Saturday, November 25, 2006

NFL Refs Affect Total Points Scored? (Peter Morelli and High-Scoring Games)

Do you think the total points scored in a game is affected by which referee crew is calling the game? In theory, this might make sense -- not every referee crew calls penalties exactly the same way so if one crew rarely calls offensive holding and teams are aware of it, the offensive lines can get away with what another crew would consider holding. As a result, in theory, both offenses would score more than expected and the total points scored will be higher than expected.

The idea that a referee's tendency would give a hidden advantage to both offenses is not new. For example, in Michael Lewis's book The Blind Side, he described on page 215 how offensive lineman Steve Wallace was happy to see his favorite ref Jerry Markbreit in a key game because Markbreit lets left tackles get away with a lot, including being a bit farther back from the line of scrimmage than they are technically supposed to be. The assumption is that Markbreit would give the hidden advantage to both offensive lines and therefore his games might tend to have more total points scored than expected.

But I am not sure you would see a direct correlation between penalties called and total points scored. For example, if a referee crew calls offensive holding tightly but is well-known for rarely calling defensive pass interference, then you'd expect the total points scored to be lower than usual, even though the referee crew calls less penalties than usual.

Post your guesses and how you think we can study this.

As an enterprising attempt at tackling this, we took an unrepresentative sampling of the NFL referee statistics just as a test -- the two refs that have the highest penalties accepted after 11 weeks and the two refs that hvae the least penalties accepted after 11 weeks (Gene Steratore, Ed Hochuli; Walt Coleman, Peter Morelli). We then compared the total points scored versus the average you'd expect from each team's offense. (To simplify our test, we ignored how strong the defenses were.)

The results?
Looking from week 1-11, there is a little variance in total points scored per game: Peter Morelli 24.3, Walt Coleman 21.4, Ed Hochuli 20.9, Gene Steratore 20.5. Hmmm, the two refs who call the least penalties have higher total points scored than the two who call the most penalties.

There is some variance between total points scored and what we would expect based on each team's offense: Peter Morelli +4.1, Gene Steratore +1.2, Walt Coleman +0.6, Ed Hochuli +0.4. Hmm, Peter Morelli who calls the least penalties stands out here.

How about a correlation between the net difference from weeks 1-5 and 6-11? There is a correlation of 0.11 for net difference (and a correlation of 0.36 for total points ignoring which teams were playing).

So far, I think there may be an effect for some, but not all, refs on the total points scored in the game. The biggest trend I see so far is that if you Peter Morelli is the referee, expect more points scored than usual!

Data for Peter Morelli with actual total points scored versus expected points scored (average of offenses after 11 weeks) week-by-week: 34, 41; 61, 37; 35, 47; no game wk 4; 54, 40; 44, 40; 58, 50; 63, 41; 44, 45; 34, 41; 37, 31.

Or for Peter Morelli, actual compared with expected, week-by-week: -7, +24, -12, bye, +14, +4, +8, +21, -1, -7, +6. Six higher than expected and four less than expected (one of which was just -1).


Post a Comment

<< Home