GMT

    its very interesting to note that both Marshall and Ambrose have same number of wickets with similar average in their peak 10 years, just a slight variation in their adjusted average due to peer average difference during their periods. even more interesting is Marshall's best was 1980 to 1990 and Ambrose's 1990 to 2000. a clear indication of a shift in the bowling leadership of the team [[ And Roberts was from 1974 to 1983. Ananth: ]]

  • Saurabh on September 16, 2012, 2:37 GMT

    Can you give a sneak preview of your lara article? Desperately waiting for that. [[ Already given in response to Arijit's comment!!! Ananth: ]]

  • Rohit Mehta on September 16, 2012, 2:20 GMT

    As always Ananth, your article is in a class of its own. Hats off for overcoming the logistical challenges associated with such an article. Again, as rightly pointed out, Imran ahead of SFB made my jaw drop! Couldn't agree more with you and Waspsting about Ajmal. But perhaps these awards tend to be biased towards the batsmen. Alas!

    Back to the article, Murali's 588 in a decade is perhaps a record which might never be broken. Also Warnie and Pigeon's best decades coming at the same time was perhaps the single most important cog in the Aus wheel, along with their awesome Top3, as well as the greatest wk/bat of all time- Gilly. [[ Also the fact that there was no weakness with Australia-2000. Left arm pace bowlers are nice to have but not necessary for team balance purposes. And it is better to have two quality leg spinners than an off-spinner for the sake of one. Ananth: ]] Also, Sir Gary's achievements as a batsman alone warrant the all-time #2 status, proved so beautifully in the numbers above; perhaps the greatest cricketer ever.

    We must wonder how these numbers might have been if there were as many tests/teams in the early 1900's! And as you said, if there were no wars. Oh, for the romance known as Test Match cricket.

  • dale on September 15, 2012, 23:55 GMT

    Ananth:You have managed to brilliantly implement a "weight" formula which levels the playing field and makes it easier to appreciate all the great players across all era.Bradman is Bradman. Sobers with an exceptional average at the superb rate of 61 runs per inning while serving as a stock bowler - 40 overs per Test is truly a magnificent performer. Hobbs is the other stand out batsman,and together with Weekes they are the only other batsmen who scored at a rate of over 60 runs per innings. Will people finally accept the fact that Imran is one of the very best fast bowlers of all time ? In addition to his bowling Imran also supplemented his bowling with solid batting stats by scoring @ 37 runs per inning while maintaining a very good average. Lance Gibbs' stats validates his reputation as the finest off spin bowler during the sixties and further justifies his inclusion in an all time West Indies XI and also as one of the top five exponents of the art of off spin bowling. [[ Pl see reply to Sifter. Gibbs has lost out on many counts. But he was atruly world-class spinner and earned his wickets when 300 was a way-out target. Ananth: ]]

  • Sifter on September 15, 2012, 22:53 GMT

    Ah if only Healy had bowed out a couple of years earlier, then Adam Gilchrist could have made the list...I was looking for the best keeper who actually kept for ALL of a decade ie. NOT sangakkara or flower, maybe Stewart as well might miss that qualifier. And just to add, I really like these type of lists - 10 years is a fine career for any cricketer, and it's going to be a real shame that Sachin Tendulkar's ridiculously long career is going to distort analysis among cricket fans from here on. Also bears mentioning how low Shane Warne is. His 1993-2003 average of 25.00 is very close to his career of 25.41. Take out his poor first year, and his last 4 years must have also been very close to 25 as well. Strange how a man so consistently injured, suspended, shrouded by controversy and supposedly 'de-mystified' by various teams/technologies still managed to get years 11-14 almost as good as his best decade. [[ I feel I may, in the months to come, re-visit the Peer analysis covering the entire career of players. That would be the ultimate table. But the 10-year period is a very nice slice of time: long enough to add weight and short enough for us to see 70+ and 18+ from players whose names do not start with 'B'. Ananth: ]]

  • arijit on September 15, 2012, 20:12 GMT

    Ananth, there’s one thing I always think of mentioning whenever you write a post on bowlers but keep forgetting: it’s that the extra runs from no-balls and wides began being credited (discredited?) to bowlers’ figures only from July 1983, skewing comparisons across generations. I don’t expect there’s the individual data to “correct” this, but it’s something we should remember while appreciating the 10-year peak of Imran or the career averages of Marshall, Ambrose, Wasim, Waqar, et al. Marshall, in particular, bowled a lot of no-balls. I remember that those days, often half of the first 35-40 runs scored by the Windies’ opponents would be extras. [[ What does it matter. Would the no balls/wides debited to the bowlers cause anything more than a difference of 0.1 to the bowling averages. I will anyhow do a study of NBs/Wides debited from whenever the data is available. Ananth: ]]

  • arijit on September 15, 2012, 20:10 GMT

    Walsh’s wicket count expectedly rises in his later years but it’s intriguing that his averages were better in his earlier years. Does it mean that operating with great bowlers kept his tally down? Also, is it too much to ask for a special inclusion of George Headley and his stats? He misses out from most tables because of the small number of Tests he played. But his career spanned more than two decades --- and what a player he was. Ten centuries in his first 17 Tests before the ill-advised post-war comeback spoiled it a bit. [[ Instead of including Headley I could show his adjused average as 64.93 (66.72*35.92/36.91). (Test 187 to 274 - 1930 to 1939). He would be there in the top-6. But then what about Graeme Pollock. Ananth: ]]

  • arijit on September 15, 2012, 20:08 GMT

    Ananth, a few minor observations: The batsmen’s 10-year peak seems to generally come early before they taper off a bit towards the end, but the bowlers seem to do better in their later years. I think this was discussed in relation to one of your previous posts on bowlers too. The aggregate tables for both seem to lie towards the later years, but this probably has to do with the increasing frequency of Tests played. I had expected that for most batsmen, their best-10-year averages would be like Sobers’s: some 15-20 above their career averages. But the list was an eye-opener: for most, the gap is much narrower. Just shows that consistency is what made them great. While Lara’s low not-out count affects his career average, his 10-year peak has the further handicap that his slump (1997-2003) came bang in the middle of his career. [[ There was no real slump in Lara's career during 1997-2003. As the following table illustrates, the good years alternated with the average years. 1997 12 21 0 859 40.90 1998 9 15 1 608 43.43 1999 8 15 1 832 59.43 2000 9 17 0 497 29.24 2001 9 18 0 1151 63.94 2002 7 10 0 351 35.10 2003 10 19 1 1344 74.67 For the selected period the runs are around 5700 and the average is close to 50. Ananth: ]]

  • Dinesh on September 15, 2012, 18:47 GMT

    Aah My bad. Just checked there's a colum with the number of innings played. Can you please add the Same in "Ordered on Runs" sheet as well. It will be easier to find the RPI.

  • dinesh on September 15, 2012, 18:43 GMT

    Ananth: In your reply to Pavan you said Tendulkar scored 5500 runs in other 13 years. I think its 8000 runs if you remove his best Decade in terms of average and its 7500 if you remove his best decade In terms of Runs. It probably is Typing mistake. [[ Yes, a mistake. Ananth: ]] Havent Gone through the Excel yet. Will post further after looking it.

    Boy what a decade Imran had. bringing down his average from 32.xx to 22.xx. That takes some doing.

    Ananth a small request. Can you add the number of tests column s well so that it can be a bit easier to find out who played more tests and probably their numbers got inflated a bit due to that [[ Will add Tests to the Decade Runs table. Ananth: ]]. I aint pulling any one's achievements down. Just wanted to compare it in this way as well.

  • Ram on September 15, 2012, 18:35 GMT

    Pretty much confirms that Imran is the best bowling all rounder ever. (where as, Sobers is the best batting all rounder)

  • Khalil Sawant on September 15, 2012, 18:32 GMT

    The fact that Imran should trump SF Barnes in the list speaks everything

  • Gaurav on September 15, 2012, 18:20 GMT

    Really great analysis Ananth,lots of work done for this,i just have one suggestion,if with the run accumulated/wicket taken you can also give how many tests/innings it took a player to accumulate/take those run/wickets it will also be great & easy to compare that how much per match it took a player to get those runs/wickets [[ I will add the Tests played in the Decade Runs table but not the average since that would distract from the contents of the table. Ananth: ]]

  • Abhishek Mukherjee on September 15, 2012, 18:01 GMT

    Excellent work, as always. Can you extend this to all-rounders? Might be an interesting study. [[ Possible since I have complete batting and bowling data avalable in a single place. Ananth: ]]

  • getsetgopk on September 15, 2012, 16:33 GMT

    That is one fabulous piece of work Ananth, very well done indeed and yes I was a bit surprised too, to see Imran Khan leading the bowling table with the best average. Without this effort would have never known. On the recent Saeed Ajmal debacle, why cant the ICC hire your services for a few weeks to see who is who and what is what? Stats dont lie and I would urge the ICC to think about it. Once again thanks for this fantastic piece, expecting alot more from you. [[ I myself feel that numbers do not really tell everything. The approach is correct. Have a team of impeccable jury members to decide. However they should look at the implications before it goes public. And the jury should be given complete tables so that they do not vote out of ignorance. Ananth: ]]

  • Vikram on September 15, 2012, 16:17 GMT

    Awesome work Ananth. The one instance where the best in batting really competed against the best of bowling has to be Richards vs Imran and what wonderful series they were. Neither of them could really outdo the other and the series were always drawn. I have never really taken to Cook but that achievement is something. However the biggest eye opener for me is Sobers. An allrounder to be the closest to DGB is some achievement. As for Imran, given his performances against Indian, it actually makes me feel good. Finally, WI as a team have 4 of the top 12 batsmen and 3 of the top 12 bowlers. And does it prove that SA did under-achieve, having the best batsman and the best bowler but never really cracking it during the 90s/00s? [[ I think the general perception is that South Africa failed in key moments of the limited overs competitions. Tests: I am not sure. Barring a match or two in which they could not complete the task (has happened to many other teams), they have tended to be there at the top, almost always in the top-3 over the past 20 years. Of course, if Morkel had hung on for 10 minutes more, they would have moved to Test no.1 during 2010 itself. Ananth: ]] Some really interesting questions. And finally, good to see SRT being there, I moved out of India in 2003 and moved from cricket t football, but till the time that I did watch cricket, SRT was a good batsman. As for BCL, overdue a recognition.

  • Pavan Kumar on September 15, 2012, 15:32 GMT

    Excellent Article as always Ananth..

    Sobers as second to Bradman for his ten best years was not surprising.His ten best years starting from 1958 when he scored his first century till the next years where he scored 2/3 of his centuries and 60% of his runs

    Sachin's best 10 years show that he was the best batsman in the world during his peak years. India could have won more matches if he had better support in both the bowling and batting departments

    Murali as always an amazing performer. Almost 75% of his wickets were captured during his peak 10 years. On he flip side remaining 8 years he captured only about 200 odd wickets. [[ Probably true of almost any player. Just couple of examples. Ponting has scored only 3300 runs in the remaining 7 years. Tendulkar has scored only 5500 runs in the other 13 years. Warne is better capturing 220 wickets in the remaining 5 years. Ananth: ]]

  • unni on September 15, 2012, 15:23 GMT

    Grant analysis, Ananth. I would count this as one of your best effort. Just because, 10 years would define an average career for a good player. Except that number you didn't use any other hard-numbers. The usage of the average of all players for peer analysis is a very nice approach. This brings in a nice number, which is derived from the available data itself. Technically, the first problem would have vanished if you had defined the problem as the best 100 (or n) consecutive tests for a player. But, I agree that this would sacrifice the aesthetic beauty of a human understandable block of 10 years. [[ 100 is too high a number when you realize that Weekes and Walcott played less than 50 Tests. 10 years seems to be a long enough period to hang our hat on. It takes care of almost any good player. Clem Hill is a good example. Ananth: ]]

  • KIJOHI on September 15, 2012, 15:09 GMT

    Personally, as an Australian, it's great to see Sir Richard Hadlee recognised as the masterful bowler that he was. One can only dream of the world XI teams built from such analyses.

  • Entertainment on September 15, 2012, 14:10 GMT

    Sports is purely about entertainment, numbers come distant second. Due to this addiction of stats players become selfish, put self above team, old timers carry on and on. Players should be remembered by the memories they left behind not numbers. After his retirement a certain individual rightly asked "Have I entertained"? I think every cricketer should ask this question to themselves if they want sports to survive. [[ Not sure what you want to say. You must realize that you have come into a Cricket Analysis blogspace. Surely this is what you should expect to find. Why complain. If you find that you have walked into the (wrong) analysis zone, you should quietly exit and move to the entertainment zone. Also please tell me what memories of Imran Khan or Sobers, a 25-year old youth of today would have. Ananth: ]]

  • Pankaj Joshi on September 15, 2012, 13:54 GMT

    Not taken time to digest article but ...Oh dear ... Milind's comment was so telling. It is on record that the Nobel committee collectively kicked themselves in the backside for not awarding Gandhi in 1947. Imran's presentation is phenomenal Ananth, thanks for the a(h)nalysis - and to think that we two generations have lived our lives hearing all about the great Botham and the great Hadlee. Truly twas bliss to have a white skin. Well at least the legacy is self-explanatory. For all three cases. QED

  • Hemant Brar on September 15, 2012, 13:41 GMT

    Well for me it is not that big a surprise to see Imran at the top. I was expecting him to be definitely in the top 3.

    I always rated him just a little below Sobers as an all rounder. In the last decade of his career (which is more or less his best decade), he averaged around 18 with ball and around 50 with bat.

    I wanted to check where does he stand in the Batsman Best Decade List and found he is at number 44 (Lara at 36, Zaheer Abbas, Asian Bradman, 43, Gooch 47, de Silva 53, Anwar, Sehwag, Laxman, Azhar well below him in the list).

    And he was a great captain/leader. And then add to it what he has done as a human, cancer hospital. Role Model.

  • Zulfiqar Ahmed on September 15, 2012, 12:59 GMT

    Very critical effort,you have done a wonderful exercise, really extracted the essence of greatness and summed it up in numbers the cricket belongs to. Being a fan of Indian team though it feels hollow not to see any bowler representing the sub 25 adjusted bowling average table.Nonetheless, batting giants are always there to recuperate this limitation.Bravo. [[ These are facts, Zulfiqar. The best Indian bowling average of 28.7 of Bedi is 74th in the Bowling average table and there are 8 Pakistani bowlers ahead of Bedi. India has always been a batsmen-centric team. Ananth: ]]

  • Ali Shah on September 15, 2012, 12:55 GMT

    Great article Ananth. Imran was a beast at his peak in the 1980s. If I remember correctly he has the highest peak of all post war bowlers. And it is a testament to Imran's skill and persistence that he was still so very good when he came back from a career threatening injury. Imran's bowling achievements get forgotten probably because he was one of the greatest captains the game ever saw and that is how he is usually remembered for Pakistan got very good bowlers even later on but never really found such a captain as he was. I really wish he will be Pakistan's next Prime Minister and usher in an era like that of his cricket captaincy. [[ What is remarkable is that imran is ahead of the pair of giants who played after him. I also wish Imran leads his party to victory as he led his team in 1992. Pakistan badly neads a decade of peace and quiet and a strong person at the helm. Ananth: ]]

  • Gerry_the_Merry on September 15, 2012, 11:47 GMT

    Elegant, Ananth. Imran Khan is indeed the Superstar here. Not only for the reason you have highlighted, but in the batting tables, though with a time lag, and with only 2378 runs, he delivered 51.63.

    However, as expected, his best average period lags his highest productivity period by 1 - 2 years...by 1989 he was not bowling a whole lot.

    Very surprised to see that Brian Lara's best average 10 years were enough to put him below Inzamam and Hanif Mohd...! But #4 by runs scored is certainly as expected. [[ Lara's curse, not that he would care at all, is the number of not outs. A 17-year career with 6 not outs translates to a 10-year period 4 not outs. He would go way up on a RpT or RpI basis. But I do not really want to take off Batting Average for this analysis. Ananth: ]] DCS Compton almost makes it to the top 20, but I suppose you have not built in home/away etc. [[ Home/Away splits would be red herrings. These are clearly defined career-segments and should stay that way. Ananth: ]] Ponting and Dravid take the top 2 positions by runs scored in the averages table also, making their achievement creditable.

    By productivity Kallis makes it to the top 50 bowlers also in addition to being #2 on runs.

    But hats off to GC Smith - 8th on productivity as a batsman alone, besides being captain in almost 100% of the tests played.

  • Waspsting on September 15, 2012, 11:41 GMT

    "There is a real, real surprise at the top. This should make people sit up and take notice... Who could have imagined that Imran Khan would have leapt over many other fancied bowlers? "

    Right here, me. The only guy i thought might edge him was Barnes. (already knew Marshall's peak didn't last a full 10 years)

    I've said it before, i'll say it again, Pakistani players almost never get the credit they deserve. Miandad, Imran, Waqar, Mohammed Yousuf - even Wasim Bari. Wasim Akram is the exception.

    (We have a beautiful example of this with the recent story of Saeed Ajmal - BEYOND RIDICULOUS whats happened there). [[ Yes, I concur with you all the way. No player can hold a candle to Saeed Ajmal's achievements, ACROSS FORMATS. Ananth: ]] One of the many things I take from your excellent, objective analysis' are what does "getting the credit you deserve" than mean anyway?

    And my speculative answer is its as much a popularity contest, a measure of whose media is strongest, double standards etc. as it is a measure of actual performance.

    Imran on top - a blow for realistic assessments than. [[ I still cannot get over it. Imran's 10 years are BETTER than SF Barnes's !!! A period 1979-1989 has better bowling figures than a 1907-1914 period. I think time for all of us to stop and think. Ananth: ]]

  • Vyasa on September 15, 2012, 10:58 GMT

    Anant, Fabulous work. I'm happy with you giving equal weightage to bowlers and batsmen. My 2 cents:

    1) 10 years is a long time in cricket. It would encompass 2-3 series against all countries, home and away. By applying a cutoff of number of tests/ runs/ wickets, one could get an estimate of the number of years a person would take to achieve those numbers. The objective of such an exercise is two fold- one, the modern day has a more packed calendar. This would highlight (although a bias towards run making and wicket taking per year already shows it) it better.

    2) If the 10 year window could be shortened to say, 7 years (arbitrary) or so, one could search to see how long individuals have been productive by using the same measures (no overlap between tests allowed). One can get a measure of longevity as well.

    3) Is is a coincidence that in decades where the bowlers were good, the top batsmen did not show up those numbers and vice versa? What if we look at say, 50+ players?

    Vyasa. [[ I think the 10-year period is a must because it carries with itself the weight of a long period. And I can show you 7 year periods during which players played fewer than 25 Tests. Then the wholed analysis would be diluted. As I normally do, I have set the bar higher than warranted. We can separate the very good players from the great players. And the word great will sit well. Ananth: ]]

  • No featured comments at the moment.

  • Vyasa on September 15, 2012, 10:58 GMT

    Anant, Fabulous work. I'm happy with you giving equal weightage to bowlers and batsmen. My 2 cents:

    1) 10 years is a long time in cricket. It would encompass 2-3 series against all countries, home and away. By applying a cutoff of number of tests/ runs/ wickets, one could get an estimate of the number of years a person would take to achieve those numbers. The objective of such an exercise is two fold- one, the modern day has a more packed calendar. This would highlight (although a bias towards run making and wicket taking per year already shows it) it better.

    2) If the 10 year window could be shortened to say, 7 years (arbitrary) or so, one could search to see how long individuals have been productive by using the same measures (no overlap between tests allowed). One can get a measure of longevity as well.

    3) Is is a coincidence that in decades where the bowlers were good, the top batsmen did not show up those numbers and vice versa? What if we look at say, 50+ players?

    Vyasa. [[ I think the 10-year period is a must because it carries with itself the weight of a long period. And I can show you 7 year periods during which players played fewer than 25 Tests. Then the wholed analysis would be diluted. As I normally do, I have set the bar higher than warranted. We can separate the very good players from the great players. And the word great will sit well. Ananth: ]]

  • Waspsting on September 15, 2012, 11:41 GMT

    "There is a real, real surprise at the top. This should make people sit up and take notice... Who could have imagined that Imran Khan would have leapt over many other fancied bowlers? "

    Right here, me. The only guy i thought might edge him was Barnes. (already knew Marshall's peak didn't last a full 10 years)

    I've said it before, i'll say it again, Pakistani players almost never get the credit they deserve. Miandad, Imran, Waqar, Mohammed Yousuf - even Wasim Bari. Wasim Akram is the exception.

    (We have a beautiful example of this with the recent story of Saeed Ajmal - BEYOND RIDICULOUS whats happened there). [[ Yes, I concur with you all the way. No player can hold a candle to Saeed Ajmal's achievements, ACROSS FORMATS. Ananth: ]] One of the many things I take from your excellent, objective analysis' are what does "getting the credit you deserve" than mean anyway?

    And my speculative answer is its as much a popularity contest, a measure of whose media is strongest, double standards etc. as it is a measure of actual performance.

    Imran on top - a blow for realistic assessments than. [[ I still cannot get over it. Imran's 10 years are BETTER than SF Barnes's !!! A period 1979-1989 has better bowling figures than a 1907-1914 period. I think time for all of us to stop and think. Ananth: ]]

  • Gerry_the_Merry on September 15, 2012, 11:47 GMT

    Elegant, Ananth. Imran Khan is indeed the Superstar here. Not only for the reason you have highlighted, but in the batting tables, though with a time lag, and with only 2378 runs, he delivered 51.63.

    However, as expected, his best average period lags his highest productivity period by 1 - 2 years...by 1989 he was not bowling a whole lot.

    Very surprised to see that Brian Lara's best average 10 years were enough to put him below Inzamam and Hanif Mohd...! But #4 by runs scored is certainly as expected. [[ Lara's curse, not that he would care at all, is the number of not outs. A 17-year career with 6 not outs translates to a 10-year period 4 not outs. He would go way up on a RpT or RpI basis. But I do not really want to take off Batting Average for this analysis. Ananth: ]] DCS Compton almost makes it to the top 20, but I suppose you have not built in home/away etc. [[ Home/Away splits would be red herrings. These are clearly defined career-segments and should stay that way. Ananth: ]] Ponting and Dravid take the top 2 positions by runs scored in the averages table also, making their achievement creditable.

    By productivity Kallis makes it to the top 50 bowlers also in addition to being #2 on runs.

    But hats off to GC Smith - 8th on productivity as a batsman alone, besides being captain in almost 100% of the tests played.

  • Ali Shah on September 15, 2012, 12:55 GMT

    Great article Ananth. Imran was a beast at his peak in the 1980s. If I remember correctly he has the highest peak of all post war bowlers. And it is a testament to Imran's skill and persistence that he was still so very good when he came back from a career threatening injury. Imran's bowling achievements get forgotten probably because he was one of the greatest captains the game ever saw and that is how he is usually remembered for Pakistan got very good bowlers even later on but never really found such a captain as he was. I really wish he will be Pakistan's next Prime Minister and usher in an era like that of his cricket captaincy. [[ What is remarkable is that imran is ahead of the pair of giants who played after him. I also wish Imran leads his party to victory as he led his team in 1992. Pakistan badly neads a decade of peace and quiet and a strong person at the helm. Ananth: ]]

  • Zulfiqar Ahmed on September 15, 2012, 12:59 GMT

    Very critical effort,you have done a wonderful exercise, really extracted the essence of greatness and summed it up in numbers the cricket belongs to. Being a fan of Indian team though it feels hollow not to see any bowler representing the sub 25 adjusted bowling average table.Nonetheless, batting giants are always there to recuperate this limitation.Bravo. [[ These are facts, Zulfiqar. The best Indian bowling average of 28.7 of Bedi is 74th in the Bowling average table and there are 8 Pakistani bowlers ahead of Bedi. India has always been a batsmen-centric team. Ananth: ]]

  • Hemant Brar on September 15, 2012, 13:41 GMT

    Well for me it is not that big a surprise to see Imran at the top. I was expecting him to be definitely in the top 3.

    I always rated him just a little below Sobers as an all rounder. In the last decade of his career (which is more or less his best decade), he averaged around 18 with ball and around 50 with bat.

    I wanted to check where does he stand in the Batsman Best Decade List and found he is at number 44 (Lara at 36, Zaheer Abbas, Asian Bradman, 43, Gooch 47, de Silva 53, Anwar, Sehwag, Laxman, Azhar well below him in the list).

    And he was a great captain/leader. And then add to it what he has done as a human, cancer hospital. Role Model.

  • Pankaj Joshi on September 15, 2012, 13:54 GMT

    Not taken time to digest article but ...Oh dear ... Milind's comment was so telling. It is on record that the Nobel committee collectively kicked themselves in the backside for not awarding Gandhi in 1947. Imran's presentation is phenomenal Ananth, thanks for the a(h)nalysis - and to think that we two generations have lived our lives hearing all about the great Botham and the great Hadlee. Truly twas bliss to have a white skin. Well at least the legacy is self-explanatory. For all three cases. QED

  • Entertainment on September 15, 2012, 14:10 GMT

    Sports is purely about entertainment, numbers come distant second. Due to this addiction of stats players become selfish, put self above team, old timers carry on and on. Players should be remembered by the memories they left behind not numbers. After his retirement a certain individual rightly asked "Have I entertained"? I think every cricketer should ask this question to themselves if they want sports to survive. [[ Not sure what you want to say. You must realize that you have come into a Cricket Analysis blogspace. Surely this is what you should expect to find. Why complain. If you find that you have walked into the (wrong) analysis zone, you should quietly exit and move to the entertainment zone. Also please tell me what memories of Imran Khan or Sobers, a 25-year old youth of today would have. Ananth: ]]

  • KIJOHI on September 15, 2012, 15:09 GMT

    Personally, as an Australian, it's great to see Sir Richard Hadlee recognised as the masterful bowler that he was. One can only dream of the world XI teams built from such analyses.

  • unni on September 15, 2012, 15:23 GMT

    Grant analysis, Ananth. I would count this as one of your best effort. Just because, 10 years would define an average career for a good player. Except that number you didn't use any other hard-numbers. The usage of the average of all players for peer analysis is a very nice approach. This brings in a nice number, which is derived from the available data itself. Technically, the first problem would have vanished if you had defined the problem as the best 100 (or n) consecutive tests for a player. But, I agree that this would sacrifice the aesthetic beauty of a human understandable block of 10 years. [[ 100 is too high a number when you realize that Weekes and Walcott played less than 50 Tests. 10 years seems to be a long enough period to hang our hat on. It takes care of almost any good player. Clem Hill is a good example. Ananth: ]]