Matches (13)
SL v AFG (1)
PSL 2024 (3)
NZ v AUS (1)
WCL 2 (1)
CWC Play-off (3)
WI 4-Day (4)
Stats Analysis

How do you rank the best captain in an individual Test?

The Captaincy Performance Index takes into account the team's performances and the captain's individual contribution to the team

Captain Ajinkya Rahane's century greatly contributed to India's series-levelling win in Melbourne in 2020-21  •  Getty Images

Captain Ajinkya Rahane's century greatly contributed to India's series-levelling win in Melbourne in 2020-21  •  Getty Images

A couple of weeks ago, an article of mine, on Test captains, was published in the Cricket Monthly. While that feature covered this fascinating subject in an anecdotal manner, this article will look at the subject from an analytical point of view, using the measure I created to evaluate the performance of Test captains at individual Test level, called Captain Performance Index (CPI).
A Test captain's role in cricket, unlike those of captains in, say, football, Davis Cup tennis, and basketball and baseball, comes with real responsibility. A cricket captain - setting aside the matter of how much advice he receives from the coach - bears responsibility in a number of areas. He selects the team, more or less; determines the sequence of bowlers used; decides the batting order; sets fields; decides whether the team plays safe or aggressively; and a lot more.
Inarguably, the Test captain has the most responsible and difficult task among captains in all sports.
So how does one measure a captain's performance? Results, of course, are important here, starting with the captain's own batting, bowling and fielding performances. Factors such as the venue of the match, the relative strengths of the two teams, how much experience the captain's team has, and who won the toss, come into play as well.
We can look at the numbers relating to these factors in different ways to arrive at an assessment of how the captain performed. But there is one more factor to understand. If a team wins by a margin of, say, under ten runs, we could say that the captain marshalled his troops at key moments. How do we measure a captain's impact in one- or two-wickets wins? It is the batters at the crease who have to perform. The captain, in the pavilion, can do nothing but chew on his fingernails. But he has to get some credit for such close wins for decisions taken earlier.
There are three major bones of contention.
The first is that most of these close wins could be attributed to the team rather than the captain, barring personal performances. That's true, but in my view the captain is like the CEO of a company. He takes important decisions on behalf of the team, he is one with his team, so it does not matter that some of these measures are identified strongly with the team. If the captain wins, the team wins and if the team loses, the captain loses. I am going to treat the captain and the team as roughly synonymous. (And my methodology ensures that "non-playing" captains do not have a great chance of getting high CPI values.)
The second is that it is impossible to assign values to captaincy decisions that might have had a profound impact on the course of the match. Say, the captain places two short covers and gets a batter caught driving, or places two leg slips to a sharply turning offspinner and gets his man, or deliberately leaves the midwicket area open and gets the batter top-edging - we have to accept that these are part of the captain's contribution but cannot really be measured.
Finally, how do we really measure the overall game-changing strategic contributions, such as those England captain Ben Stokes has made recently? (The coach, Brendon McCullum, is like the policy-making chairman of the company - let us keep him out of our analysis.) It is impossible to recognise these strategies in an objective manner and we have to be satisfied that these methods have been rewarded by a sequence of ten wins and two losses in 12 matches.
What is Captain Performance Index?
CPI is calculated through a combination of what the captain achieved for the team, in terms of the result and the margin of result, and how he performed on the field in his individual capacity. These numbers are indexed by four factors: how the teams match up on their strength matrix, their experience quotient, the location of the match, and the result of the toss. The calculations are simple and are explained below. A CPI value of 90.0 is about the highest that can be achieved.

1. Result points

1a. Result: A win is allotted the maximum of 40 points, a tie gets 25 points, and a draw receives 20. A loss is allotted four points because after all, the losing captain too has played, and often fought hard. And he could well have lost a close match, and so deserves these nominal points.
Recently, some readers have questioned some of these numbers. Why 40? Why not 50? Why toss?
Most of these are relative numbers. These are fixed so that the CPI maxes out at, say, a nice round number like 100. And regarding the weights, these are the result of my analysis for over two decades and the inputs provided through thousands of reader responses. I am open to considering well-thought-out alternatives; say, a fifth multiplying factor, if you have one. Or another base-point factor.
1b. Margin: The margin of victory, as perceived in the Team Performance Points measure, is allotted a maximum of ten points. The biggest win is England's 675-run win over Australia in Brisbane in 1928-29, which is allotted 9.42 points. The teams that did not win receive their fair allocation. Thus, a captain who loses by a narrow margin receives significant recognition.

2. Performance points

These are for a captain's own performance on the field, based on the percentage of team contribution he made. Care is taken that the performances are substantial and a lightweight one, such as taking two out of three wickets to fall in an innings (where the target is reached or there is a declaration or the match ends), is not rewarded out of proportion. The captain's individual performances carry around 15-25% weight.
2a. Batting: Ten points.
2b. Bowling/fielding: Ten points. (Fielding points account for wicketkeeper-captains).

3. Index values

Index values are used to multiply the Base Points. The range of the parameter is an indicator of the weight it has and its importance.
3a. Relative Team Strengths: 0.667 (very strong team) to 1.333 (very weak team). Maximum 2.97, and minimum 0.33. These are extrapolated to between 1.33 and 0.667.
3b. Location of Test: 0.875 (Home), 1.00 (Neutral), 1.125 (Away).
3c. Team Experience: 0.925 to 1.075. This is based on the sum of Tests played by the members of the teams - 600 to 0. Actual value of maximum Tests is 850 (India-2008); the 171 historical values (6%) above 600 are normalised to 600. I have deliberately used the absolute values of the teams rather than the relative values (like the team strength) since I strongly feel that a shortage of experience hits a team badly irrespective of who they play against. The captain of an inexperienced team has to be given credit for his team's lack of caps.
3d. Result of Toss: Winning - 0.975, Losing - 1.025.
The product of all four multiplicative indices (MF1) is used to adjust the Result-related values. The product of the first two multiplicative indices (Location and Team Strength - MF2) is used to adjust the Performance values since the Performance does not depend on Team Experience or Toss.
Captain Performance Index:
CPI = Result points * MF1 + Performance points * MF2.
Overall, it can be seen that this is a rather simple, easy-to-understand measure.
Let us now move to the tables.
The best performance by a captain was, by a mile, Rashid Khan's coup d'etat against a much stronger Bangladesh, away from home, with an experience quotient of a mere 13 Tests from the 11 Afghanistan players. Okay, I concede that Rashid won the toss and he performed like a champion himself: 51 (off 61 balls), 5 for 55, 24 (22 balls), and 6 for 49. Truly a great all-round performance. The net effect - a huge CPI value of 91-plus, ahead of the next best by more than ten points. Then comes Imran Khan, who won by an innings against England at Headingley in 1987 after losing the toss and fielding quite a weak team. He took 3 for 37 and 7 for 40 and scored 26.
South Africa, led by Trevor Goddard, stunned the strong Australians in Adelaide in 1963-64 by a comfortable margin, which fetched Goddard over 80 points. Archie MacLaren, leading a team containing four debutants and a low experience index, beat a much-fancied Australia side in Sydney in 1901-02. He got close to 80 points. Sunil Gavaskar led the way with a hundred in a tough away win in Auckland in 1975-76 and secured over 78 points. Most of us are familiar with the ninth-placed Test - Mominul Haque secured one of the greatest away wins against a much stronger New Zealand side last year in Mount Maunganui. He scored 101 runs and took two wickets.
Indian readers will be happy to see that Ajinkya Rahane's MCG win in 2020-21 is featured in 14th place. His outstanding match-winning 112 helped the team immensely. And the Brisbane win to seal the series also finds a place in the top 30. Rahane is the only captain to have two entries in the top 30.
I have five classifications in this table on Special CPIs. Since each has four entries, I will only briefly cover one in each classification. The lowest CPI for a win is that of Marvan Atapattu, when a very strong Sri Lanka, with a very experienced team (575 caps), won the toss and won comfortably against Bangladesh in Colombo in 2005. The key factor here is that Atapattu contributed a sum total of 11 runs out of 457, which goes some way to explaining why he secured a low 27 CPI points.
The most points for a drawn Test was when John Reid, captaining a considerably weaker and inexperienced New Zealand, secured a draw at The Oval in 1958. Reid scored 27 and 51 not out and took two wickets. He secured a very high 48 points.
A draw at Lord's in 1954 against a considerably weaker and inexperienced Pakistan fetched England captain Len Hutton fewer than 13 CPI points.
On the other hand, a decidedly inexperienced Zimbabwe ran Sri Lanka close and lost narrowly in Colombo in 2017. Graeme Cremer got over 30 CPI points for this loss, in which he took nine wickets and 61 runs - more points than Atapattu got for his win, which shows how sturdy the whole CPI concept is.
Finally, Mike Gatting, who did nothing in the match, gets a low 3.7 points for a innings loss against a much weaker Pakistan at Headingley in 1987 (the match for which winning captain Imran got 80.73 points).
Who was the best captain, based on average CPI per Test? It is no surprise that Don Bradman leads the table. He achieved a win per cent of 62.5 and an average CPI of over 36.4 CPI points per match. Richie Benaud follows with 33.8 and Ian Chappell is next with 33.5. In sixth place comes a surprise. Abdul Hafeez Kardar, though he won only six out the 23 Tests in which he captained Pakistan, has a relatively high average CPI of 31.9. There are a few reasons for this. The first is that the teams Kardar captained were mostly very weak and with almost no experience. And some of those six wins were memorable, like the one at The Oval in 1954.
This would also explain why successful captains like Steve Waugh, Ricky Ponting and Clive Lloyd do not have high average CPI values. They captained strong and experienced teams and the results they produced were mostly expected. While a win is a win, it can be seen that the CPI values can be dramatically different from one win to another. The list of the top captains contains some unexpected names like Dimuth Karunaratne, Shaun Pollock, and John Goddard.
At the other end of the spectrum - low average-CPI values - three Zimbabwe captains prop up the table. Then comes the real surprise. As I said in the Cricket Monthly article, David Gower had a very poor overall career as a captain. His average CPI value is just over 20. Another relative surprise is India's Mansur Ali Khan Pataudi. The win-percentage values tell a story. In general, the weaker teams have low values.
The table above is an extension of the average CPI values. I have presented the table by team so that the top captains by teams, based on CPI, can be viewed. The Australian top five have already been presented.
Hutton was the best English captain. His average CPI was just above 31 but he had a sub-50 win percentage. Mike Brearley is second. His win percentage was higher, at 58. However, as everyone will be aware, his batting contributions were minimal. He scored only 1100 runs in those 31 matches - a meagre contribution indeed.
For India, Virat Kohli is the best captain, by a wide margin. A win percentage of nearly 59 is very good indeed. His CPI average is an acceptable 30.6. Then come Rahul Dravid, Sourav Ganguly, and MS Dhoni - but with considerably lower average values. Dravid's win percentage is quite low.
For West Indies, Lloyd and Viv Richards follow Goddard. It must be understood that Goddard had much weaker teams than the two more illustrious captains. Pollock leads the table for South Africa, followed by Hansie Cronje and Graeme Smith, who averaged 27.2 across no fewer than 109 Tests. Akram was the best Pakistan captain, followed by Kardar, Misbah-ul-Haq and Imran. All have averages exceeding 30. Despite Kane Williamson's high win percentage, the fact that he had a strong and experienced team meant that Geoff Howarth leads the table for New Zealand. Karunaratne leads for Sri Lanka, way ahead of Angelo Mathews and Mahela Jayawardene.
Now for a set of matches in which the CPI values for the two captains vary by a mile. The highest difference occurred in a match we have already discussed. When Rashid secured over 90 points for Afghanistan's win against Bangladesh, the local captain, Shakib Al Hasan, secured a mere 8.6 points, a huge 83-point difference. Everything that worked for Rashid worked against Shakib. At Headingley in 1987, Gatting secured 3.8 points, while Imran got 80.73 points.
In most Tests, one captain gets the benefit of potentially getting more points because it is an away game for his team, and one gets the toss benefit. Unless the teams are evenly matched on the team-strength factor, one team will benefit. It's the same with the experience factor. And the captains perform at varying levels. In other words, the base factors could vary considerably, as also the multiplication factors. Given all these variations, it is a miracle if the two CPI values almost match. The way the values in the table above cancel each other out is fascinating. The match scorelines are very interesting.
At Newlands in 1992-93, Kepler Wessels and Mohammad Azharuddin managed to secure the same number of CPI points in a well-contested match. India, set 215 to win, were 29 for 1. In Chennai in 1981-82, India had the edge over England, but the other factors helped England attain parity. At The Oval in 1909 a tight match ended almost perfectly balanced, with England at 104 for 3, chasing 313 against Australia. In Lahore against West Indies in 1980-81, Pakistan led by 72 runs but slipped in the second innings and had a lead of 228 with three wickets in hand. The captains did not do much. In Napier in 2009-10, New Zealand had the edge at the end of the match against a well-matched Pakistan. Both captains did well.
Finally, a table to identify the best ten-Test sequences any captain has had.
Johnny Douglas had a ten-Test sequence in 1911-12 in which he accumulated over 450 CPI points. It was a win-dominant sequence with several away victories. Imran had a good unbeaten sequence in 1987; though he had only three wins, most of the draws were in away Tests. But what really helped him was his performances in these matches - 400 runs and 45 wickets. Dean Elgar captained South Africa to an excellent 440-plus point sequence in which he had wins over West Indies, India, New Zealand and Bangladesh. Bradman had a similar sequence between 1937 and 1946. However, it must be remembered that England were also quite comparable in those times.
Readers might wonder what has happened to all those successful sequences of West Indies in the 1980s and Australia in the 2000s. The truth is that these were almost certainly the strongest teams of their era and they just bulldozed their opponents, home or away, The captains had hugely experienced teams under their command. Many of their wins were clocked at CPI values of around 30. The bottom line is that just about any top player could have captained these teams and reached levels close to what those captains achieved. For the record, Waugh's best sequence was 349 points, Lloyd's 356, and Ponting's, 342.
There's another way of looking at the Australian streaks. I have looked at each of the 32 Tests and perused the key numbers at the beginning of the Test. The numbers clearly indicate that in every Test plus the two Tests that ended the streaks, Australia were favourites to win.

Conclusion

It can be seen that the CPI, although it has strong team-centric features, gives a good idea of how a captain has performed - both in leadership and player roles. Rightly, the result carries a higher weight. However, the captain's performances on the field have sufficient weight to clearly identify performing and non-performing captains.
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 me a request for inclusion, providing 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 in 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