Thursday, October 25, 2007

A Season-Long Experiment

My rooting interest lies with the WKU Hilltoppers. I have spent the last two months lamenting the schedule. It is not a terrible schedule in terms of quality, mind you, it's the timing of the games that has me worried. See, much of the Sun Belt is also in the Bible Belt, and that means that Wednesday night games conflict with church for a healthy portion of our community (we have four Wednesday night tips this season). That means fewer big home crowds for the Tops. Further, the Toppers play a whopping nine games away from home on Saturday nights. My question is, do these things matter?

I am not a mathematician or a statistician, so I may not be able to prove causation, but we can certainly see if there is a correlation between crowd attendance and losing on the road. I am interested to examine several things:

1. Does playing on the road on Saturday (presumably facing larger crowds than weeknight games) increase a team's chance of losing a road game?
2. Ken Pomeroy has demonstrated that home court advantage in a broad sense is often overblown. But, I'm not sure his data shows us that spiked attendance at "big games" does not matter. Of course, Duke and Kentucky and several other teams are almost always going to have a big crowd on hand. We already know it's tough to win there on any night. I am more interested in...
3. Do teams perform better statistically when the crowd is bigger? Do they "pull the upset" with a bigger crowd on hand? For example, does San Francisco "overperform" when Gonzaga comes to town and a big crowd is on hand?

It should be interesting to track this over the course of the season. Of course, the numbers will not mean much, statistically speaking, until late in the season when the RPI and Sagarin can help provide some baselines.