It is the summer holidays, a time at which one’s thoughts move from idiotic economic policy and commentary, to enjoying the annual ritual of getting the garden under control while listening to live cricket.
So naturally, blogging has to go into summer mode as well. The cricinfo blog, of course, is locked permanently in summer mode. A recent post there concerns the importance of controlling for pitch quality in any statistical analysis of performance. Essentially, we are dealing with the econometric issue of misspecification in which independent variables are correlated with an error term if we cannot include a variable for pitch quality in analysis.
The author of this blog post proposes a method for using information from the match to control for pitch quality, by combining the runs per wicket and balls per wicket in the match. This looks like a useful first step but I see two limitations: First, the method for combining these two statistics is ad hoc rather than resting on foundations from a theoretical model of the processes driving outcomes; and second, the method seems to overreach, by ascribing all of the variance in scores between matches to the pitch quality and none to relative performance, when in reality performance and conditions both contribute to the observed variance in scores across matches.
My former student, Scott Brooker, and I have recently put out a paper drawn from Scott’s Ph.D. thesis, in which we propose a theory-based method for estimating from historical data the pitch quality in ODI matches. It will be of interest only to cricket loving stats geeks, but if you are a member of that fairly large community, do check out our paper here. We will be drip feeding more work from this research agenda into working papers over the next few months, including a paper on how incorporating pitch information can lead to improved rules for adjusting targets in rain-affected matches.
Postscript: Needless to say, my co-blogger at Offsetting has even less interest in cricket than he does in rugby, but rest assured that is a deficit that his colleagues have been trying desperately to correct over several years!