Time for week 2 of Does Your Team Suck. Do the Broncos suck? In the world of college football, the transitive property seems to be entirely forgotten and rightfully so. Houston is a playoff contender? Navy beats Houston. Air Force beats Navy handily. Air Force loses a close one to New Mexico. All hail New Mexico, a true playoff contender (who happened to lose to Rutgers). Does Rutgers suck?
All of this merely goes to show how difficult it can be to truly assess the quality of a team. It should be remembered, that this rating is meant to assess the quality of a teams wins, it does not include expectations about future games. For example, it is quite likely that Alabama will win its next 5 games; however, those games still need to be played, so this rating system would not make such an assumption. This should be kept in mind when reviewing the results. Eventually there should be some correspondence between body of work and quality of team, but there is a lot of football to be played.
For an intro to the model, please look at DYTS intro
Several people believe that losses should be included. While the model doesn't penalize losses, losses do change the model by affecting conference strength assessments. This can have a profound effect. In summary, a teams loss has systemic effects even it they aren't direct. This also makes calculations easier.
Again, for simplicity, I have decided not to include scoring in the model. While this would certainly improve things (because it would more closely correspond with the eye test), it makes things more complex than I would like at this time. I may include scoring differences next year. Including scoring can be problematic if you have a coach who eases up in the second half and plays backups or if you have a coach who likes to run up the score against hapless opponents.
I appreciate most of the feedback given about the intro. Egnowit was especially helpful at pointing out some glaring problems with my initial attempt.
I fixed a couple of these problems, but I have yet to address the problem that arises when a team (using the eye test) is 1-0 in conference but has performed poorly OOC gives its winning opponents more credit than they "seem" to deserve. This is not unexpected; the model is based on wins and not expectations. I haven't found a satisfying way to include this in the model yet, but I may include it next year. (If I'm still interested).
The conference strengths have been adjusted to reflect this weeks results. Here is what I obtained:
The 2nd column represents the absolute strength of a conference based on its OOC wins (adjusted for conference size). The 3rd column represents the change from the previous week's rating. The 4th column is the most important. It represents the strength of a conference after adjusting for quality of wins. The 5th column indicates the change from the previous week. The last column shows ordinal changes in ranking.
The Big 10 grew much weaker this week, but is still #1, slightly edging out the ACC, who grew the most. The Big12 strengthened but is still much weaker than expected. The biggest result for us is that the MAC is stronger than the MWC. It's not by much, but it means that Western Michigan's in-conference wins are more valued than ours.
Using the 4th column above, I calculated the strengths of each team's wins. Here is what I obtained:
|San Diego State||MW||0.1240||36||↓5|
|New Mexico State||SB||0.0728||62||↓12|
|San José State||MW||0.0266||97||↑20|
|Texas San Antonio||CUSA||0.0200||T102||↓9|
The model has obtained some surprising results. I especially enjoyed how different it responds than the typical poll rankings. BSU obviously weakened because of Beaver and Cajun losses and because their conference strength weakened. Tennessee got embarrassed by Alabama, but is still ranked #2. The model still respects their 5 wins.
Western Michigan has jumped the Broncos (BW BW to BW). They have one more win than we do, but still. Hopefully, our conference strength can improve (we can do that with a victory against BYU this week). North Carolina and Utah get some much due respect. Stanford (which started this whole mess) is certainly not a dumpster fire, although Mcaffery's injury could hurt them later.
California is somehow still hanging around. (Oh yeah, they beat Utah, who is currently atop the Pac-12 South.) Georgia lost to Vanderbilt and gets a nice boost in the rankings. Syracuse is the biggest gainer with their victory over VT. Does VT suck? Does Syracuse suck?
Hope you enjoy the results!
I failed to include Charlotte's win over FAU last week.
I mistakenly used FAU instead of FIU when calculating the OOC strength of Indiana.