Last month, I was at an
event at the Mumbai Literary Fest where Ramachandra Guha made 2 strong (and
mildly controversial) assertions
- Anil Kumble won India more matches than Sachin Tendulkar
- Anil Kumble would have made a better captain than Sourav Ganguly
Maybe it was because he came from my home state, Karnataka. Maybe it was because he held a degree in Mechanical Engineering from one of Karnataka’s best engineering colleges, RVCE. Maybe it was because he acted and bowled like a tear-away fast bowler (something Indian cricket fans sorely missed) – glaring at the batsmen and wiping out tails. Maybe it was because he once bowled with a broken jaw.
While the second assertion made above is tougher to verify, I decided to spend some time digging through the data to verify the first assertion.
Methodology
I looked at all of India’s
test victories in the period 1993-2008 in which both Sachin and Kumble played.
For each game, I identified the top 3 batsmen and top 3 bowlers in the team who
contributed to the victory. For rank-ordering the batsmen, I simply looked at
the total number of runs scored in the game (first + second innings). For rank-ordering
the bowlers, I assigned 3 points for each top order wicket, 2 points for each
wicket-keeper’s/all-rounder’s wicket and 1 point for each tail-ender’s wicket.
I then calculated a
cumulative score to assess each player’s value by assigning 3 points for games
in which they were the best batsman/bowler, 2 points for games in which they
were second-best and 1 point for games in which they were third-best. The
assumption made for this scoring system is that scoring runs and taking wickets
are both equally critical to success in test cricket (hence the same scale).
While one could argue that the 3-2-1 scale itself is arbitrary (i.e the
relative importance of a player’s performance is not always linear), I believe
it is at least indicative of the importance of a player to the team.
Results
Sachin and Kumble played
together in 40 Indian victories – 28 at home and 12 away. Their contribution to
the team in these victories is given in the tables below. As per my cumulative
score scale, Kumble beats Sachin 90 to
50. Surprisingly, Kumble’s contribution in Away wins is also superior.
I also did the same analyses for Sachin, Dravid and Kumble. They played together in 29 Test wins. Shockingly, Kumble’s contribution (63) equals that of Sachin and Dravid combined (64)!
However, in a country
gifted with abundant batting riches but a paucity of quality bowlers, our test
team misses King Kumble more than we miss God Sachin. And it is time we recognize
his contribution to Indian cricket.
PS: One could argue that
batsmen only get one chance for a mistake while bowlers can make several leading
to greater variability in performance, but maybe that only goes to reinforce
the importance of good bowlers in tests. Furthermore, one could argue that
India’s poor bowling resources gave Kumble more opportunities to pick up
wickets. But again, that only reinforces how critical he was to our success.
PPS: In a future blogpost, I will analyse drawn games in which Sachin performed well, but India failed to close out the win due to a poor performance by Kumble and the bowlers.
Backup data and analysis: https://www.dropbox.com/s/r6d369yzydqz0va/Sachin%20vs%20Kumble%20vs%20Dravid.xlsx?dl=0
PPS: In a future blogpost, I will analyse drawn games in which Sachin performed well, but India failed to close out the win due to a poor performance by Kumble and the bowlers.
Backup data and analysis: https://www.dropbox.com/s/r6d369yzydqz0va/Sachin%20vs%20Kumble%20vs%20Dravid.xlsx?dl=0
9 comments:
This is cool ! Proves Kumble's importance to victories.
I agree that Kumble trumps Sachin at home, but not with the margin :)
Excellent! How do you get these stats? I have been thinking about playing with cricket stats for a while but they're generally not easily downloadable!
I wanted to do something along the lines of: http://arxiv.org/pdf/1208.1204.pdf
Cheers!
Purushottam
:) Purshya - I did it by brute force. Opened every scorecard one by one. I don't think one can download data of this nature from Cricinfo. Statsguru has some interesting statistical cuts though.
awesome statistical analysis!
the 2nd point raised by ram guha could have been a subjective one though.
Wonderful. Can you please also include VVS Laxman and do the analysis. Looking forward for your next blog. Thanks.
i agree with the claim as far as the 1990s and Home Tests go (just intuitively). that he was so often the highest wicket-taker on those dust-bowls in the 1990s with 3 spinners is enough!
can't help thinking that the 3-2-1 ranking method (for best batsman or best bowler) introduces too many 'edges' for a few more runs scored (regardless of what time in the match they were scored) or a single wicket more taken (this could be a tail-ender's wicket) than the next best.
Especially on runs, agree about the potential skewing of 3-2-1. The ideal comparison would have been to see %age of runs scored - and then compare with %age of wickets taken. But that would have taken ages to do with the raw data difficult to pull from a database.
Just to clarify, tail-enders wickets were weighted as one-third that of a batsman's wicket - so the skewing due to inconsequential tail-ender's wicket does not hold.
Anyway - the purpose of this article was just to argue that Kumble deserves more praise and the data, whichever way you look at it, I think shows that.
totally :) Kumble does deserve a lot more praise, no question. no one seriously argued that did they?
your system does introduce skewing even for bowlers. let's say Harbhajan got 6 in a match and Kumble got 7 - one more wicket of a tail-ender, else same as Harbhajan. your system gives him 3 while Bhaj only gets 2. that difference is not reflective of their small difference in contributions.
other than that there are the standard truisms. bowlers win you matches since you need 20 wickets, batsmen need to get enough and can only set things up, etc.
it would be interesting to see such a comparison for more balanced bowling attacks like the great Aussie team (McGrath, Warne, Gillespie and co) or the great WI teams of the 70s and 80s or the 90s Pak teams with Waqar, Wasim and Saqlain/ Mushtaq Ahmed. and then compare each with the best batsmen in respective teams (Ponting,Richards, Inzy). they may end up a lot closer.
Post a Comment