Jump to content
SiouxSports.com Forum

AFCA coaches' poll


Recommended Posts

A tidbit of info that I've actually found enlightening from a Bison site is the Massey Ratings. The Massey power ratings include all schools, from BCS to the NAIA level. The Massey system seems to properly weigh the NCC, so it's got to be a good system. :glare:

Here are the top D-II schools by Massey

Overall Wkly

Rank Change School WLT Rating

86 + 6 Grand Valley St 1 0 1.319

111 +43 North Dakota St 2 0 1.069

116 + 9 Valdosta St 2 0 1.056

136 + 2 North Dakota 1 0 0.941

137 +35 South Dakota St 1 0 0.935

143 +27 St Cloud 2 0 0.885

148 + 9 Saginaw Val 1 0 0.856

150 - 1 Carson Newman 2 0 0.843

153 - 2 Pittsburg St 1 0 0.828

157 + 8 Indiana PA 1 0 0.819

160 -44 NW Missouri 0 1 0.805

The Div-III school, Mt. Union, that is the primary target for Div III reforms which hurt CC, Clarkson, RPI, is rated above a number of Div IA schools.

Link to comment
Share on other sites

Hold it here. There are a couple of obvious questions:

1. How in the world is SCSU rated below three other NCC schools?

2. In 2002 there were 117 DIA schools and 123 DIAA schools for a total of 240. You're telling me that current DIIs UND and NDSU are rated ahead of roughly half of them?

Link to comment
Share on other sites

1. Massey isn't a poll, he reads in scores, does some regression analysis, and spits out a rating. NDSU is rated so highly because Montana was rated very high before. If Montana and Tusculum keep winning (and the teams that they beat keep winning), NDSU will continue to rise. SDSU had another big win. I think that the combination of those two wins may have bouyed the whole conference somehow.

2. Yeah, Massey rates 699 schools so NDSU is rated pretty highly. For example, the four schools immediately below NDSU in the Massey Ratings are DI-A schools and NDSU would be the 12th rated DI-AA school. What does it mean? Nothing - especially this early in the season.

If you really want to freak out, check out who is number 1 in the BCS poll: Holy Cow

Link to comment
Share on other sites

I've checked out Massey's ratings before because he has been doing college hockey for a long time. College football presents unique problems in that hundreds of football teams often play only 10 or so games each, far too sparse a matrix to get a significant result from a Bradley-Terry comparison. Further, people want comparisons immediately, when each team has only played 1-2 games!

So, he attempts to mathematically immitate the process that human pollsters use to produce "meaningful" polls with sparse data. He tries to correct for too little information at the beginning of the season by including last season's results and reducing their weight as the season progresses. Also, for each game he calculates a likelihood of that outcome being repeated (based on the scores). Recognizes the problems that can create (some teams could always play close games, or play up/down to certain opponents), he measures each team's actual deviation about the likely outcome and incorporates that, so a team which always finds a way to win gets credit for that.

Basically, over enough games, his model should settle into something close to a Bradley-Terry analysis, but in the short term he tries to correct for sparse data by incorporating many of the factors that human pollsters consider other than just wins/losses. Such "corrections" are necessary in football because there aren't enough contests to assume that random events are reasonably evenly distributed between helping a team win/lose.

Link to comment
Share on other sites

If you really want to freak out, check out who is number 1 in the BCS poll: Holy Cow

What's that? Valdost... !?!?! .....

Oh dear God! I'm blind!

Kidding. :glare:

Trying to rate college football with this limited sample size (few games) and huge, mostly empty, matrix (10 games per team with ~700 teams) is a statistical nightmare. I can't imagine how long it would take to make a Bradley-Terry-type calculation stabilize.

There are B-T and similar calculations done during the college hockey season (don't complain Bison fans, it's the Sioux board :huh: ) and I don't look at much less believe those until after the first of the year (mid-way through the season).

Link to comment
Share on other sites

Well, I don't have any idea what a Bradley-Terry thingie is (impressive) but I do know that I almost melted my PC trying to figure out how many steps it would take for each DI-AA team (plus NDSU) to declare they had beat a team, that beat a team, etc..., that beat the DI-AA champs. If these power ranking calculations do the same type of recursion, I don't see how they come up with the rankings so quickly.

For the record it took NDSU 15 steps last year (worse than almost any DI-AA team). I'm not sure this qualifies as interesting but the team Valdosta beat that was supposedly a top 25 team in DI-AA needed the same number of steps. I kind of did a triple-take when I saw that they were rated - they were worse than NDSU last year.

Link to comment
Share on other sites

Borrowed from a USCHO page:

The (Bradley-Terry) rating method was first invented in 1929 by a German named Zermelo to evaluate the results of a chess tournament in which a full round-robin was not completed. In 1952 a pair of Americans, Bradley and Terry, unaware of Zermelo's work, rediscovered the method while trying to model the outcomes of taste tests, and this rating system came to be called the Bradley-Terry method. In the 1990s, an English statistician named Kenneth Butler, studying in Canada, decided to apply the Bradley-Terry method to US college hockey, and when prodded for a name chose Ken's Ratings for American College Hockey (or KRACH).

Pages that explain the B-T method (in various levels of detail):




Some of those pages are better than Sominex. :huh::glare:

Link to comment
Share on other sites

  • 2 weeks later...

The Sioux are back in the poll at number 25:


I can't understand this love affair the pollsters have with UMD. They've lost to USD, who just lost to a dIII team. Their one win over an NCC team was over probably the worst team in the league, MSU-Mankato. Their other win was over a pretty bad Southwest St. I realize UMD had a good year last year, but that loss to USD should have made it clear that this is not the same team as in 2002.

Link to comment
Share on other sites

I'm also wondering why UMD is so loved by the pollsters. UMD's win over USD was definitely diminished by USD's loss to a DIII team. It is laughable when the polls come out, regardless of what sport. Though they claim the polls are not biased, you know that the votes being cast are in direct correlation to the attitudes and feelings coaches feel towards other programs.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...