Sunday, April 3, 2011

ICC World Cup 2011: Statistical Analysis of Cricket Matches

India just won the ICC Cricket World Cup 2011 today, beating Sri Lanka by 6 wickets. They had to chase 275 and surpassed the runs with 10 balls to spare. It was a really good match all along, with both the teams trying to deliver their best.

Being in the United States, where all the rage in sports statistics is with respect to baseball (or so it seems, although, I will have to agree that the entries in INFORMS Blog Challenge on OR and Sports are quite diverse, with one entry on cricket as well), I do not get to hear a lot about statistics and Monte Carlo simulation in relation to cricket. Just like baseball, cricket is ripe for an infusion of heavy statistics, and in fact, it might be already there. There are plenty of statistics available to whet anybody's analytic appetite, see for example the Statsguru section of ESPN Cricinfo website, where you can create customizable queries to get any statistics of your choice. Anyways, I decided to take a look at the literature on this topic, and was happy to notice quite a few interesting papers. Here I list a few, which I found interesting:

[1] M. J. Bailey & S. R. Clarke. Predicting the match outcome in one day international cricket matches, while the match is in progress. Journal of Science and Sports Medicine. 2006. Vol 5, p480–487.
This paper uses multiple linear regression analysis to try to predict various outcomes of a match with the data that is already available.

[2] Swartz, T. B., Gill, P. S., and Muthukumarana, S. Modelling and simulation for one-day cricket. The Canadian Journal of Statistics. 2009. 37(2). p143–160.
This paper goes pretty close in simulating ODI matches, to answer various questions, like how many runs will be scored etc.

[3] B. M. de Silva, G. R. Pond & T. B. Swartz. Estimation of the magnitude of victory in one-day cricket. The Australian and New Zealand Journal of Statistics. 2001. 43. p259–268.

Naturally, most of the these articles (including all of the above ones) have appeared in journals dealing with statistics. The other avenue for related research are the journals of Operations Research. There are quite a few articles appearing in OR journals regarding modeling of batting orders, bowling orders etc. One of the most famous article appearing in an OR journal regarding cricket is the one regarding the Duckworth-Lewis criteria, which is already often used in practice to determine the winning team (or reset batting target) when there are unexpected interruptions in a match, e.g., rain. The article reference is as follows:

F.C. Duckworth & A. J. Lewis. A fair method for resetting targets in one-day cricket matches. Journal of the Operational Research Society. 1998. 49, p220–227.

Could we have simulated the result of the World Cup Cricket final match which happened today? One thing is for sure, even if we would have had the correct result from simulation, it wouldn’t have reduced the fun of watching it by any bit.

Update (11/16/2011): I have a few ideas about modeling cricket matches using Monte Carlo simulation, and would like to collaborate on a research paper. If anyone reading this is interested, please get in touch through comments.

No comments:

Post a Comment