Check up on that notion with stats

Stephen Fleming scores a big Test double in the ongoing Test series. Fleming’s 99 from his last Test tour in SA and that one-day ton against the South Africans in the last World Cup immediately spring to mind. This man loves scoring against South Africa. I run the Cricinfo stats filter to confirm, and am promptly told that I drew a blank there. For someone with an overall career batting average approaching 39, Fleming fared a meagre 30.15 runs (573 aggregate) against South Africans prior to this match and the 262 at Capetown stands out like Table Mountains in the Newlands stadium backdrop.

I turn to his ODI stats in desperation. Fleming averages only a point and a half more than his career average (32.07) against his imagined ‘bunnies’, and that World Cup century remains by far his best score playing South Africa. Just two filtering exercises on Cricinfo’s statsguru and ‘pop’ goes a confidently held notion.

What went wrong there? I regroup from the statistical pasting after a while and sit back recalling the circumstances of the three Fleming innings mentioned above. The nature and background of those knocks went into creating that ‘SA-basher’ image of the NZ skipper. I realised I had jumped to a certain general conclusion based on a few specific samples. But isn’t that the method we often apply while judging sportspersons in any field?

We, the sports viewing people, are often reasonably correct in our judgement of players and teams we follow on a regular basis, even without going into stats filters. The samples (live player performances, say) are gathered aplenty and fed into that processor residing above our eyes. Our mind does a decent job at creating ‘impressions’. It is better than most computers at our disposal in filling up the odd missing input data all by itself. Those unique add-on tools named ‘perception’ (P) and ‘interpolation’ (I) never cease to amaze us in the way they keep functioning without our realising it. (Remember the optical illusion test your friend emailed you the other day?)

The gap between such ‘impression’ or ‘notion’ and reality, however, gets wider as the number of samples dwindle and P-I dominate the proceedings. This difference tends to come through in our judgement of such players, largely from other teams, that do not qualify as ‘idols’ to us. Reason: too few sample performances get fed, with too much job left to be done by P & I. Most of these samples got fed at all because they managed to attract our attention by their exceptional natures, and these may not truly represent the general pattern of the entire field.

Let us understand more about this aspect through an imaginary example. An American man passionate about baseball visits India during that feverish 2001 Indo-Australian Test series and gets an introductory dose on this longish sport from his enthusiastic local friend. Now this gentleman again happens to visit Australia to celebrate his New Year holidays there in December 2003. When he returns to his homeland in January he carries with him an impression that this languid player named Vangipurappu Venkat Sai Laxman would never be allowed to step into a baseball field but nevertheless can become the greatest thing that happened to this ‘cricket’ game. He reserves his best for the best opposition.

Cut to March 2006. The US guy fits a 2-day Indian stopover in his South Asian business trip, lands at his old friend’s place and enquires about latest developments in the game of cricket. Imagine the look on the guy’s face upon being told that the same Laxman, not yet too old by cricketing standards, played all of one ball and scored no runs in a full 3 test series against a second strength team.

That hastily-compiled beta version of the example was perhaps not too comprehensive in design. Let us take a real case study, one of those strange general notions that persist in India despite being contrary to the facts, ones that manage to stand the test of time and stats. I am talking about Ajit Agarkar’s perceived decline as an ODI bowler. It will surely be interesting to find out how he compares to the brilliant Irfan Pathan in the era they bowled together till date, a period when each of Ajit's many returns to the national side raised sniggers.

Brief Result: Agarkar’s bowling average of 30.9 at an economy rate of 5.09 does not compare too unfavourably with Pathan’s corresponding figures of 29.7 and 4.88. Unbelievable, isn’t it? If this can happen to cricketers we follow keenly, there is little chance of others (i.e. the ones from ‘another world’, as Different Strokes reader Jay observed recently) to be spared such delusions.

Professor Ang is on his way folks. I’m waiting for a soon-to-be-publicised manhunt named ‘Deano ki Khoj’ [Manhunt for Deano] on Indian sports television to spread my wings. Meanwhile, I dedicate the essence of my latest thesis to my reader friends: To avoid skewed notions, it is better that we take the help of statistics while assessing players, especially those that we do not watch regularly.

Come on, think about it students, and let me know what cricketing notion of yours got busted after a brief stats study. Your immense contributions to the cric-tattoos post make us expect responses on similar lines.