ponyboy wrote:But Couch and Smurzer you can play that game with data on both sides, can't you?
An example. I'd love to prove that JJ's been the greatest coach in the history of modern sport and so when I put the comparison with Briles together, well it turned out that Briles had clearly outperformed JJ. My temptation was to exclude JJ's first rebuilding year (2008) and then this season (2013) as outliers that didn't reflect reality, the truth as I saw it. To do so would have improved our average Sagarin in the JJ years by 20 whole points. But that wasn't fair and what you're doing isn't fair, IMO. Avoid confirmation bias and include all data.
How would you like me to revise it? I actually expected JJ's years to look much better than Bennett's based on record but it isn't the case. The similarities were a surprise to me too. I chose to look at the data thus way to try to compensate for out of conference strength which messes up the "average opponent ranking" when the OOC teams vary so much.
What stats would you consider "fair"?