Ranking the captains

This article was inspired from an online discussion where the posters were debating who has been the better of the two captains for India in Test Cricket, Sourav Ganguly or MS Dhoni. Many different arguments were presented to argue in the favor of one over the other: overall win-loss record as a captain, victories against weaker/stronger opponents, performance outside India/sub-continent etc.

Previously, we have presented a method for ranking teams and players purely based on the one-on-one match-up between the teams/players. Here, we apply the same technique to rank captains in International Cricket. The basic idea remains the same: we look at the head-to-head record between the captains in a particular format of the game, and based purely on this record, we generate a ranking of the captains. Apart from the fact that this method is absolutely free of any parameters (and therefore the most objective possible), the other advantages include:

  1. This method automatically takes care of the encounter against the “weaker” teams. That is, if a captain has an inflated win-loss record because many of his victories were against the weak teams, then that may not necessarily lead to a higher ranking for that captain. On the other hand, wins against stronger teams will tend to earn the captain a higher ranking.
  2. It also takes care of the fact the fact that teams may have been strong or weak at different times in the time period that we consider. This is because teams are usually led by different captains during their peak phase (captains with good records tend to stay) and down phase (captains with bad records are fired).

You can find the rankings of the captains for all international matches year 2000 onwards on the Captain Rankings page. The page lists the ranks of the captains based on their win-loss-draw record for the time period 2000 – present, but it is possible to generate the rankings for any time period within this interval by choosing the appropriate time period in the menu on the right hand side.

So what do these ranks tell us about the Ganguly verus Dhoni debate? In both the Tests and the ODIs, Dhoni is ranked better than Ganguly. In fact Dhoni is ranked 5th overall in the Test format for the time period 2000-2013, 3rd in the ODIs and 5th in the T20 format. No other captain does as well in all the three formats of the game (Graeme Smith is a close second), although Ricky Ponting does beat Dhoni hands down in both the Test and the ODI format.

This analysis may still not be enough to convince the detractors of Dhoni, but there is little doubt that Dhoni has indeed been a terrific skipper for Team India, and he is as good (if not better than) as Ganguly was.

11 thoughts on “Ranking the captains”

  1. Hrishikesh Kelkar

    Dont you think that another parameter in your analysis should be how the team was performing a few years prior to the appointment of a particular captain. This will tell you whether the captain has inherited a good team or a bad one.

    1. Hi Hrishikesh,

      That’s an interesting suggestion. I am not sure how to precisely define “few years prior to appointment” – should we look at the record of the team in the last one year? two years? And secondly, how exactly do we take into account the past team performance in ranking the captains? I would much rather keep the model simple and free of any artificial parameters.

  2. Interesting analysis. Does this also adjust for the fact that Dhoni has played way too many Tests at home than overseas (compared to say Ganguly and also probably other captains)? If not, request you to present comparative rankings including this adjustment.

    I am assuming winning at home is easier than away. So, a win at home should be less valuable than a win away.

    1. Thanks for your suggestion Amit. I don’t have the win/loss record of the captains broken down by home/away matches, but when I get time I will see which of Dhoni/Ganguly comes out to be better in terms of away records.

  3. I have to agree, that

    i) home conditions bias should be accounted for;
    ii) the other way to perhaps adjust for the “team inherited” question might be to look at the ODI/Test rankings of the players that made up the teams one year prior to a player taking on captain duties (agreed this would be tougher for ODI as the relative turn-over is higher for ODIs vs. Test) and adjust weights based upon relative difference a particular player made as captain.

    1. That’s an interesting suggestion Anshuman, but I am still not sure how do we “adjust weights” based on the ranks of the player constituting the teams. And as you have mentioned, there is always a turnover of players from teams, so it is not clear what the appropriate time period is that we should choose for this analysis.

  4. Gautam Gambhir comes at #2 ? Number of matches played should be also counted. Or at least have some kind of cut off. Too few data points can lead to random results

    1. Hi Piyush,

      The model is supposed to take care of the small sample size, and it does take into account the total number of victories of a captain. The reason why Gambhir gets ranked so high is because he never lost or had a draw match. This skews the rankings heavily in his favor.

  5. Siddharth Kirtikar

    Shashi this is fabulous. I can’t think of a better measure, but nonetheless… some points for possible further improvements:

    (1) Home/Away issue definitely makes a big differences these days. (TL;DR : PAK was expected to loose in SA, and they lost 3-0. Now, SA is likely to loose against PAK later this year. If we have a different PAK captain in the coming series, it is going to make a huge difference in your current method of ranking.)

    (2) Evaluating Captains as “Manager on the field”, rather than “A Brigadier in a war”, i.e. rather than relying solely on ternary data source, how about team’s performance relative to prior expectations (like stocks).
    [TL;DR : I have seen a lot of one-sided Test matches over last 2 years, either because of Team composition, or pitch conditions. For example, odds between Dhoni-Clarke were quite different in AUS, than in IND. It’s hard to evaluate performance of say Clarke in IND solely based on 3 losses. The question is how well he managed his limited resources. How about we use your win/loss analogy (used in Comparative ranking of Players), and apply it to the Captain’s decisions (i.e. how effective was he in utilizing a particular bowler or under utilizing a better bowling option, on that particular day and pitch). Eg. Dinda went for 63 in 4 Overs, despite better bowling options. http://www.espncricinfo.com/ci/engine/match/598011.html

  6. Hi Shashi,
    This is wonderful work that you have put up here. The guys who have commented have raised very good points on how to improve the model. I don’t know if this makes sense, but do you think the rakings of other players in the team might affect the captain’s ranking. For example, when Ponting took over from Steve Waugh as captain, he had pretty much the best players playing for him in every aspect of the game. I am not undermining Ponting in any way, but could it be that anyone who is captaining such an elite team would lead them to a large number of victories? That is, hypothetically if Ponting had the same magnitude of success with “lesser” players, then his ranking should definitely be higher, rite??
    Btw, I could be wrong.. Hope you are keeping well..


    1. Hello Chintoo,

      Thanks very much for your comments. I agree with what you (and others who have commented on this article) have suggested, that the strength of the team prior to the captain taking over should be accounted to the ranking. For the sake of simplicity, I have stuck to looking at only the heads-on wins/loss/draw records. Regarding Ponting specifically, he had an extended period of great success in Test Cricket, for which he should be duly recognized – irrespective of the strength of his team. Remember, a good team also needs a good leader to become really great.

Comments are closed.