clock menu more-arrow no yes

Filed under:

Can SOS be misleading?

New, 1 comment

One of my Sunday rituals is to take a gander at the latest Sagarin ratings (and similar statistical offerings) before submitting my draft blogpoll ballot. According to the latest iteration, Rutgers comes in at #68 in the country based on what they've accomplished to this point, having played the #140 schedule in the country.

I'm not trying to posit that Rutgers has played particularly well to this point, as they haven't. However, I think there is a bit of a quirk in these computer polls that can distort the picture of everything that's going on, to some extent.

As an example, let's compare the schedules to this point of Rutgers and USF. They were chosen because both have a win over an ACC team, and have played similarly bad schedules, featuring two FCS teams.

Rutgers

(8) Cincinnati L 47-15
(228) Howard W 45-7
(135) FIU W 23-15
(115) @Maryland W 34-15
(196) Texas Southern W 36-0
(18) Pittsbugh L 24-17
(144) @Army W 27-10

USF

(187) Wofford W 40-7
(178) @Western Kentucky W 35-13
(212) Charleston Southern 59-0
(44) @Florida State W 17-7
(100) @Syracuse W 34-20
(8) Cincinnati L 34-17
(18) @Pittsburgh L 41-14

Rutgers and USF have played fairly comparable schedules up to this point. Rutgers had the worse loss to Cincy, but keep in mind that Tony Pike missed most of the second half of that game, and that Dom Natale's turnover meltdown sent the opener out of hand. In comparison, Rutgers looked far, far more competitive against Pitt. Based on the more-recent history, Rutgers is the better team right now.

As far as I can tell, what's really skewing the Sagarin ratings are the fact that FSU is ranked much more highly than Maryland is. However, I'm not that sure if there's that much of a wide gulf between the teams. Maryland narrowly lost to MTSU, while FSU pulled away from Jacksonville State late. A few bounces here and there, and both games turn on their heads. At some point, there is a lower boundary or floor with poor teams. Alabama would beat #80 about as bad as #200.

The purpose of this post isn't to rail against computer rating systems. I think that they're actually far more accurate than human polls, even if each still has its own particular blend of oddities and eccentricities. By the end of the season, with the benefit of a larger, more accurate sample size, everything should look a lot more intuitive. However, these current results are very difficult to square with what's actually going on in games at the moment. USF may have the better resume to this point, but they're not 32 rankings and 7 points on a neutral field better.

Which is, not to single out USF, although I love picking on their program at times. At the moment, all human and computer polls have their share of strange, unintuitive results. We'll just have to wait until the end of the season, and see how everything plays out to know where exactly the chips fall.