November 20th, 2009
2009 NCAA College Football, Round 13
Just for grins, here is the graph of the college football season so far. This is every game between all 719 teams. First, I resolved all beatloops. Then, I took out all the teams that weren’t in Div1A, while still respecting the game outcomes. Meaning, if A->B->C->D, but B and C aren’t in Div1A, I absorbed them and changed the beatpath to A->D. Alabama is clearly ranked #1 according to this. No one in the Pac10 is even in the top 25. Click to embiggen. Wake Forest, huh?
November 20th, 2009 at 7:59 am
Wow, look at the Big East.
November 20th, 2009 at 7:59 am
Imagine if Wake had beaten Georgia Tech (go Jackets!); they’d be even higher. Good thing Tech went for it on 4th down.
November 20th, 2009 at 9:38 am
Looking at this chart more closely, your’re missing four FBS teams: Brigham Young, UTEP, Louisiana-Monroe, and Louisiana-Lafayette.
November 20th, 2009 at 11:59 am
Interesting. The non-conference games make for some interesting relationships that look really stable, so you end up with strange things like Baylor having a beatpath to Ohio St, and things like that.
My own alma mater (Iowa St) has managed to almost disconnect themselves from the rest of the Big 12 North, but are still under beatpaths from two teams they beat (Baylor & Nebraska). Still, they’ve had a better than expected season. It helps to keep expectations really low.
November 20th, 2009 at 12:27 pm
Yeah, in some ways this shows the limits of this approach when applied to a sparse data set. 30 teams having a beatpath to OSU is just weird.
November 23rd, 2009 at 6:31 am
Man, that graph is rough on my colleges. NMSU and UNM at the dead bottom.
November 24th, 2009 at 10:33 am
Ah-ha.. thanks for the tip, Chris. I’ll look for them if I run this again this season.