Tuesday, October 25, 2011

Should One Play Effect an Entire Season?

On Monday I calculated the latest Top 25 NCAA FBS Production Model ranking and it has Wisconsin as the 2nd most productive team ("offense" and "defense") in the nation - as of last week. In today's environment this may seem strange as a loss propels teams down the rankings in college football. Yet here is a very clear example of how one event makes a difference in the final result - Michigan State won and Wisconsin lost.

The model that I am using though is trying to capture all the plays during the season to determine how productive teams are relative to the other 119 teams in the NCAA football bowl subdivision, and while this is a large event, it gets mingled with the other 446 plays by Wisconsin's defense and the other 491 plays by Michigan State's offense.

My question is: should one play effect an entire season? I am not sure that this is all that efficient in terms of measuring productivity, and hence believe that this one play over-emphasizes the measure of each teams performance on the field.

For those interested - here is the play and the resulting decision - which I agree was the correct call. OK, it is fairly dramatic.