With week three over (and week four just set to begin), we can construct a non-biased power rating that might give us some clues as to how things are shaping up.
Of course, the conclusions drawn here don't rest on solid footing yet. Many teams have played just two games, and even those who have three games faced opponents who have played just two games. That's not enough to construct a reliable foundation for comparative rankings. At this point, our Bayesian modifications to our pre-season rankings are a better bet.
But the time will come to completely throw away pre-conceived notions, and when we do, what will we see? Here's a sneak peek (that is guaranteed to be wrong for many teams). The following is the top 25 among all teams from all divisions; we use that because most FBS teams have played an FCS team and without including those teams and their opponents, we have even less to go by.
Strength power rating; unbiased top 25 through week 3, 2013:
num Team record rating offense defense
1. Georgia Tech 2-0 92.65 73.50 [ 3] -19.14 [ 4]
2. Baylor 2-0 89.84 91.91 [ 2] 2.06 [ 56]
3. Washington 2-0 80.48 59.00 [ 36] -21.47 [ 3]
4. Florida St 2-0 77.49 53.37 [ 69] -24.11 [ 1]
5. Oregon 3-0 74.72 71.04 [ 6] -3.67 [ 27]
6. Navy 2-0 70.94 62.67 [ 17] -8.27 [ 13]
7. Augsburg 1-0 68.90 95.01 [ 1] 26.11 [356]
8. Indiana 2-1 64.86 72.73 [ 4] 7.87 [ 96]
9. Duke 2-1 64.53 61.08 [ 25] -3.45 [ 29]
10. Illinois 2-1 64.49 67.71 [ 7] 3.21 [ 61]
11. Alabama 2-0 63.26 67.51 [ 9] 4.24 [ 70]
12. UCLA 2-0 62.88 52.32 [ 73] -10.55 [ 11]
13. Oklahoma St 3-0 62.20 61.58 [ 22] -0.61 [ 41]
14. Florida 1-1 61.79 46.15 [124] -15.64 [ 6]
15. Montana 2-0 61.68 48.34 [102] -13.34 [ 9]
16. Missouri 2-0 61.01 53.31 [ 70] -7.70 [ 15]
17. Towson 3-0 59.48 57.52 [ 43] -1.95 [ 34]
18. Miami FL 2-0 58.09 59.33 [ 34] 1.23 [ 54]
19. Utah St 2-1 57.57 61.32 [ 23] 3.75 [ 64]
20. Washington St 2-1 57.39 53.69 [ 63] -3.70 [ 26]
21. Wabash 1-0 57.20 61.71 [ 21] 4.51 [ 73]
22. LSU 3-0 56.56 62.70 [ 16] 6.14 [ 83]
23. Auburn 3-0 56.39 56.95 [ 46] 0.56 [ 49]
24. Maryland 3-0 55.80 51.88 [ 77] -3.92 [ 24]
25. Cincinnati 2-1 55.50 60.91 [ 26] 5.41 [ 77]
Clearly you can see that for every possible keen insight, there is an obvious blunder. First we note that almost all of the very top teams have played just two games (it's much easier to look out-of-this-world for two games than three). Georgia Tech and Baylor have been great, no doubt, but their rating exceeds last year's rating for Alabama (who ranks #11 here). Oregon is the top 3-game team at #5, which seems pretty realistic, and Washington and Florida State might be top five teams. How good is Navy? We don't know but based on their two games they might be very good. But they're just 2-0. They beat Indiana, the top team with a loss in this early snapshot. Note that Duke and Illinois in turn lost to Georgia Tech and Washington, and it's still possible to have "fair" rankings.
Then there are the 1-game wonders like Augsburg (Augsburg?) showing up at #7, and Wabash at #21. We'll go out on a limb and guess those teams aren't really top 25 quality. Nor Towson, despite their 3-0 start, or Montana, no matter how well they do in the FCS.
The rest of the top 25 is full of FBS teams already, which is a sign of things starting to coalesce. There are clues, however, that it hasn't quite cooked long enough, such as Alabama's #70 defensive rating. When you think of it, it makes some sense—they gave up 42 points in their last game. And we haven't seen enough of Texas A&M to know that they have a great offense (remember, we don't know anything about Johnny Football from last year, just this year's scores).
Looking further down the top 732, we see some interesting points. Louisville at #29, and USC at #31—will they be moving up or is that how good they are? Georgia at #35 and Clemson at #37 should both be moving up I would think.
We find #38 Eastern Washington (2-1) one spot ahead of Oregon State (2-1), whom they beat in overtime. Ohio State and Wisconsin are #49 and #50...for the Big Ten's sake, they should advance from there. Being rated low at this point is nothing to be worried about if you've played well; even #100 Notre Dame can be sure that's not their real ranking, but it does reflect how pedestrian their play has been. And at #213 Texas could be in trouble. Not that they'll remain there, but you can see how they got their, losing badly to BYU, who lost to Virginia, who got smoked by Oregon.
Early-season unbiased power ratings are interesting more for the questions they raise than predictive value. Sometimes they shine the light on an unheralded team, while other times a team with a fluke game looks like a big winner before their subsequent play forces a regression to their (actual) mean. We'll be taking an extra look at Georgia Tech, Baylor, and Washington this week, no doubt (Navy doesn't play, unfortunately): which ones will live up to their early promise?
Comments