What's your model for strength of schedule?

C

CAARHawk

Banned
I have had a problem with SEC teams getting credited with playing in the toughest conference when they only play 8 conference games. The B1G also plays 8 conference games, but the P12 agreement adds a 9th game against major conference foes.

The polls definitely don't seem to take SOS into consideration, except perhaps giving too much credit to speculative SOS based on early wins against too highly early season opponents.

So my question is, taking all factors into consideration, how would you set up a SOS model that was equitable across all conferences?
 
Put me in the category that SOS for football is not a reliable factor. There are too many variants and too few games for predictability. The best measure of the strongest teams is conference champions. Then it becomes a beauty contest among them. Naturally fans will look at what teams did out of conference for comparison when they should be looking at the overall conference record against out of conference teams.
 
CaarHawk- they get credited for playing in the toughest conference because the last 6 national champs come from that conference. When has that ever happened before? 1 conference winning that many in a row is impressive. They have the most players drafted in the NFL each year and they have the most teams ranked in the top 10-15 each year. They had 3 in the top 5 at the end of the year in 2011. Those are facts.

Want to do something about it? Improve the reputation of our conference by actually winning bowl games on New Years Day (we have 1 win in the last 2 years on that day) and by winning BCS games (Michigan did this year) and BCS titles. We have been smoked the last 2 games we were in (Ohio State).

We had a golden opportunity 6 years ago when Michigan and Ohio State were ranked 1 and 2 Thanksgiving weekend. Both teams went on to BCS bowl games and got beat badly.
 
I dont know. I do know that playing an acc and a couple of sunbelts, then having a bye in Oct, just to come back and play two games before playing Chattanooga, is a pretty easy slate. That pretty much just says, all we need to do is win our conf games and for good measure, we will throw a bye in here (Oct) and another easy one in after going into Nov.

Not the same as having a B12 and a P12 in your non cons and playing 5 games in Oct. Now is it?
 
Last edited:
Main flaw in current sos is lack of meaningful number of BCS crossover conference games. There just aren't enough games played each year between the major conferences to really inform which are really the best conferences. I suppose the only way to improve this area would be to require each BCS conference team to play at least two BCS teams from two other conferences each year, with one of those being an away game.
 
Old Dude brought up a great point. The SEC often schedules a cupcake in mid to late November...every freaking year. Based on the system we have in place where 1 slip up in November sinks your whole season....playing a damn FCS opponent in November completely eliminates the opportunity to lose that game when things get dicey around the nation.

And also against an FCS team, you can play guys to keep them sharp as opposed to a bye, and similar to a bye rest some other guys that need an extra week to get healthy. Its a double edged sword, yet both edge is good for you.

When Oklahoma St tripped up against ISU on that friday night....who did Bama play the next day? Bama played the ever imposing.....Georgia freaking Southern.

In 2012, the week before the Iron Bowl, Bama has conveniently scheduled something like Western Carolina. Cupcakes in November for the SEC when THE REST OF THE NATION is mired deep in conference play against BCS opponents.

The SEC has earned their championships obviously by winning the games...no doubt about it, but some questionable scheduling practices certainly haven't hurt them in the month of November, while every other conference leader seems to run into some late season buzz saw teams.
 
So nobody can come up with a feasible model?

I was thinking about having some kind of RPI type multiplier built in that would reward teams playing tough games in November and actually hurt teams playing cupcakes. Right now, a team beating a cupcake jumps a team losing a close game to a strong team, that should not happen. At best, it should be a wash. But I would prefer the team playing the cupcake to actually get less credit, because of the other advantages they get for scheduling that way.
 
CAARHawk said:
I was thinking about having some kind of RPI type multiplier built in that would reward teams playing tough games in November and actually hurt teams playing cupcakes.
Click to expand...

At least one of the BCS computer rankings (Massey) give weight to games that are later in the season, for better or worse. The problem with the multiplier you suggest is that if it were implemented, the SEC would simply sandwich their FCS-opposition games earlier in the season.

People love to hate on the computer rankings (which clearly aren't perfect), but they're unbiased and never lie*. Basically they weigh each win/loss as an extension of an opponent's SOS - winning against one of the best teams equates to one win, while beating one of the worst FBS teams could be 1/10 of a win. Preseason rankings don't exist either, so teams beat a 7-6 Texas A&M, as opposed to a #7 ranked team.

I would highly recommend reading the Colley Matrix methodology to anyone interested in how rankings are compiled, seeing as they offer all their formulas online. Despite all their problems, I prefer this way of ranking teams when compared to human polls

Edit: The polls have an edge in differentiating between closely ranked teams, which is beneficial for things like deciding who is deserving of the BCS bowls. There are often a lot of oddballs towards the bottom of the Top 25 however, showing a problem with how the rankings are compiled
 
Last edited:
You must log in or register to reply here.

Latest posts

Share this page

Top