Comments on: Week 11 Power Rankings – Happy Thanksgiving from Football Perspective http://www.footballperspective.com/week-11-power-rankings-happy-thanksgiving-from-football-perspective/ NFL History and Stats Thu, 22 Nov 2012 22:32:40 +0000 hourly 1 https://wordpress.org/?v=6.9.4 By: George http://www.footballperspective.com/week-11-power-rankings-happy-thanksgiving-from-football-perspective/#comment-3935 Thu, 22 Nov 2012 22:32:40 +0000 http://www.footballperspective.com/?p=4475#comment-3935 Happy thanksgiving to everyone. On the subject of ranking systems I am now effectively running 4 (Least Squares – Winston style, Weighted Least Squares, Massey 1997 Thesis style, Colley in a Massey style sense – e.g. points instead of wins to generate a rating) and then averaging these out when making predictions (e.g. taking the average % chance to cover). Not using capping in any of them so points scored is creating a bias (disappointingly).

In terms of who is top of each (all essentially hindsight predictions);

Least Squares: San Francisco, New England (clearly with a gap to), Denver
Weighted Least Squares: Houston, Denver, Chicago
Massey 97 style: San Francisco, New England (clearly with a gap to), Denver
Colley style but with points: San Francisco, New England (clearly with a gap to), Denver

Re: the weighted results the maths seems to go heavily against some of the odd blips/losses San Francisco has had where they lag a distant 6th.

Things that the ratings/spreadsheets liked for week 12 (one of which would have already resulted in a tie disappointingly);
Denver (-10.5) 86.4% chance of covering
New England (-6.5) 76.2% chance of covering
Seattle (-3.0) 73.1% chance of covering
Houston (-3.0) 72.3% chance of covering

One of the things I liked about the above is essentially in the case of the New England and Houston games you have maths against other factors (e.g. the effects of traveling on a short week) and so far the other factors are 1-0.

]]>