Excellent article on football statistics ("Understanding Uncertainty")

[scruffbag]

Hi, I thought I'd post a link to the following site as it features some fantastic explanation on the application of statistcs to football league data. The site's developed by David Spiegelhalter, a Professor of Statistics at the University of Cambridge. Apologies if this has already been posted before: http://understandinguncertainty.org/node/45 http://understandinguncertainty.org/node/56 http://understandinguncertainty.org/node/61 One of the most interesting things I think that's highlighted from the analysis is that, based on the data from 2006/07, statistically there was no significant difference between 17 of the 20 Premier League teams. I'm just starting to develop my own models for football, and although I'll be using more than one season's data for this (which improves the capacity for obtaining "statistically signficant" results), I think this could have implications on how to account for team strength. Anyone else got any views on this? Cheers

[aliando]

Re: Excellent article on football statistics ("Understanding Uncertainty") as a lay man, not a statician, the premise makes sense - trick is turning stats on previous data into profitable future predictions. get that right and you'll be banned from every bookie very quickly, only have exchanges to ply your trade with and have an avid following ;)