By: George

George — Tue, 27 May 2014 15:30:26 +0000

I’ll take a go at the oddness of the Titans at the Chiefs line. I’ve started “solving” (Winston/Stern method) for things other than just offensive and defensive ratings (trying to build a better model). If you start solving for other things such as Individual Homefield Advantages (HFA) which average out to around 3, and then solve for Home and Away Random Error factors averaged to 0 (along the lines – admittedly this is for other sports – of what Stephen Clarke and the crowd at Swinburne University added to their Australian Football model, which also included individual HFA’s) you will note Kansas City (admittedly this is over a sample of 8 games and there appears to be no correlation year on year) had the worst HFA last year of (give or take) -6.85 (max HFA last year I had Buffalo at 14.17 which is why I now have them rated last). The Chiefs home random error I also had at -3.85 (so total negative effect of playing at home of around 10 points).

Therefore I am assuming the bookmakers are probably trying to factor in that Kansas aren’t that great at home, and even if their HFA regresses toward the mean, the chances are that it is still likely to be negative along with any random error associated with it. The major caveat with all of this is we are solving these over a small sample size (e.g. 7 or 8 home games) the however is that it reduced the sum of the squared error in my model by 6-7% which I took as being significant (which I don’t know if it’s the case but it’s a start).

By: Chase Stuart

Chase Stuart — Tue, 27 May 2014 15:20:42 +0000

In reply to Bryan Frye. Very cool stuff. It makes sense to see Denver ahead of Seattle here due to SOS.

Comments on: Vegas Has The Seahawks As the Best Team in 2014

By: George

By: Chase Stuart