Rating Test cricket’s captains: Individual performance

By matth / Roar Guru

In my previous article I developed a method to rank Test cricket’s most successful captains by team results.

In this article I will examine the individual performances of captains in more detail, primarily to see if they thrived or crumbled after assuming the captaincy and all the pressures that go with it.

The Don
The captain with the best raw batting average is, of course, Don Bradman. Bradman’s average as captain was slightly higher than his career average, at 101.51. Truly remarkable of course. However, there are some insights to be gleaned from within this gargantuan statistic.

Bradman averaged 110.25 as captain at home, compared to 85.63 away, which is a pretty standard difference. In the Ashes, Bradman averaged 90.07 as captain. Not too shabby.

Bradman certainly set the tone. In wins as captain he averaged 124 and if he won the toss, his performance in his team’s first innings was 120.77. He was equally as good at the back end of games. His average as captain in his team’s second innings is a ridiculous 157.

However, if Bradman lost the toss his efforts in his team’s first innings was a mortal 49.76. So if the opposition grabbed the best of the conditions, this greatly affected Bradman’s influence, whether Australia were sent in or batted first. So if you were an opposing captain against Bradman, win the toss or go home.

(PA Images via Getty Images)

It is in losses as captain that Bradman becomes truly average. Bradman averaged just 30 in losses, putting him 43rd on the all time list. Of course the thing to remember here is, if Bradman reached 30 then he likely went on and Australia likely won the match.

Statistician Charles Davis once showed that Bradman was as susceptible as any other batsman to getting out early (i.e. within the first ten runs), but once he got in, no batsman was ever harder to remove.

The rest
The big five captains with the next best batting averages are mostly the usual suspects. In descending order after Bradman we have
Steve Smith, Kumar Sangakkara, Kane Williamson, Mahela Jayawardene and Sir Garfield Sobers.

That’s a fairly handy list and all of these players performed better as captains than when playing for someone else. These improvements ranged from 3.1 per cent for Sobers up to 27.6 per cent for Williamson.

Captaining at home
Steve Smith has the best average of this group when captaining on home soil – a Bradman-esque 88.51. Note however that there have been better. New Zealand’s Ross Taylor averaged 97.66 when captaining at home.

Others to exceed Steve Smith’s mark are Pakistan rogue Saleem Malik (92.83), English Bodyline exponent Douglas Jardine (96.25), silky South African Hashim Amla (93.66), the West Indian crab Shivnarine Chanderpaul (90.37) and current Sri Lankan opener Dimuth Karunaratne (96.42).

Leading away
Of the big five, Sobers had the best away average as captain – 59.36. This was actually slightly better than his home average. Only two non-Bradmans have bettered that mark: diminutive Bangladesh wicketkeeper batsman Mushfiqur Rahim (59.65) and 1930s South African skipper Alan Melville (63.22). We will come back to Melville’s remarkable career a little later.

(Photo by S&G/PA Images via Getty Images)

Performing in wins
Of the big five, Smith is also the best in wins (84.2), but there have been plenty even better, including such players as England portly opener Graham Gooch (86.6) and Pakistan’s even more portly leader Inzamam-ul-Haq (94.1).

The winners of the winners (?) though are Shivnarine Chanderpaul and South Africa’s Dudley Nourse, who both averaged well over 200 in their admittedly very infrequent wins as captain. They each won just a single Test, both from well over ten attempts.

Holding it all together in losses
It is Sri Lankan Kumar Sangakkara who takes the title in losses, his mark of 52.16 being almost 20 runs better than any of the other big five. There are only two other captains in history to do better and they are both Australians.

As might be expected Steve Waugh refused to accept defeat and averaged 55.06 in losses. He was just shaded by early 20th century captain Monty Noble, who averaged an amazing 56.57 in losses as captain. This is remarkable given his overall average as captain of 38.15 and as a regular player of just 25.12. Noble was certainly a man for a crisis.

Winning the toss and sticking the boot in
In terms of winning the toss and taking advantage, both Smith (87.85) and Sangakkara (94.28) were masters. Only Pakistan opener Zaheer Abbas (97) and South African 1950s captain Jackie McGlew (87.66) can match those numbers.

Losing the toss but producing the goods
When losing the toss and having to set the tone in less advantageous conditions, then Jayawardene (80.41) and Williamson (70.6) come to the fore, but they are no match for first innings specialist Smith, who averages 94.85 when losing the toss. Two other standouts were Pakistan triple centurion Hanif Mohammad (166) and the post isolation South African leader Kepler Wessels (86.25).

(Photo by Phil Walter/Getty Images)

Third and fourth innings performers
Sobers (68.82) and Sangakkara (76.33) were the third and fourth innings performers, showing their skills against slow bowling. Cricket’s most successful captain, Warwick Armstrong, nearly matched the Don in this category, averaging a whopping 153.5 in the later innings. The only other player to match these three was superb Kiwi batsman Martin Crowe (78.18).

Doing it in the Ashes
Since we are Australians and get a kick out of kicking the English, it’s important to note that there is an Ashes summer on its way and who better to belt the English than Steve Smith, who averages 137.4 in Ashes contests when leading. That is 47 runs better than Bradman! Bob Simpson is the only other player to average more than 57 as a captain in the pressure cooker of an Ashes contest – 81.3.

The pressure of the little urn can get to some captains. Michael Clarke averaged 34.11 as an Ashes captain and Alastair Cook an underwhelming 29.41. Who would have thought that the hapless Glenn McGrath bunny Mike Atherton would have a four-run better Ashes record as captain than Alastair Cook?

Spare a thought for beefy Ian Botham. Botham was a man completely unsuited to captaincy and he averaged a stomach-churning 6.8 as an Ashes captain.

But batting performances are not everything when it comes to defeating the old enemy. Two of England’s greatest Ashes captains were Douglas Jardine and Mike Brearley – and the each averaged a tick over 22.

(Photo by PA Images via Getty Images)

The great improvers
Some players are crushed by the captaincy and others are inspired. I have already mentioned that Kane Williamson is a 27.6 per cent better batsman as captain than before he took the mantle. That’s worth 13.5 runs per innings to his team and they are now world champions.

There have been 30 other batsman to improve by even more as captain. Some of those started from an incredibly low base, for example Indian left-arm spinner Bishan Bedi improved his batting average from 7.09 to a heady 13.36. Others were truly impressive and here are some examples.

Clive Lloyd was arguably his team’s greatest captain. A production line of intimidating speedsters helped, but so did his own contributions increasing from 38.7 to 51.3.

Ian Chappell was another to thrive as leader, with his average increasing from 37.3 to 50. Virat Kohli, another noted competitor, performs 42.5 per cent better as captain, increasing his average from 41.13 to 58.6

English opener Graham Gooch was a whopping 63.4 per cent better as captain, with his average moving from 35.93 to 58.72, and Warwick Armstrong went from 35.66 to 56. This, coupled with his similarly impressive bowling improvement from 35.81 to 24.47, makes Armstrong one of the highest performing captains of all time.

On top of these great captains there are four players who more than doubled their batting output as captain.

The great Pakistan all rounder Imran Khan averaged 25.43 with the bat as a bowling all-rounder before taking on the captaincy. As captain of the most chaotic team in world cricket, Imran averaged 52.34 with the bat, scoring five centuries. In addition he improved his bowling average from 25.53 to 20.26, which is the second best all time bowling average for a captain. This puts Imran Khan up there with Warwick Armstrong as the best all-round performer as captain in history.

(Photo by S&G/PA Images via Getty Images)

Another bowling captain, Zimbabwe’s Heath Streak, improved his batting from an average of 16.01 to a more healthy 36.17, a 126 per cent increase. Unfortunately, given he was probably Zimbabwe’s greatest pace bowler, his bowling average went from 25.81 to 34.8 in the same period.

At the very beginning of Test cricket, Australia’s Billy Murdoch played a couple of Tests as a regular player and averaged just 6.33. He then led his country for much of the next decade, averaging 35.08 and scoring Test cricket’s first double century.

Alan Melville
Finally there is South African captain during the 1930s and 1940s, Alan Melville. Melville started his career by captaining on debut and celebrated with a duck. He was averaging just 12.67 after his first three Tests. But from that point Melville’s batting blossomed, scoring four centuries and two fifties in his next five Tests.

Those performances came in a remarkable run either side of World War Two. It started in the fourth Test of the 1938-39 home series against England when Melville hit 67 in his only innings. In the final Test of the series Melville struck 78 in the first innings and then scored his maiden century in the second, 103. This was the infamous timeless that went on for ten days before being called a draw as England were about to miss their boat home.

An eight-year break ensued due to war, but somehow Melville maintained his form as he led his team to England in 1947. In the first Test at Trent Bridge Melville, now 37 years old, hit a South African Test record 189 in the first innings and then backed it up with 104 not out scored in a single session in the second innings, only the second South African to achieve this feat, while also being the first to hit centuries in each innings of a Test.

Moving to Lord’s Melville scored 117 in the first innings to make it four centuries in a row. Only the great West Indian Everton Weekes ever scored more consecutive hundreds, with five. Australia’s Jack Fingleton and India Rahul Dravid are the only others to match Melville’s four in a row.

(Photo by Ashley Allen/Getty Images)

Melville’s form then went back to where it had come from. From a career to date average of 87.0 he scored just 198 runs from his final nine innings. His final Test was the only one he played without the captaincy and he scored just 15 and 24.

Melville’s career is all the more remarkable given he broke three vertebrae in his back in 1929, before he even represented his country. This was after being denied a place on the 1929 South African tour of England because Melville’s father wanted him to attend Oxford instead. While there Melville captained and starred for the University but also broke his collarbone in a collision with his batting partner and suffered from appendicitis. Nevertheless he went on to captain the Sussex county side before returning to South Africa.

But even so, Melville’s career should have ended in 1939. A recurrence of his back injury while serving with the South African armed forces during the war forced Melville to wear a steel jacket for almost a year. Then, despite recovering and leading his side to England in 1947, on tour just before his four-century streak he broke a bone in his little finger and suffered a thigh strain.

This was one tough captain and he was name a Wisden Cricketer of the Year as a result. The tour took its toll though. By the end Melville was exhausted and had lost 27 pounds. He retired from first class cricket soon after but came back, only to fracture his wrist. Even then he recovered to play one final Test before finally giving the game away.

Highest score by a captain
It is the great West Indian Brian Lara that holds the record for highest score as captain, 400 not out. I guess we know why they held off declaring.

(Photo by Joe Mann/Getty Images)

There have been seven other triple centuries by captains, the most notable is probably from Kiwi Brendon McCullum. New Zealand gave away a 246-run first innings lead to India at Wellington in 2014. McCullum made his way to 302 at an uncharacteristically leisurely pace in the second innings to lead his team to safety.

Going the other way
Captaincy does no favours for some players. The great West Indian leader of tied Test fame, Frank Worrell, averaged 53.13 but this reduced to by 23.5 per cent to 40.42 once he took on the leadership. Some years later, one of his successors was Richie Richardson.

Richardson had the unfortunate luck to captain a West Indies team during its decline from world’s best and Richardson went from a floppy hat wearing, carefree stroke maker averaging 47.9 to a tortured soul managing just 35.18 runs per innings.

Michael Vaughan presided over England’s great rise in the mid 2000s, but it came at the expense of his own batting. Also hampered by chronic injuries, Vaughan went from averaging 50.98 pre captaincy to 36.02 afterwards.

A more recent example is our current captain Tim Paine. Paine averaged 41.66 in pre-sandpaper days but the workload of keeping and captaining during a tumultuous time for Australian cricket has seen this drop to just 28.4. His team did, of course, retain the Ashes in England for the first time since 2001 and Australians tend not to have to twist themselves in knots to like their team, so it’s not all bad.

There are six players whose batting averages halved or worse as captains. Only one played since 1961 and that is Ian Botham. A pre-captain average of 36.74 became a McGrath-esque mark of 13.14 in 21 innings as leader. To make sure of his captaincy legacy, Botham’s bowling record worsened by nearly 20 per cent and England didn’t win a single match under his leadership. A truly disastrous effort.

Bowling captains
There have only been 50 captains to take at least ten wickets while in the role. Plenty of these were batsmen who bowled a bit, such as Viv Richards and Allan Border. Only 25 captains took at least ten wickets at two or more per Test.

Of these 25 the best bowling average belongs to 1950s Pakistan bowler Fazal Mahmood. Mahmood was probably his country’s first great bowler and averaged an excellent 19.14 as captain. This was 29.2 per cent better than his non-captaincy mark of 27.03 and is the greatest improvement for a true bowling captain in history. Warwick Armstrong improved 31.7 per cent but only took 1.7 wickets per Test.

(PA Images via Getty Images)

Seven other captains have maintained a bowling average of less than 25 while in charge. Of these, the great Imran Khan was nearest to Mahmood, averaging 20.26. He did this over 48 Tests as captain (Mahmood led for ten Tests), which is nearly double anyone else with a sub-25 average and nine Tests longer than any other bowling captain. During his captaincy Imran took five wickets in an innings 12 times and ten wickets in a match four times.

Imran’s record in every category is top five. That is: averages both home and away, in wins, in losses, in his team’s first bowling innings whether winning the toss or losing it, and in the third or fourth innings. There are simply no weaknesses. Imran’s winning average of just 12.61 came from 14 victories and he is undoubtedly the best performing captain with the ball in Test cricket.

Other exceptional bowling captains include South Africa’s Shaun Pollock (21.36 from 26 Tests), England Bob Willis (21.59 from 18 Tests) and the Pakistan duo of Wasim Akram (23.35 from 25 Tests) and Waqar Younis (23.47 from 17 Tests).

At the other end of the spectrum, only three bowling captains saw their average decline by more than ten per cent. They are Ian Botham (refer above for his miserable captaincy experience), Heath Streak (also see above, he basically switched from bowling to batting) and India leg spinner Anil Kumble. Kumble is most famous for being one of only two players to take all ten wickets in a Test innings. Once he became captain his bowling averaged skyrocketed from 28.73 to 39.49.

One player not mentioned here is Indian pace bowler Kapil Dev. His bowling record as captain was a creditable 26.35, a 14.4 per cent improvement on his non-captaincy efforts. Kapil holds the record for the best bowling performance in an innings while captain: nine wickets for 83 runs against the great West Indian side in 1983. He also took eight wickets for 106 against Australia at the Adelaide Oval in 1985. Imran Kahn and Sri Lankan spinner Rangana Herath are the only other players to take eight wickets in an innings as captains.

The Crowd Says:

AUTHOR

2021-08-08T14:04:46+00:00

matth

Roar Guru


I don’t have a problem with shaving some outliers from scores to provide a more representative average, but like everything else in cricket, context is everything. And like everything else in statistics assumptions can have a significant effect. So for example does every score have to be examined for its context? But then there is arbitrary decisions made on what should be included or excluded and why. And then the cut off for exclusion is arbitrary. 100? 150? Top 10 percent of scores regardless of context? Scores in draws, but what was the context of the draws? Take an extreme. If we lower the threshold for excluding scores too far are we trying to bring everyone down to a level? I’ve been taught to always start with a question set the parameters and assumptions and let the answers fall where they may. My concern is two possible situations: 1. Where a person starts with a desired answer and sets assumptions, cut offs and parameters to ‘prove’ that answer. Inevitably they are successful. Or 2. Arbitrary cut offs snd assumptions are put in place. A conclusion is drawn and treated as ‘fact’ when that ‘fact’s existence is entirely dependent on the assumptions made in the first place. So sure, take out some outliers to ‘normalise’ an average, but we have to recognize that this means we are manipulating data.

2021-08-07T00:35:36+00:00

Renato CARINI

Roar Rookie


*Are large scores a problem

2021-08-07T00:32:08+00:00

Renato CARINI

Roar Rookie


matth. I gave your concerns some more thought and it seems to me that the difference in our positions comes down to this: Are large scores are problem when it comes to the batting average? Most statisticians believe that outlier scores have an effect on the mean and present an inflated and misleading figure. And it seems to me that this is a generally accepted problem when it comes to the cricket 'batting average'. Am I correct in concluding that, in your opinion, large scores are NOT a problem when it comes to career batting averages?? I would love to confirm your position on this.

2021-08-06T02:40:04+00:00

Rowdy

Roar Rookie


Bradman, as a player of the 30s, yes to those issues. Had he been born in the 90s he'd be born of these times.

2021-08-06T02:35:11+00:00

keith griffen

Guest


I'd like to have seen Bradman fare against a modern playing schedule, social media, playing in the subcontinent, managing the ball tampering scandal, sledging to name but a few issues, including playing reverse swing, extreme pace and spin. All will take it's toll on averages.

AUTHOR

2021-08-06T02:01:48+00:00

matth

Roar Guru


I’m not arguing for Bradman over or under Trumper, I’m examining the method. Bradman and Trumper were obviously the preeminent batsmen of their respective ages by a fair way. Only maybe Hobbs comes close. That is what makes the current era so interesting. Smith has potentially been the first batsman since those times to be a significant gap ahead of 2nd best. (Viv could also have claims for his era but Chappell and Gabaskar were both very good indeed).

2021-08-06T01:35:40+00:00

Renato CARINI

Roar Rookie


I used Lara's extreme example to make a point. I could have used hundreds and hundreds of less extreme but still pertinent examples Adam Voges, David Warner, Wally Hammond, Len Hutton, Billy Murdoch, Graeme Smith,Sanath Jayasuriya, and of course, Bradman.

2021-08-06T01:28:52+00:00

Renato CARINI

Roar Rookie


I know you think Bradman was superior to Trumper but it's hard for you to explain why Trumper performed so much better over their first hundred runs. At least in the first innings. The fact that Bradman pulls ahead after 100 is revealing. The Don only outperforms when Trumper is no longer interested. As Arthur Mailey said, Bradman's appetite for big hundreds makes him a more valuable player than Trumper but it does not make him the better batsman. This should be obvious.

2021-08-06T01:10:55+00:00

Renato CARINI

Roar Rookie


Why 100 and not 150? Because 100 has long been recognised as a THE most significant benchmark. The batsman's symbol of success. Not 150, 180 or 200

2021-08-06T01:05:43+00:00

Renato CARINI

Roar Rookie


At least, I can conclude that player X is performing better over their first hundred than player Y. And it would be hard to argue otherwise.

AUTHOR

2021-08-06T00:38:04+00:00

matth

Roar Guru


Why 100 and not 150? Why not different caps for 3 day tests vs 4 and 5? Why is overwhelming scoreboard pressure in longer matches not given any credence. Why is Lara’s obvious outlier of a series treated as a reason to place such a rule in for hundreds of series all with different characteristics. I know you see this as an objective measure, I see it as subjectively deciding on a method and cut off to look for a certain result. But hey if we all agree on everything there would be no articles to write snd no comments :happy:

2021-08-06T00:12:07+00:00

Rowdy

Roar Rookie


I reckon Bradman would’ve risen to the top even in today’s milieu. He was a very driven person. ——- I am not a fan of Bradman but l respect his contribution enormously. “Good batsman” as my 2nd favourite cricketer of the era would say with menace in his tone.

2021-08-05T23:46:56+00:00

Keith Griffen

Guest


Bradman is the best of his time no doubt and has no comparison almost from that era.

2021-08-05T04:47:38+00:00

Once Upon a Time on the Roar

Roar Guru


The only foreign country Bradman played in was England.

2021-08-05T04:39:58+00:00

Rowdy

Roar Rookie


I meant his peers of the 20s to 40s

2021-08-05T02:43:34+00:00

Renato CARINI

Roar Rookie


Hi Matth. You have raised a few different points here: scoring rate, performance in drawn games, and the value of runs scored after passing 100. I am only going to address the last point. Firstly, in my latest analysis of Bradman (which I'm not sure you read): https://www.theroar.com.au/2021/07/10/australias-greatest-victor-and-the-don/ I did not cut off all Bradman's hundreds, only those where his team established a colossal first innings lead. In eight first innings 100s, his team established leads of 380, 365, 504, 304, 290, 384, 404 and 344 so I think it perfectly reasonable to cut these hundreds off at 100* If these runs were not 'soft' then I'm not really sure what is. Second, the main problem I have with counting large hundreds (those made in front-running situations, only) is that these outliers will give the reader a false impression when it comes to figuring out their average. For example, in 2003-04 Brian Lara scored 23, 0, 36 and 400* in his four first innings. Lara had very little influence on the series but his first innings average is 153 However, capping at 100, his average falls to 53 which is a much better indicator or his performance, IMHO. The same goes with first innings scores of 280*, 0, 0 ,0 ,0 versus 70, 70, 70, 70, 70 Both batsmen have the same average but their match-winning influence is miles apart. With my capping method, the first player's averages falls back to 25 which, I think, is a more accurate reflection of his performance and match influence. THAT is the problem that I have with NOT capping scores (in obviously front-running situations). It is not that runs beyond 100 are worthless but they compensate for failures elsewhere to obfuscate rather than enlighten. Besides, capping at 100 is certainly an objective process and affects each batsman equally. On balance, the capping method seems to take us closer to the truth than simply counting all scores indiscriminately.

2021-08-05T00:45:08+00:00

Keith Griffen

Guest


Players aren't playing an opposition with only 1 style of play and are also playing cricket in 3 formats 300+ days a year and are playing professional sport not just for honour

2021-08-04T21:46:18+00:00

Rowdy

Roar Rookie


Well why was everyone else not doing a Bradman?

AUTHOR

2021-08-04T06:09:38+00:00

matth

Roar Guru


The one thing from you analysis of Bradman that I've never understood is how you discount his runs over 100 as useless and yet he played in a high scoring era, and 23 of his 29 tons resulted in wins. So you argument typically seems to be that because he had scored enough to win any further runs are discounted, even though the tests were still won? given these were not three day tests, surely the scoreboard pressure Bradman exerted had its own effect in achieving those wins? In those four draws where Bradman scored centuries, the first was a triple where is can be argued Australia didn't leave themselves enough time. In the second England survived 135 overs in their second innings to force the draw, so Bradman's time wasn't a factor as much as trouble bowling the opposition out. in the third Bradman scored an unbeaten 2nd innings century to salvage a draw from a weak position. In the fourth he scored 102 off 135 balls chasing victory.

AUTHOR

2021-08-04T06:02:01+00:00

matth

Roar Guru


This is all true, but doesn't explain how all the other players of the era, with the same advantages and disadvantages that Bradman had in the era were to a man at least 50% worse than him (by averages). comparison to peers is really the only reliable rating of cricketers and on that mark, no batsman has been further ahead of his peers.

More Comments on The Roar

Read more at The Roar