NFL Referee Statistics: Correlations from Season to Season
Which statistics for NFL referees seem to hold up from season to season? Let's test this by looking at the correlation across the spread of referees from one season to the next. For example, if the ranking of which referees have the highest accepted penalties per game is the same from one season to the next, we would see a strong correlation from year to year.
This is based on a comparison of the 2007 season to the first 9 weeks of the 2008 season and also a comparison of the 2006 season to the 2007 season.
Total accepted penalties per game has a strong correlation from year to year: 0.503 from 2007 to 2008 so-far and 0.392 from 2006 to 2007. Same for total accepted penalty yards per game with 0.516 for 2007 to 2008 so-far and 0.394 for 2006 to 2007.
So it seems that some referee crews year in and year out consistently call more penalties than other referee crews. Maybe some are sticklers for the rules and others let things slide.
There is also a correlation across the years for the number of accepted penalties called against the visiting team (0.275 this year and 0.123 last year), accepted penalty yards against visiting teams (0.236 this year and 0.332 last year), accepted penalties against home teams (0.298 this year and 0.489 last year), and to some degree accepted penalty yards against home teams (0.088 this year and 0.417 last year).
There is no correlation from year to year for the home team's win rate so far (this year 0.009 and last year -0.172). One point, though, is that there is such a small sample size. After the first eight weeks, Peter Morelli went from the referee where the home team did the worst in 2007 (33% win rate) to the opposite -- the referee after the first nine weeks where the home team has done the best (93% win rate!). If we took out Peter Morelli, the correlation would jump from 0.009 to 0.335 -- so let's see how it shapes up the rest of the year. Considering the correlation between 2006 and 2007 was negative, I am not expecting it to climb by much. But the adventures of Pete Morelli's home win rate continue...
This is based on a comparison of the 2007 season to the first 9 weeks of the 2008 season and also a comparison of the 2006 season to the 2007 season.
Total accepted penalties per game has a strong correlation from year to year: 0.503 from 2007 to 2008 so-far and 0.392 from 2006 to 2007. Same for total accepted penalty yards per game with 0.516 for 2007 to 2008 so-far and 0.394 for 2006 to 2007.
So it seems that some referee crews year in and year out consistently call more penalties than other referee crews. Maybe some are sticklers for the rules and others let things slide.
There is also a correlation across the years for the number of accepted penalties called against the visiting team (0.275 this year and 0.123 last year), accepted penalty yards against visiting teams (0.236 this year and 0.332 last year), accepted penalties against home teams (0.298 this year and 0.489 last year), and to some degree accepted penalty yards against home teams (0.088 this year and 0.417 last year).
There is no correlation from year to year for the home team's win rate so far (this year 0.009 and last year -0.172). One point, though, is that there is such a small sample size. After the first eight weeks, Peter Morelli went from the referee where the home team did the worst in 2007 (33% win rate) to the opposite -- the referee after the first nine weeks where the home team has done the best (93% win rate!). If we took out Peter Morelli, the correlation would jump from 0.009 to 0.335 -- so let's see how it shapes up the rest of the year. Considering the correlation between 2006 and 2007 was negative, I am not expecting it to climb by much. But the adventures of Pete Morelli's home win rate continue...
posted by Rex at 1:21 PM
1 Comments:
This is interesting, considering the composition of the ref's crew changes every single year......
Post a Comment
<< Home