The Greatest Test all-rounder: Who wins the numbers race?

When Ben Stokes called time on his international career in June, it was hailed in some quarters as the final act by one of the greatest all-rounders of all time. Such weighty claims did not seem entirely unreasonable. After all, Stokes is one of only three cricketers with a giant double of 7000 runs and 200 wickets in nearly 150 years of Test cricket. Sir Garfield Sobers and Jacques Kallis, the other two in this elite club, have long been the top contenders in the perennial “greatest all-rounder” debate. But what about the rest?

Stokes’ calibre is also highlighted by the fact that only one all-rounder comes close to matching his pedigree among active cricketers. Look at this list and you’ll see that Ravindra Jadeja, who turns 38 in December, tops the Test all-rounders rankings currently, and not many in that top 15 would match the definition of a genuine one. The ones who do, haven’t played enough as yet.

But when it comes to determining the best all-rounder ever from a statistical perspective, how would you do it in a truly holistic manner? In this study, we aim to narrow down the top all-rounders in Test cricket history by analysing their overall performances, and how they stack up compared to their teammates and other contemporaries in either discipline.

But we’ve also recognised that the ‘best all-rounder’ debate is linked to what makes a great player. This includes questions such as: what does this player mean to his team, and to his team’s success? How consistently good are they - are they remembered for shining brightly in a couple of high-profile Tests, thus skewing their ability in our imagination? Or did they put in high-quality performances with bat and ball across a large portion of their career? Were they able to perform in different conditions, or are their numbers bolstered by having wild success at home?

The metrics we have used to determine the best all-rounder of all time are laid out in this first article of a two-part series, along with the results. Further explanation, along with insights and answers to some of the questions that this might throw up for readers, will follow in Part Two. Let’s dive in.

To narrow things down, we have started with a sizable cut-off - we are only considering players with a double of 2000 runs and 100 wickets in Tests. 36 players have achieved this to date. Of those, we have eliminated six who have crossed the threshold in their secondary discipline only because of their longevity. The ones who average under 20 with the bat - Shane Warne, Anil Kumble, Stuart Broad, Harbhajan Singh and Tim Southee - and the only one with a 45-plus bowling average, Carl Hooper, have been discarded, narrowing it down to a list of 30.

An all-rounder would typically bat in the top eight of a line-up and serve as one of the first six bowling options as a bare minimum. For context, and a better understanding of the above filtration, the batting average for No. 8 across 2634 Tests played to date is 21.2. Likewise, the sixth bowler in the lineup has averaged 43.6.

The notable omissions other than the players mentioned above are South Africa’s Aubrey Faulkner and Australia’s Monty Noble and Warwick Armstrong - all from the early 20th century. They had stellar all-round numbers, but narrowly missed the above cut-off. Mushtaq Mohammad, Lance Klusener and Brian McMillan are a few other significant exclusions due to shorter career spans. While this may seem unfair to the earlier players who might have had limited game time due to circumstances outside of their control, unfortunately we will never know whether they could have maintained their impressive records across a longer time span.

A generic way to rank all-rounders is often the difference between the batting and bowling averages, as reflected in the table above. However, that might not necessarily substantiate enough. A genuine all-rounder would, by definition, walk into most sides given his expertise and excellence in his primary discipline. Then comes how impactful he is with the other discipline, and how the two skillsets unite in terms of the overall impact.

For the broader parameters, the percentile for each of the sub-categories is calculated, and then the mean of those gives us a final leaderboard with respect to all the players in the mentioned cut-off. For example, for volume, the first broad category, the mean percentiles of the total runs scored and wickets taken have been calculated with respect to all the 3281 Test cricketers to date. Sachin Tendulkar (15921) and Muttiah Muralitharan (800 wickets) top the respective charts. Kallis, who is the fourth-highest run-scorer in Tests (percentile: 83.47) and bagged 292 wickets (percentile: 36.5), tops the overall list here out of the 3281 players courtesy of his stellar double. By adding his two percentiles together and dividing by two, he comes out with a mean percentile of 59.99. This value ranks the highest among all 3281 players, and grants him a perfect score of 100.

The percentiles in parameters which feature just the 36 players grants a perfect 100 to the best in the set, and zero to the lowest one. The ones at the bottom in every parameter are among one of the six discarded players out of 36, and so barely anyone in the final set of 30 takes a hit in the final calculations.

As explained above, Kallis tops this parameter not just in the shortlist, but overall. Among the nominated players, Ravichandran Ashwin ranks second here (ninth among the 3281 cricketers), a position he claims largely because of his 537 wickets. That gives him a percentile score of 74.29. The next four in the set are the ones with a giant mix of both - Kapil Dev, Ian Botham, Sobers and Stokes.

The cut-off for the exclusive batting and bowling parameters remains 2000 runs and 100 wickets respectively. A grand total of 349 players have scored 2000 or more runs in Tests till date, while 206 players fall in the latter category. At 99.94, Bradman remains the benchmark for batting averages (100 percentile), and Sobers and Kallis occupy the top two positions among the all-rounders.

The percentage of runs scored and wickets taken compared to the rest of the teammates are indicative of a player's overall impact for his team. Bradman tops this list as well, contributing 24.28% of team runs, while Sobers (15.52%) ranks the best among the nominated set (at 33rd overall). This basically separates players who have done the heavy-lifting for their respective teams from the rest. Point in case, Sunil Gavaskar ranks 15th here - the best among Indians - having contributed 16.55% (one-sixth) of India’s runs in a 16-year-long career. Brian Lara, the torch-bearer for West Indies through the 90s and until mid 2000s, ranks third (18.87%).

As for contributions in wins, we factor in the percentage of runs and wickets that a player has contributed to the victories that he has been a part of. Imagine a bowler has featured in 5 wins with his side bagging 20 wickets in all of them, and five of those 100 dismissals are runouts. If the said player has picked 19 wickets out of those 95 (excluding the runouts), his bowling contribution across these matches would be 20% (19 out of 95). Likewise, we calculate the individual batting impact with respect to the team’s combined aggregate across those wins.

The top five here make for an interesting observation - Vinoo Mankad, Hadlee, Sobers, Shakib Al Hasan and Botham. Mankad and Shakib were first-generation greats for their respective countries. Hadlee was instrumental when New Zealand got into winning ways for the first time in their history, with a sharp rise in the Eighties. Until the end of 1979, New Zealand had won just one out of the 45 Test series they had featured in (the 1969 Pakistan tour). Between January 1980 and the end of Hadlee’s career (England tour 1990), they won 10 out of 21 (excluding one-off Tests).

The noteworthy names way down the list in this set are Kapil Dev (22nd) and Stokes (24th). In stark contrast to Hadlee, a fair number of Kapil’s standout match performances came in either draws or defeats. Wins over Pakistan and West Indies in Chennai in 1979 and 1980 respectively were among the notable exceptions, as was the famous tied Test at the same venue in 1986.

As seen in the table at the top of the article, Sobers bags a perfect 100 for the difference in averages here, while Moeen Ali sits at the other end with a percentile of 14.52 (Tim Southee, one of the six discarded players, ranks last).

This is a rather intriguing one. Consider a batting average of minimum 40 and a sub-30 bowling average. Both make for great reading in isolation. In this exercise, we have factored in the longest sequences (most consecutive Tests) wherein a player managed this exact double, starting at any given point of his career. Then, we calculate what percentage these matches constitute of the individual’s overall career appearances. The hot streak is the key here, and the averages are immaterial, for that could be misleading especially in case of lower sample sizes.

Jadeja boasts the longest such streak in Test cricket history, spanning 82 games, starting with the 2014 New Zealand tour until the first Test against South Africa in Kolkata last year. Imran maintained that for 81 Tests, starting with his seventh appearance in 1976 until his penultimate one during the 1991/92 home season against Sri Lanka. Imran’s streak constituted 92.05% of his career appearances (81 out of 88 Tests), and hence he tops this parameter, followed by Jadeja (90.11% - 82 out of 91). Looking at the percentage of a player’s career that he maintained the averages gives those all-rounders who played fewer Tests in older eras an equal chance.

Keith Miller (67.27%), Tony Greig (63.79%) and Vinoo Mankad (61.36%) round off the top five in terms of the percentage of matches when they maintained this high standard. Even if Jadeja, the only active cricketer among them, plays another 30 Tests, he’d still stay at second on this parameter. Kapil could maintain this streak in just 15 straight Tests over a career spanning 131 games, and ranks 27th out of 30, while Shakib takes the seventh spot.

For this, we have compared the batting and bowling averages for both home and away Tests for each of the 36 players in the cut-off i.e. four sub-categories. Then, all players are ranked on the four parameters individually, and a mean percentile for the four gives us the final leaderboard.

Sobers and Kallis averaged 50-plus both in home and away Tests, while Hadlee stands out on the bowling front with an average of 22.96 in New Zealand and 21.72 away. However, it’s Imran again who pips everyone else overall with impressive numbers across the four sub-categories.

Have questions about the methodology or why a particular player ranks higher or lower than you expected? Mail them, along with your name, to submissions@cricbuzz.com and we’ll take up the best questions in the second part of this series, with due credit to the contributors.