Pac-12 FB Strength

[dystopiamembrane]

Massey’s CFB composite, a composite of 80+ rating systems, offers the following list of P5 conference relative strength (highest rated team and number of schools in top 25 in parenthesis).

1. SEC (1-Georgia; 5)
2. Big Ten (2-Michigan; 7)
3. Big 12 (5-Oklahoma State; 3)
4. ACC (14-Wake Forest; 4)
5. Pac-12 (15-Utah; 2)

Note: The Pac-12 continues to be shown to be closer in strength to the MW and AAC than it is to the other P5 conferences.

[Xanthis]

The ACC is garbage being propped up by Clemson’s past success and 8 conference games. The PAC definitely has to get better as a whole, but its still an apples to oranges comparison. 

[dystopiamembrane]

I would assume that close to zero of the 80+ rating systems use previous season data in calculation at this late period of the season.

[dystopiamembrane]

#14 10-2 Wake Forest, #16 9-3 Clemson, #17 10-2 Pittsburgh, #19 9-3 North Carolina State

[Xanthis]

Isn’t part of the SoS calculated based on rankings? Rankings which are extremely biased and still rooted in prior years since beggining rankings matter quite a bit throughout the whole season. Slap another conference game in there and maybe Wake or Pitt is 10-2. 

The PAC had wins against LSU and Ohio St, but BYU running through the PAC 12 teams on their schedule seems to have erased those. Utah’s loss to SDSU isn’t as bad as perceived. The perception of the PAC is silly and it’s mostly because of the PAC 12 network, late games, and a couple of years without marquee non conference wins. This year the conference gets those wins, but is still looked at as weak. 

I’m not familiar with Massey’s stuff, but don’t the computer models usually have the PAC 12 better than the human ones?

[dystopiamembrane]

The link to the composite is in my original post. Scroll down to the bottom of the page to see the Conference ratings.

[Johnny]

All computer rankings start with a ranking based on last year’s data and human opinion and it goes from there. It’s all stupid

[dystopiamembrane]

Yes, in the beginning most ranking systems are seeded with previous season data. However, this late in the season, these systems should all be using only current season data. Also, the size of the data set that is a college football season is far too large for my mind to rank teams. I am the stupid one. I rely on computers for that.

[2008 National Champ]

Bill Connolly’s (sic?) never completely eliminate previous year’s data since returning production and 5 year recruiting are part of S&P and SP+. I like the work he has put into it but my main problem with his systems is that they use multiple year data to evaluate a single season.

A true resume analysis is what you are describing but I don’t believe that every ranking used in the Massey composite limits itself to that.

[dystopiamembrane]

PS Does it seem strange to anyone else that there are people questioning whether or not computers are the best way to mine data?

[Xanthis]

Give me the computers always. Sure they aren’t always right, but they aren’t biased like us. 

[dystopiamembrane]

[2008 National Champ]

computers are nothing more than fancy calculators. the answers they give are only as good as the data entered and the predetermined weighting of the values.

To Dystopian’s point, computers are the most efficient way to mine data. The relevance of the data mined is always suspect as it depends on the bias inherent in whomever created the key word search.

[Author]

Posts