When it's time to decide which four teams are the best in 2014 we're going to be asking which "proved the most" by winning against better teams. Here's a pre-season look at how tough it will be for teams to be included in that discussion.
None of the good ways to measure schedule strength will be available until lots of games have been played, so all we have to go on is opponents' ranks, and until the pre-season computer rankings are published all we have is last year's. I am using the value in the Maj column - the best rank for which more than half of the computers have the team ranked that highly.
Ordinal ranks are not an ideal metric - there's a lot more of a difference between #20 and #1 than there is between #70 and #50 that isn't captured with a simple average. For full schedules a reasonably good alternative is to find the rank that characterizes more than half the opponents. For example Arkansas has seven opponents in the top 27, and by any measure has the toughest 1AA schedule. Even that isn't very useful for non-conference schedules with only three or four games.
For an average that accounts for the rating non-linearity that is lost when just rankings are available I wanted something that "looks like" the rank-based interest formula but with only the opponent's rank as a variable. I came up with
|875000 - 15625×Rank + 200×Rank2 - Rank3|
Based as it is on last year's opponents' rankings, it is no surprise that 19 of the 25 toughest schedules will be played by SEC or Pac 12 teams. Predictably, only two teams from those conferences appear in the top 25 non-conference schedules.
SOS Based Upon Opponents' 2013 Computer Ranks
NonConference SOS Based Upon Opponents' 2013 Computer Ranks
© Copyright 2014, Paul Kislanko