Week 5, NFL scoring stats.
To explain the columns above, Median is a median point spread, and can be used to get a feel for how good a team is without overly weighting a blowout win or blowout loss. HS is Brian Burke’s Homemade Sagarin, as implemented in Maggie Xiong’s PDL::Stats. Pred is the predicted Pythagorean expectation. The exponent for this measure is fitted to the data set itself. SOS, SRS, and MOV are the simple ranking components, analyzed via this Perl implementation. MOV is margin of victory, or point spread divided by games played. SOS is strength of schedule. SRS is the simple ranking.
I don’t like working stats the first couple weeks of the season, because they don’t mean very much. But for those who might want some week 4 stats, here they are.
For those with sharp eyes, Pro Football Reference’s 2013 stats, specifically their SOS and SRS stats, have been busted. This is a screen shot of page http://www.pro-football-reference.com/years/2013/ around 8:30am on this October 11, 2013.
The simple ranking (SRS) is supposed to equal MOV + SOS. On page above, it does not. (seen October 11, 2013).
Simply put, MOV + SOS is supposed to equal SRS. It, in general, does not. Normally my SRS values are within one tenth of a point of PFR’s, but not today, and the failure of the calculation on their part is pretty plain to see.
Nathan Oyler, via Twitter, started asking me the following question, how can my code be used to calculate offensive SRS and defensive SRS? I looked a bit and there doesn’t seem to be a good definition for the term. Some comments by Chase Stuart (1) lead to the notion that you start with “points for” and “points against” stats, and that gives us some useful clues.
If SRS can be broken into two components, OSRS and DSRS, for offensive and defensive SRS, then so can margin of victory. The sum of offensive margin of victory and defensive margin of victory have to equal margin of victory, so it then becomes possible to define such a term.
If you take “points scored” or “points for” and add them all up, divide by the total number of games played, you get the average number of points scored, by one side, in a game. We’re going to call this the average score for now. You can then define the offensive and defensive mov in this way:
offensive mov = (“points for” – games_played(team)* avg_score ) / games_played(team)
defensive mov = ( games_played(team) * avg_score – “points against”) / games_played(team)
These definitions work the way they should. Add the two together and you get margin of victory. Put these terms in place of the MOV in a SRS calculation, and you have a way to calculate an offensive SOS and a defensive SOS. There are still some wrinkles, and those belong in another publication. That’s enough for now, though.
1. See Chase’s reply to Racer1.