When Alabama lost to Texas A&M this weekend, it left no undefeated SEC teams. In response, my friend Chris Haddock, a family practice doctor and raging Georgia Bulldogs fan, posted this on Facebook:
UGA has one win against a BCS Top 10 opponent. Florida has three. Alabama has two. Comined Kansas State, Oregon, Notre Dame have ZERO. Nada. Zilch. If the SEC Championship Game winner doesn’t play for the national championship then it is a tainted title to whichever of these undeserving non-SEC teams wins.
Since the SEC has won the last 6 national championships, many will share this sentiment.
And our numbers agree. Alabama only dropped to 2nd behind Oregon in our rankings, and 6 of the top 10 teams come from the SEC. With this many good teams in a conference, you expect every team to lose at least once.
The question becomes whether the SEC champion, either Alabama or 7th ranked Georgia, makes the BCS title game. For this to happen, two of these 3 teams must lose.
- Team rank: 1st.
- Probability of remaining undefeated: 58.7%.
- Toughest remaining game: 76.3% chance to beat Oregon State.
- Team rank: 3rd.
- Probability of remaining undefeated: 64%.
- Toughest remaining game: 80% chance to beat Texas.
- Team rank: 8th.
- Probability of remaining undefeated: 41.4%.
- Toughest remaining game: 44% chance to beat USC.
Somehow, I don’t think I’ll be able to publish Chris’s next post if Notre Dame squeaks past USC and into the national title game ahead of the SEC champion.
Last week, Alper Akanser, a Ph.D. student at Georgia Tech, took our undefeated probabilities and determined the chance that 0, 1, 2… teams would remain undefeated. This week, his code gives a 42.7% chance that 0 or 1 teams remain undefeated, opening the door for the SEC champion.
That’s pretty good odds for the SEC.
All the proves is confirmation bias. If people think the SEC is very good and then SEC teams lose a game to each other, they don’t get knocked down because everyone thinks its such a tough conference that one loss is ok.
South Carolina has lost two games and their best win is against Georgia. The next best win would be what? Missouri, Kentucky, Tennessee? Georgia’s best win is Florida but they lost badly to a South Carolina team that hasn’t proven a thing against anyone of substance. You can continue to use this logic.
Ed Feng says
Well, I certainly have my biases. And so do my friends down South.
But my rankings do not. There’s nothing human about the 5 SEC teams in the top 10 right now.
Crap, I made a mistake. There are 6 SEC teams in the top 10. I skipped LSU.
I can respect a quant based approach and its certainly better than the always subjective eye test, but things seem to get muddied when the only above average team a particular team plays is against another team that does the same thing. Sounds confusing so I’ll lay it out like this.
How do we know South Carolina is good? Their 6 conference wins have come against teams with a combined conf record of 13-21. Eliminate Georgia and its 6-20. But Georgia is a good win. Or is it? Their 7 conf wins have come against opponents with a combined SEC record of 15-33. Eliminate Florida and it’s 8-32. So South Carolina is good because they beat Georgia who beat Florida?
This seems to a lot of transitive property at work. Yes, the SEC has some very good teams, but it’s ridiculous to say that the top 10 teams in the country are all in that conference when the conference itself is very top heavy. So each team loses a game to another top team and its fine, ignoring the fact that the vast majority of their wins have come against bad teams. We don’t know how good the SEC is because for the most part, they don’t play anyone of merit outside the conference and pad their schedules with bottom tier FBS and even FCS schools.
Ed Feng says
The 6 SEC teams in the top 10 results from 2 factors:
1. large margin of victory in out of conference wins. Bama over Michigan comes to mind first.
2. these 6 teams with big margin of victory over the lower teams in the conference.
While the SEC certainly doesn’t challenge itself out of conference as well as it should, they do have big games coming up: South Carolina against Clemson, Florida against Florida State. We’ll see how these games turn out, as well as bowl games.
Asked another way: What out of conference wins are making the sec so strong? Presumably your metrics wouldn’t reward for in conference unless those were validated by out of conference results. Can you point to a few(?) games whose results are propelling the sec to the top? If one or two had gone differently, would it affect you rankings in any material way?