| RANK W/OUT WEIGHT | OFFENSE RANK | DEFENSE RANK | RECORD | SS | ||||
| 1 | NEW ENGLAND | 1 | 1 | 18 | 10-3-0 | 28 | ||
| 2 | DENVER | 5 | 8 | 4 | 10-3-0 | 21 | ||
| 3 | HOUSTON | 2 | 4 | 7 | 11-2-0 | 30 | ||
| 4 | SAN FRANCISCO | 4 | 9 | 2 | 9-3-1 | 25 | ||
| 5 | ATLANTA | 3 | 11 | 16 | 11-2-0 | 31 | ||
| 6 | NY GIANTS | 6 | 6 | 8 | 8-5-0 | 12 | ||
| 7 | SEATTLE | 7 | 13 | 3 | 8-5-0 | 17 | ||
| 8 | CHICAGO | 8 | 27 | 1 | 8-5-0 | 10 | ||
| 9 | GREEN BAY | 9 | 7 | 19 | 9-4-0 | 10 | ||
| 10 | BALTIMORE | 10 | 12 | 14 | 9-4-0 | 19 | ||
| 11 | PITTSBURGH | 11 | 21 | 6 | 7-6-0 | 27 | ||
| 12 | CINCINNATI | 12 | 15 | 15 | 7-6-0 | 29 | ||
| 13 | WASHINGTON | 14 | 3 | 21 | 7-6-0 | 4 | ||
| 14 | INDIANAPOLIS | 13 | 16 | 28 | 9-4-0 | 31 | ||
| 15 | MINNESOTA | 15 | 18 | 23 | 7-6-0 | 14 | ||
| 16 | DALLAS | 16 | 14 | 17 | 7-6-0 | 2 | ||
| 17 | TAMPA BAY | 17 | 10 | 20 | 6-7-0 | 24 | ||
| 18 | ST LOUIS | 19 | 20 | 12 | 6-6-1 | 9 | ||
| 19 | SAN DIEGO | 18 | 28 | 5 | 5-8-0 | 23 | ||
| 20 | CAROLINA | 21 | 17 | 10 | 4-9-0 | 1 | ||
| 21 | DETROIT | 20 | 2 | 26 | 4-9-0 | 6 | ||
| 22 | NY JETS | 22 | 31 | 11 | 6-7-0 | 8 | ||
| 23 | CLEVELAND | 24 | 24 | 13 | 5-8-0 | 22 | ||
| 24 | NEW ORLEANS | 25 | 5 | 30 | 5-8-0 | 5 | ||
| 25 | MIAMI | 23 | 26 | 22 | 5-8-0 | 18 | ||
| 26 | PHILADELPHIA | 27 | 23 | 25 | 4-9-0 | 14 | ||
| 27 | BUFFALO | 26 | 22 | 31 | 5-8-0 | 26 | ||
| 28 | TENNESSEE | 29 | 25 | 27 | 4-9-0 | 7 | ||
| 29 | ARIZONA | 28 | 32 | 9 | 4-9-0 | 3 | ||
| 30 | KANSAS CITY | 30 | 30 | 24 | 2-11-0 | 20 | ||
| 31 | OAKLAND | 31 | 19 | 32 | 3-10-0 | 16 | ||
| 32 | JACKSONVILLE | 32 | 29 | 29 | 2-11-0 | 13 |
That's right. Tom Brady and Payton Manning are united again on the two best teams in the NFL.
What's more interesting is that the dismantling of a top ten defense by the Patriots has catapulted them from fourth to first.
But you may ask, "But how could you put them at first if they have such a low rated defense?"
Well my first answer is that I didn't decide on the final rankings specifically at all. I created a system by which I plug in weekly stats and the model computes the rest. My subjective biases are all within the model when weighting the importance of one stat versus another. My biases are not based on week to week impressions of bits and pieces of game highlights. Everything is, "just the stats ma'am".
But you may then ask, "But how could your lousy system make such a blunder by placing the Patriots ahead of two arguably better balanced teams with equal or better records? You must have some offensive bias. Yes that must be it. You set up a model where the offense is more important so that you can downgrade the Steelers rankings."
Certainly anyone who knows my feelings on the Steelers would agree that this theory has some face validity.
But my reply would be: My system is operating just fine thank you. And I'll ignore that snarky comment and point out that the Patriots so dominate the offensive categories that they overwhelm all other categories of measurement. I weight overall offense and defense score equally in the power ranking.
Never since the perfect Patriots team of 2007 has a team so dominated a side of the ball. The Patriots have a perfect score on offense. The next ranked offense (of the Detroit Lions) is 25% less in its score value.
To remind everyone, the one thing I do with all the scores is what I call a normalization process. I take all the differing measurements and distill them down into a like number. It's a way to take apples and oranges and make them comparable for this exercise. Additionally the normalization process doesn't just sort out and standardize the scores for different categories, it also allows for the distribution of the data to reflect through to the final ranking. It allows for just this case where one team has a large lead on the next competitor. In many ranking systems people just rank the offense on 1 to 32. It implies that each step down is equal. That the difference between the first and second team is the same difference as between the 30th and 31st. But in a real application of statistical measurement this may or may not be the case.
Which it isn't in this one. It almost never is the case that data is evenly distributed. That's the thing they don't really talk about in Urban Planning statistic classes.
To put this dominance in another light. The difference between the number one Chicago defense and the number two San Francisco defense is only 5%.
On the bright side the Browns have clawed their way into mediocrity. Yea!
I still want Shurmer canned.
Cheers!
No comments:
Post a Comment