Heavy Lifting - thoughts and web finds by an economist
     I also contribute to Division of Labour Load HL's Front Page
Monday, September 18, 2006

On the Top 25 rankings

I have contended that removing the strength of schedule component to the BCS rankings at the end of the NCAA football season is not that important. Why? Because pollsters take the strength of schedule into account when making their rankings.

Anecdotal evidence is now available.

Last week UGA was ranked #10 in the AP poll. We won over Alabama - Birmingham 34 to 0. Three teams in the top ten lost over the past weekend, Notre Dame, LSU, and Florida State. With three teams losing ahead of you, you would expect UGA to climb to #7 or #6. Nope, we climbed one spot to #9.

Meanwhile, Michigan, who beat Notre Dame (#2 last week), jumped from #11 to #6, and Louisville, who beat Miami (FL) (#17 last week), from #12 to #8.

What to take from this? The pollsters seem to have rewarded UGA with the minimum possible given that three teams ahead of UGA lost over the weekend. On the other hand, Louisville and Michigan are strongly rewarded for winning and for winning against harder opponents.

[Updated Sept. 19]: If we look at the USA-Today Coaches poll, we see that Georgia moved from #9 to #7, while Louisville moved from #12 to #9 and Michigam moved from #13 to #6. In the Coaches' poll, UGA jumped over Texas whereas in the AP poll Texas stays one ahead of UGA despite Texas having one loss.

This suggests that the Coaches reward winning moreso than who you win against. This suggests that the writers are actually taking into account more information than the coaches when making their ranking decisions. This runs counter to conventional wisdom which claims the coaches have more and better information on which team is better and the writers have their regional biases. I haven't proven the bias/no bias issue, but anecdotal evidence might point in the opposite direction.

[end update]

Thus, at least with the AP poll, perhaps I am correct that strength of schedule doesn't need to be included as a separate element in the BCS calculation.

Two thoughts:

1. If you could get your hands on the data, your hypothesis about one poll being a better reflection of schedule strength than another would seem to be empirically testable.
2. My main problem with the polls is how much they are influenced by starting position (i.e. first pre-season poll). If Team A is ranked first overall and Team B is ranked tenth overall pre-season, and both teams win the same number of games versus similar competition, Team B can never seem to pass Team A in the polls, even if they are playing a little bit better football than A.

Love your blog...keep up the good work.

Post a Comment


Le Chai - galerie du vin



Posts that contain Craig Depken per day for the last 90 days.


Heavy Lifting's Main Page
Email Me
Atom Feed

Heavy Lifting

Great Links

Money I Found Today

Heavy Lifting - Firehose style (56k warning)

Recent Posts

- On the Top 25 rankings



Site Meter Blogroll Me!


Modified maystar design
powered by blogger