Suresh Menon is Editor, Wisden India Almanack, and author, most recently, of Bishan: Portrait of a Cricketer
Stats and the truth aren't the same
Posted on: 09:10 AM IST Jul 10, 2012 IST
Cricket fans believe they have the answers to recurring questions. Who is the greatest batsman? Don Bradman. The greatest allrounder? Garry Sobers. The greatest captain? Mike Brearley. There is comfort in the familiar.
In some cases, we are told figures are all that matters, and Jacques Kallis' 12, 379 runs, 276 wickets and 181 catches in Tests is superior to anybody else's. Otherwise we are told that figures don't matter, and there will never be another player like Sobers. Somehow, through all these arguments - and what is sport without its arguments about the greatest? - one element has generally not had much focus on it. Captaincy. The cricket captain is a hands-on man who decides strategy, changes tack when required, and sometimes goes on pure instinct when his textbook training might pull him in the opposite direction.
Yet, for the greatest captain there are few candidates - Bradman, Brearley, Richie Benaud, Frank Worrell, Steve Waugh, Ian Chappell, Nasser Hussain. One reason is the lack of a proper evaluation system. Are victories alone enough? There is too the nagging suspicion that great teams make great captains and not vice versa.
A new algorithm - developed by Satyam Mukherjee, a scientist at Northwestern University, Illinois - discussed in a blog on The Economist, uses Google's PageRank algorithm. Mukherjee compares captains using a head-to-head to judge the quality of victory.
There is no room in the Top 10 in this system for Bradman, Brearley or Worrell. Steve Waugh is rated No.1 with Graeme Smith (No. 2), the only non-Aussie in the top five. More interestingly, there is room for only one captain from the pre-War (1945, for younger readers) years, Bill Woodfull who clocks in at No. 12. Half the captains belong to the contemporary era, which is understandable because there are more matches and therefore greater statistical information to judge from.
In recent years, there have been many attempts to look at the vast amount of statistics in cricket from a different perspective than the traditional, and find explanations. The Impact Index - again, developed by an Indian, Jaideep Verma - brings to cricket a version of the 'Moneyball' techniques made popular in the book and film of that name. There is a surface gloss to these departures from the long-standing.
But these - as well as the classical, let me hastily add - are equally valid or equally invalid. For they arrive at their conclusions based on a series of assumptions that, like in Euclidian geometry, have to be accepted on faith. Later developments are not more correct or more wrong, merely different.
Each system is internally consistent. The manner of calculating a batsman's average, for instance, chooses to leave out the number of innings in which he remained not out. Likewise with the Impact Index or the Mukherjee System. Each leaves out a whole slew of discrete events as being insignificant; but we don't know that they are. Or they cherry pick elements that have to be accepted on faith like the axioms in Euclidian geometry.
It is this cherry-picking that is the fundamental weakness of some of the most popular recent books on economics, sports or philosophy. Elsewhere I have called it the Malcolm Gladwell Syndrome. You can always join the dots differently to get a different pattern, but sometimes you can also put in fresh dots. Context is essential, but separating the wheat of the significant from the chaff of the incidental cannot be left to statistics in any field of human endeavour.
Perhaps there is a cultural approach to captaincy that makes a difference. In India, traditionally, no captain has been a hero to his vice-captain. Cliques form quickly consisting of the mediocre and the disaffected. A captain has to be leader both on and off the field, a task many good on-field captains have found onerous.
No statistical study can account for brilliant captaincy moves that went wrong because of a dropped catch, for example. When Mohammad Azharuddin put England in to bat at Lord's in 1990, history records that it was a poor gesture because Graham Gooch made 333. History also records that Gooch was dropped at 36. What might have been ... there are too many imponderables in a single game of cricket, nay, in a single day or a single over, for the reductionist to tie up everything with a single set of figures, great fun though it may be.
An approximation of truth is not truth itself - that is the charm of sport, and the consolation of the cricketing romantic.