A few years ago Milind Pandit and I worked together on a concept for a non-contextual, objective, nuanced player-contribution analysis. The cornerstone of that analysis was that we would use only the basic numbers available on the scorecards, but it went beyond the plain-vanilla "percentage of runs scored" or "percentage of wickets taken" measures.
I must credit Milind for laying down the clear framework for this analysis. Recently I have made a few minor tweaks to it, but in essence it remains true to what we envisaged back then.
First, a diversion. I have been working on a concept, "slack runs", for some time now, which I will explain briefly here. Let me first start with a summary of all results in Test cricket.
A team that wins by a certain number of runs has a buffer of that number minus one runs. In other words, a team that wins by 100 runs has scored at least 99 runs in excess of their need. Theoretically, if this team had scored 95 runs fewer, they would have still won, albeit by five runs. These 99 runs are what I call slack runs. The teams that win by an innings and 100 runs has 100 runs to spare - and an innings. They could have scored 95 runs fewer and still won by an innings. These 100 runs are slack runs. The above table explains this concept very nicely.
The teams that won by runs, scored 608 runs in their two innings on average, and their average win was by 166 runs - which is a slack runs percentage of 27.2. The median win was by 158 runs, very close to the mean.
The teams that won by an innings, scored 483 runs in their single innings on average and their average win was by 97 runs, which is 20.1% slack runs. The median was a win by an innings and 78 runs.
I define slack runs as runs that are in the bank, so to say, but are not really needed. They are, simplistically speaking, superfluous. However, the winning team's batters have scored these runs, so we cannot question the validity of these slack runs, only the value.
Now we come to the bowlers. Are there slack wickets? Not at all. The reason is that it is the most fundamental concept of the game that a team, in order to win, has to take all 20 wickets. Barring declarations (in any case, their own choice), injuries and the two odd matches that are excluded from all serious analysis (see footnote in table above), the winning teams have taken all 20 wickets.
The bottom line is that if you take 19 wickets, you do not win. You have to take that last wicket. Fifteen teams, including Pakistan in Kingston last month, have experienced that heartbreak of losing by a wicket. So it is clear that there is no concept of slack wickets. Every wicket is important. I will come back to this concept later in the article.
The value ranges that are provided for the third and fourth innings in the table above are intended to give an idea of the range of scores achieved by the winning and losing teams. For instance, New Zealand scored 451 runs in a final innings and still lost. West Indies scored 418 in a last innings and won, and on another occasion they scored 463 and lost by an innings. Imagine India scoring 510 in the third innings, and losing. And finally, look at the high 650-plus scores in the third and fourth innings for drawn matches, scored by New Zealand and England respectively. A salute to the heroes of these matches - Nathan Astle, Brendon McCullum, Ramnaresh Sarwan, Shivnarine Chanderpaul, MAK Pataudi, Bill Edrich and others.
Test Contribution Analysis
This is determined using a top-down approach. The five key steps are:
1. Depending on the result and all related numbers, allot a maximum of 100 points for the match
2. Distribute this value between the two teams
3. Distribute the team total among three functions - batting, bowling and fielding
4. Distribute the function value between the two innings
5. Distribute the innings-function value among the players
At the top level, the points for the match are allocated. The whole concept is based on the fact that the two tied matches are complete in every way and should be allotted 100 points. All results, barring the Centurion Test of 2000 and The Oval forfeiture of 2006, are allotted 100 points.
Draws are allotted points depending on the extent of completion. The 2008-09 West Indies-England Test at North Sound, in which ten balls were bowled, is allotted 0.38 points. At the other end, the highest allotment for a drawn match is for the India-West Indies Test in Mumbai in 2011, in which both teams finished level on scores with India nine down. One more ball, and we could have had the third tied match in history, or an India win. The total points allotted here are 99.9.
The final points allocated depend on how many innings are completed, what the status of the unfinished innings is, and how far away from a result we end. A match such as the England-South Africa Test at Old Trafford in 1998 presents a real problem. The scores: South Africa 552 for 5 declared, England 183 and 369 for 9. It is possible that the match could have ended in exactly two more balls. However, there was also a possibility of a last-wicket stand of 150 and a fourth-innings collapse of 120. We work on the basis that this match was going to finish quickly, if it was extended. The total points allocated are 98.35. A surface-level conclusion based on the fact that not even the third innings was completed will not work. Hence, a nuanced conclusion depicting the real status of the match. Au contraire, a run-of-the-mill three-innings-Test with scores like 327, 491 and 263 for 4 is assigned a total of 63 points. So we make a deep study of the status of the match in draws.
The farcical two-innings contest between South Africa and England in Centurion in 2000 is allotted a notional two-innings-completion figure of 50.0 points (South Africa, 23.8 points and England, 26.2). It is fair. If nothing else, the match does not get full points. The concession of the match by Pakistan in 2006 is treated differently. It is considered the equivalent of a draw with the prevailing scores - England 173, Pakistan 504, England 298 for 4. A total of around 72.5 points are allotted (Pakistan 44.9 points, and England 27.6). Once we decide this way, it does not matter to us that Pakistan got more points. England's win was not a win on the field. As far as we are concerned, the match just stopped dead - that is all.
In the next level of points allocation, the two teams are awarded points out of the match points. The two teams in the two tied matches are allotted 50.0 points each. For the others, it is a complex calculation involving the resources used by the teams and how much they had in the tank when the match ended. This is briefly explained below. These points are called Team Performance Points (TPPs).
Innings wins: The winning team is allotted 75.0 + x points. X depends on the margin of the win and the winning team's innings score, and is allotted out of 25 points. Two identical innings wins will not produce the same number of points. For instance, in the Pakistan-Australia Test in Sharjah in 2002-03, the scores were 59, 310, 53. This win by an innings and 198 runs fetched 90.6 points for Australia. In the India-New Zealand Test in Nagpur in 2010, the scores were 193, 566, 175. This too was a win by an innings and 198 runs. India get 82.0 points for this game. The difference is due to the fact that the 198-run margin was out of two different innings scores - 310 and 566.
Run wins: This is similar to the innings win, except that the winning team is allotted 50.0 + y points. Y too depends on the margin of the win and the winning team's two innings scores, and is allotted out of 50 points. Thus it is possible that a run win might be allotted more points than an innings win. Two identical run wins will not produce the same number of points. Three wins by 72 runs, with innings totals of 228, 215, 123, 64; 144, 307, 391, 156; and 330, 212, 294, 340, produce three varying points allocations of 67.0, 60.3 and 56.4 respectively. This is because the winning teams' aggregate scores are 351, 535 and 624 respectively. The 72-run margin forms differing shares of the aggregate scores. The way the two result types are computed, it is possible for a huge run win to get more points than an innings win. It so happens that the highest allocation is for a run win.
Wicket wins: The basis is similar. The winning team is allotted 50.0 + z points. Z depends on the number of wickets lost and the target that was reached, and is allotted out of 25 points. However, the only certainty in this case is that because all four innings have been played, the winning team's point allocation will be below 75.0. In Sri Lanka's ten-wicket win against Zimbabwe last year, with scores of 358, 515, 170, 14 for 0, Sri Lanka got 74.9 points. At the other end, in Australia's one-wicket win against West Indies in 1951-52, they secured 50.5 points. The scores in that Test were 272, 216, 203, 260 for 9.
In all these cases, the losing teams get 100 points minus the winning teams' points.
Draws: These are far more complex, since we need to analyse the situation and decide which team was ahead at the time of the end of the match. This requires projecting the match beyond the actual end, especially for fourth-innings situations. The projected scores depend on the number of wickets that have fallen and the match scores. Here is the points allocation for a few draws, along with the scorelines, below.
- West Indies vs England, 1973-74: Eng 395, WI 596, Eng 277 for 7. WI 67.8, Eng 23.3
- Australia vs England, 1994-95: Eng 309, Aus 116, Eng 255 for 2, Aus 344 for 7. Eng 51.6, Aus 44.2
- Australia vs West Indies, 2015-16: WI 330, Aus 176 for 2d. WI 17.3, Aus 21.0
- Australia vs West Indies, 1968-69: WI 276, Aus 533, WI 616, Aus 339 for 9. WI 49.5, Aus 49.4
- West Indies vs India, 2006: Ind 241, WI 371, Ind 521 for 6d, WI 298 for 9. Ind 53.0, WI 45.9
- Bangladesh vs South Africa, 2015: Ban 246 for 8. Ban 13.4, SA 9.15
- Australia vs India, 2020-21: Aus 338, Ind 244, Aus 312 for 6d, Ind 334 for 5. Aus: 49.0, Ind 46.5
- England vs Australia, 1956: Eng 247, Aus 202, Eng 182 for 3d, Aus 27 for 5. Eng 62.7, Aus 28.3
Now we move on to the allocation of team points to the functions. This is, by far, the simplest of allocations. But first let us look back at the concept of slack runs. It is clear that well over 20% of the runs scored in all Test matches in decisive matches can be termed as slack runs. In addition, there are normally seven to eight batters who score most of the runs - as against five to six bowlers who take wickets. Considering this, there is a higher allocation to the batting function (55 for batting, 45 for bowling). The only exception is in innings wins. Because only one innings has been played for the winning team, and there have been two bowling innings for the winning bowlers, the allocation here is 50-50. Also, in cases of huge wins, by eight, nine or ten wickets, the bowling allocation is slightly higher than 45%. The explanations for these exact splits are too complex to add in this piece.
In the next step, the functional allotments are made to the two innings - 1/2 and 3/4. This is also a fairly straightforward allocation based on the runs scored and wickets taken.
In the last stage of allocation, the innings-functional points are allocated to up to 12 players. This is the most complicated allocation. The fielding allocation is straightforward, based on stumpings made and catches taken.
For a batter, the allocation is based on the runs scored, balls faced (actual or extrapolated), and the scoring rate. It is recognised that occupying the crease is as important as scoring quickly. In general, these two factors nullify each other and the net effect is not very significant, barring special cases.
Points for a bowler are allocated based on the wickets taken, balls bowled, and runs conceded. Importance is given not just to the wickets but the basic task of bowling tight overs. We do not just reward the bowler who took 7 for 55 but also the ones who took 2 for 31 and 1 for 24, depending on how well they bowled. Thus it is possible for spells such as 4 for 50 and 3 for 30 to be given equal points. But the main emphasis is on taking wickets.
However, the one thing that should not be forgotten is that context does not come in. My Player Performance Analysis systems (ranking the best batting and bowling performances) are contextual. The Contribution Analysis, on the other hand, is non-contextual. Only the five basic measures - runs scored, balls faced, wickets captured, balls bowled and runs conceded are used. Nothing more.
After having used these values for many years, I have come to the conclusion that the important values are the top two levels (match points and TPPs) and the last level (player contribution points). The middle two levels are only conduits for determining these values.
Let us move to the tables. There are two team-related tables and one each for batting, bowling and all-round performances.
This is a table of the high TPPs. The highest TPP received was for England's 675-run win in Brisbane in 1928. Australia were set a ridiculous target of 742 and could manage only 66. This despite the presence of Don Bradman, playing his first Test. Maybe this loss set Bradman on his historic career path. The TPP-line was 94.18-5.82.
The second-highest TPP allotted is for an innings win by England in Cape Town in 1889. Though this was not a massive innings win, the fact that England scored only 292 and dismissed South Africa for 47 and 43 raised the value.
The third-placed win is New Zealand's innings victory in Harare in 2005 - they scored 452 for 9 and then dismissed Zimbabwe for 59 and 99.
The fourth-placed match is the 1938 Oval Test - the biggest innings win in the history of Test cricket, but because Australia managed to score 201 and 123 in their two innings, England's points are reduced slightly. Bradman did not bat in either innings; 90.73 points are the reward for England for their win by a mile and some.
The next match is of more recent vintage. Australia steamrolled Pakistan in Sharjah in 2002 with a scoreline that resembled that in an 1880s match. The scores read 59, 310 and 53. Matthew Hayden single-handedly scored more than all of Pakistan's batters did in two innings.
The sixth match with a TPP allocation exceeding 90 was well the 1896 Port Elizabeth Test. Though they made scores of only 185 and 226, England managed to win by 288 runs - dismissing a hapless South Africa for 93 and 30.
The other nine matches had winning TPPs ranging from 89.56 to 87.65. Thirty-seven teams have TPPs above 85, and 156 teams have 80 or higher; 518 teams have 75 or more, 94 of these being wins by runs, the other 424 won by an innings.
The next table is an interesting one. This one features matches in which the two teams got TPPs as close to 50 as possible.
We start with the two tied Tests - the only ones in which the two teams got exactly 50 points each. There is one significant difference between the two tied matches. The 1960 Brisbane match between Australia and West Indies was the most perfect match ever played. All 40 wickets were taken and the teams ended equal on runs. On the other hand, the 1986-87 Chennai tie between India and Australia was a little imperfect. Australia lost only 12 wickets, so there were resources left unused. Undoubtedly a tie, but not a perfect one.
The match in third place is that magnificent draw in 1969 in Adelaide. At the end, Australia were 339 for 9, needing 20 runs to tie with West Indies. The high match RpW of 84.8 ensured the match had a 99.93 total value, with West Indies holding a razor-thin edge. Undoubtedly, one of the all-time great draws.
In 1987, Australia and New Zealand produced a wonderful and pulsating draw in Melbourne. New Zealand's 317 was matched comfortably by Australia, who secured a useful 40-run lead. New Zealand set Australia 247 to win and at the close, Australia were hanging on grimly at 230 for 9. The last-wicket stand occupied over seven overs.
Next is the unforgettable Headingley Ashes classic of 2019. Despite being dismissed for 67 in their first innings, England managed to carve out a terrific win by one wicket, after being 286 for 9 and 73 runs behind. Ben Stokes and Jack Leach created a win for the ages. The closeness of the contest leads to a 50.17-49.83 TPP allocation.
The sixth match featured is fit to be bracketed with the Brisbane tie among the greatest Test match ever. How do Australia and West Indies produce such classics? A low-scoring match meant Australia needed to score only 186 for a win. But Curtly Ambrose and Courtney Walsh were devastating and reduced Australia to 102 for 8 and 144 for 9. Then Tim May and Craig McDermott lasted for more than 20 overs and took Australia to within two runs of the target, only for Walsh to dismiss McDermott when they were a run short. This match was equally close - 50.18 against 49.82.
Nine other matches are displayed on the table. The TPP tallies in each differ by only one point, at most. Barring a two-run and a three-run win, the other matches were wins by a single wicket. It is clear that these were all knife-edge encounters and the TPP scoreline of 50.xx - 49.yy could easily have gone the other way. Andrew Flintoff, Huge Trumble, Dave Nourse, Kusal Perera, Inzamam-ul-Haq (twice), Gilbert Jessop, Lindsay Hassett and Mahela Jayawardene were the heroes in these games.
I will now move to the individual players' contribution patterns. Common sense tells us that the player-contribution tables, if ordered on points secured, will be dominated by wins - and ones by comfortable margins, since the team points will be over 70 and there are fewer players to share the spoils. We are not interested in looking only at these huge wins. We need to know which players contributed significantly in all types of matches. So I will look at the percentage of contribution made by the player out of his team's relevant points rather than the absolute value of contribution points.
First, the batter contributions to a match. The cut-off is nine contribution points for a reason - to get in the best performance in a losing match.
The first six performances are in drawn matches, but where all the games reached a decent stage of the third innings. Seymour Nurse scored 258 out of 417 in Christchurch in 1969 and secured 11.6 contribution points. West Indies received 38.1 points for this game since they had the edge. Nurse's contribution stood at 63.5%.
Dave Houghton's 266 against Sri Lanka in Bulawayo in 1994 was in of a total of 462, and he exceeded 60%, as did Dileep Sardesai, whose 212 was out of India's 387 in Kingston in 1971. Don Bradman and Chris Gayle, with dominating triple-hundreds, round off the top six.
The highest contribution in a win was that of John Edrich, whose unbeaten 310 out of 546 against New Zealand at Headingley in 1965 netted him 23.5 contribution points and over 58% of the team's share. Among wins, Edrich is on top because of a combination of a big innings win (more TPPs) and a high contribution to the innings.
Upul Tharanga holds the next position for his performance in a ten-wicket win for Sri Lanka in Bogra in 2005-06. That he participated in a huge opening partnership in the second innings helped Tharanga. He scored 165 out of 316 and 71 not out from 120 - nearly 57% of his team's batting output.
A notable performance from among the rest is that of Brian Lara, whose 221 and 130 in a hopeless cause in Colombo in 2001 got him only 9.7 points. However, West Indies could secure only 17.2 points and Lara's contribution is over 56% of this. This is the highest contribution by a batter in a losing cause.
Individual bowlers' share of the team contributions works very differently to the batters' shares. A bowler can take all 20 wickets - as Jim Laker nearly did. So the percentage share values could be very high.
Laker's never-again performance of 19 wickets in Manchester against Australia in 1956 is in second place, with 80.8%. In this case, the fact that the other bowlers bowled lots of overs, sharing the load, pulled Laker's score down a little.
The highest share in a lost match: in Bombay in 1966, against a strong West Indies side, Bhagwath Chandrasekhar bowled his heart out and took 11 wickets - all to no avail, since West Indies won comfortably.
Nick Cook's 11 for 83 in Karachi in 1984 is also featured. In the second innings, Cook bowled a great spell of left-arm spin for 5 for 18 while defending 64. Pakistan almost lost the match, conceding seven wickets to get to the low target. This was one of the greatest performances in a losing cause.
When I compiled the Wisden 100 list 20 years ago, I had Bradman and Hugh Tayfield at the top of my tables. When I revisited that theme a couple of years back, these two great performances moved into third places and Graham Gooch and Richard Hadlee took over. Then a year later, Kusal Perera got to the top. However, one thing is certain. Whether I use the Contribution Analysis, the Performance Ratings analysis, a plain-vanilla percentage of runs/wickets, or even a poll of readers, one performance will always stand at the top of the all-round performance table. That is Ian Botham's tour de force in Bombay in 1980. Whether I go by the total points or a percentage of team points, that stays at the top here as well.
Let us look at the performance - it is one for the ages: 6 for 58 to dismiss India for 242. Then a hundred, coming in at 57 for 4 and giving England a small lead of 50. Then another magnificent spell of 7 for 48, dismissing India for 149. The only thing Botham did not do was open for England in the last innings. His total of 35.3 points is a record, as is the 48.7% of England's TPP value of 74.3. Let us take a moment to admire this single incandescent streak of brilliance, never surpassed before, and possibly not after either.
With the lowering of the cut-off level, the second-best performance is in a lost match. In Bridgetown in 2018 against Sri Lanka, Jason Holder, the West Indies captain, stood resolutely on the burning deck, scoring 74 out of 204 and then took 4 for 19 to dismiss Sri Lanka for 154. Then he scored 15 out of 93. In the last innings, when Sri Lanka chased a modest total of 144 successfully, Holder took five of the six wickets that fell. This was nearly as great a performance as Botham's. Holder's total was 18.7 points. Out of the losing team's share of 44.7, this was an impressive 42.0%.
The third performance is another all-time classic - this time by Richard Hadlee. In Christchurch in 1984 against England, Hadlee scored a blistering 99 in 81 balls out of the New Zealand score of 307. Then he fashioned an innings win for his team, taking 3 for 16 and 5 for 28, keeping England's totals to below 100 each. The huge win, resulting in the 82.4-point allocation, also helped Hadlee.
Then comes another well-known and memorable performance, by Mushtaq Mohammad. He scored 201 out of a huge Pakistani total of 507 and then took seven wickets in Dunedin in 1973. He was a fraction behind Bates.
Aubrey Faulkner was the shining star in a win for South Africa by 19 runs in Johannesburg in 1910. He scored 78 out of 208 and 123 out of 345 and took 5 for 120 and 3 for 40 - a contribution% of around 37.5.
The last entry in the table is the famous Botham show at Headingley in 1981: he made 50 out of 174, 149 not out out of 356, supported by 6 for 95 and finally 1 for 14. It is appropriate that this table is bookended by two Botham classics.
In this table, I have looked at the contributions made by players across their careers, as a percentage of the total team points. A minimum of 20 Tests is the cut-off.
It is not a surprise that this table is dominated by allrounders from the weaker teams, and dominant bowlers. Vinoo Mankad contributed 20% of India's total points. He is closely followed by Hadlee and Shakib Al Hasan. Then come Barnes and Murali, who averaged around seven wickets per Test. They are followed by a slew of allrounders and a few bowlers. This sequence is not broken for quite a number of entries. The batter with the highest contribution percentage is, no surprise, Bradman, who contributed 13.5% of his team's total points. He is, understandably, placed only at 65th position.
An interesting nugget from this table is the average points per Test of the teams. That represents the success quotient of the teams. Note the high values for Bradman's Australia, Clarrie Grimmett's Australia (almost overlapping) and R Ashwin's India. The only other value exceeding 50 is that of England with Barnes in their ranks. Note the sub-40/30 values for Mankad's India, Fazal Mahmood's Pakistan, and recent Bangladesh teams.
In the Lord's Test of the current England-India series, Mohammed Siraj secured 9.2 points, Ishant Sharma 8.8, KL Rahul 8.3, Mohammed Shami 8.1, Jasprit Bumrah 6.9, and Rohit Sharma 6.4. These are complete match points, including batting, bowling and fielding. These reflect the player contributions precisely. It is also clear that Siraj should have been given the Player-of-the-Match award for his top-order wickets in the first innings and breakthrough wickets in the second innings. For England, Joe Root contributed 8.9 points. No one else reached five points.
In the next Test, at Headingley, Craig Overton got 14.5 points, Ollie Robinson 11.4, Root 11.2, and James Anderson 10.3. The top three batters got around six points each. Overton makes a good case for the Player-of-the-Match award (six wickets and a quick 32), but I have no problems with the selection of Robinson for his incisive spell on the fourth morning. For India, their low TPP of 21.8 points meant that no player even reached four points.
Finally, at The Oval, the Player of the Match was again wrongly awarded; this time the lapse was more pronounced than Rahul getting it at Lord's. Shardul Thakur received 11.3 points, Umesh Yadav 9.5, Ravindra Jadeja 9.3, Bumrah 8.7 and Rohit Sharma 7.7. The reasons are clearly evident on the scorecard. Important scores and key wickets. There was no doubt Rohit's was a valuable innings but it was not as effective as many other contributions. A score of 180-200 out of a 400-plus total would have justified the selection.
In summary, the top two levels (match and team points) and the last level (player points) are the important measures used in various other analyses. The middle two levels are just intermediate steps.
Calling for an all-time XV
In 2013, I conducted an exhaustive readers' poll to determine a group of 15 players to be considered for an all-time World team. There was excellent response and the results were very insightful and interesting. I now call for submissions again since new contenders have emerged, as also new measures for selection. You can mail your entries through one of three routes with the subject: "All-time XV - 2021".
- Send an email to my personal mail id, if you have it
- Send an email to the email id at the bottom of this article
- Send an email to the Talking_Cricket group, more on which is below.
When sending in your XV, provide your name, place of residence, and the list of 15 players (no more, no less). The team must be an all-terrain one. A manager/coach is optional. If you send multiple entries from one email id, I will consider the last one sent. Thus, you have the opportunity to change your selections. You don't have to justify your selections; I prefer short emails. What is important is to make your selections clear. I will include this message in the next three articles. Then, I will do a summary article, probably in January. The entry that matches the final selection or comes closest to it will be acknowledged and rewarded.
- eight batters/allrounders
- one wicketkeeper
- four pace bowlers
- two spinners
Talking Cricket Group
Any reader who wishes to join the general purpose cricket-ideas-exchange group of this name that I started last year can email a request for inclusion, giving their name, place of residence and what they do.
Email me your comments and I will respond. This email id is to be used only for sending in comments. Please note that readers whose emails are derogatory to the author or any player will be permanently blocked from sending any feedback in future.
Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems