Expert Blog

What can cricket learn from Google?

Large numbers rarely lie, and massively large numbers almost never lie. So where does cricket come into all this? Read on…

This is an intriguing title, and I’m not even sure if we’ll get somewhere. But this is a blog; not a research paper to be published in the Journal of Combinatorial Theory. So one is occasionally allowed to ramble, or wander briefly into a path leading to nothingness.

 I’ve been reading an interesting article by Chris Anderson in the Wired Magazine titled “The End of Theory: …” (if you want to go straight there, it’s at http://bit.ly/LOU8) which makes an alluring reference to ‘The Petabyte Age’. We are of course talking of data sizes; a petabyte is 10 ** 15 bytes … that’s like watching HDTV movies non-stop for 13 years with even a bio break.

Anderson is saying something that most statisticians would find impiously irreverent. He’s saying: forget your theories, models, hypotheses and such crap … simply dump all your data into the biggest computing clusters that the world has ever seen … and then let statistical algorithms find patterns!

Think of how Google conquered the advertising world. They abandoned all theories involving demographics, culture, etymology etc., and simply argued that if a lot of people say that A is like B, then A is like B … and don’t pull out your hair wondering why! Google bases its decisions only on empirical observations and correlations … and that formidable reassurance of large numbers! Large numbers rarely lie, and massively large numbers almost never lie.

So where does cricket come into all this? What will we find if we dump all the cricket statistics, all those cricket pictures and articles, all the cricket footage and replays, all those interviews with Lalit Modi when he was still smiling broadly, all the foul language that Australians use while sledging, all the sounds uttered by Harbhajan Singh beginning with the letter ‘m’, all those slow motion shots of a topless and wired-up Murali and the data about how his elbow was bent during the first, second and last trial, all video footage of Sunil Gavaskar batting with and without his skull cap, and of Vivian Richards chewing gum with or without a shaven head?

Will we get somewhere?

If we don’t, let us add the digital scans of Yuvraj Singh’s knees, the haemoglobin count of Sachin Tendulkar’s blood before and after a certain book is published, the chemical composition of all the drugs prescribed to Mohammed Asif, the variation in decibel level in Kumar Sangakkara’s appeals, the number of times Preity Zinta hugged Brett Lee, the path traversed by an errant dog when it strays into Chinnaswamy stadium …

If even that doesn’t take us anywhere, then let us put in all the telephonic conversations logged in at telecom switches, the voices of all suspected bookies, the names of all persons who use a telephone number starting or ending with 111 or 222, the timelines when the bowler bowled an ‘accidental’ beamer, bouncer or wide, and when an experienced batsman failed to ground his bat to get run out, the number of balls taken to score the first 50 runs of an innings, and the next 50, the voice feeds travelling through the Cronje earplugs, the closed-circuit TV footage of all hotel lobbies in Mumbai and Barbados, the number of times there is confusion about which captain won the toss, the frenzy and the lull in the betting markets, and the ups and downs in a cricketer’s bank balance. Also add e-mail transcripts of anyone named John, Shane or Mark, the number of minutes spent by a batsman to tie his Adidas shoe laces and the frequency with which Shahid Afridi bites the cricket ball between the tenth and 15th overs.

Surely at least now we’ve got somewhere.