Saturday, December 22, 2007

Finding Referee Trends That Hold Up Across Seasons

NFL referees have a variation among them for each statistical category. Can we look at the variations to see whether any of the variations appear not only in the 2006 season but also the 2007 season? If we can find a variation that holds up across two seasons, it suggests that the characteristic is something that depends on the particular referee and is not just by chance.

One way to try to look into this is to see if there is a correlation between a particular referee's 2006 statistic and his 2007 statistic in the same category. For example, if you chart out each referee's average number of penalties per game in 2006 and then compare it ref-by-ref to each referee's numbers for 2007, you can find out the correlation if you plotted them on a graph.

Let's look for correlations across the 2006 and 2007 seasons in a variety of categories:

Strong correlations (suggesting this does depend on the referee) -- comparing the full 2006 season to the 2007 season through week 15 for the referees who worked in both 2006 and 2007:
Average yards per penalty (0.544)
Number of penalties for the home team (0.407)
Total penalty yards (0.378)
Number of penalties (0.364)
Yards per penalty for the home team (0.364)
Yards per penalty for the visiting team (0.361)
% of games where the visiting team had more penalties called (0.332)

Weak correlations (suggesting this does not depend on the referee):
Difference in number of penalties for the visiting and home teams (-0.170)
Percent of penalties called on the visiting team (-0.158)
Home team win percentage (-0.039)

Please post your thoughts on whether you think this attempt to find correlations across seasons is some indicator of whether a characteristic may depend on the particular referee!


Post a Comment

<< Home