Professor Frink’s Power Rankings, Week 10
by tom slick on Nov 17, 2009 2:18 PM EST
Email
Print
via i37.tinypic.com
Bummed about the Panthers game? Do as Professor Frink does when HE's feeling blue! Immerse yourself in warm, soothing data. Ahhh...
RATINGS COMPUTATION (must-read for newbies):
At its heart, this is a VERY simple method of calculating ratings, using only two stats. The results aren't meant to be a perfect reflection of the real world; don't focus too narrowly on exactly where each team's totals are ranked. The two components (O and D ratings) are often more revealing than the totals, don't overlook those.
Here's how the ratings are calculated. The first step is to calculate two simple stats: Points For per Game, and Points Against per Game, for each team. These two leaguewide data sets for offense & defense both conform to a "Bell Curve", or Normal Distribution, that can be graphed. Depending on where each team lies on that graph, they have a value between 0 and 1. (This isn't something pulled from thin air. It's a very common, standardized statistical analysis technique. The purpose is to more accurately measure each team's position within the ‘population' of all league teams.) You now have every team rated from 0 to 1, on offense and on defense. Add those two numbers and you've got a total rating for each team. Last, multiply by 50 to convert the ratings onto a 0-to-100 scale, making them more readable and intuitive. There you go. Simple, right?
|
Rank |
Power Rating |
|
W-L-T |
PF/G |
PA/G |
|
Offense Rating |
Defense Rating |
Total |
|
|
1 |
90.1 |
9-0-0 |
28.0 |
15.8 |
0.856 |
0.946 |
1.802 |
|||
|
2 |
89.9 |
6-3-0 |
28.8 |
16.7 |
0.883 |
0.914 |
1.798 |
|||
|
3 |
79.3 |
5-4-0 |
24.7 |
17.1 |
0.692 |
0.894 |
1.585 |
|||
|
4 |
77.7 |
8-1-0 |
30.1 |
20.4 |
0.922 |
0.633 |
1.555 |
|||
|
5 |
74.5 |
6-3-0 |
24.9 |
18.8 |
0.705 |
0.786 |
1.491 |
|||
|
6 |
73.6 |
9-0-0 |
36.8 |
21.9 |
0.994 |
0.479 |
1.473 |
|||
|
7 |
73.2 |
6-3-0 |
23.0 |
17.4 |
0.587 |
0.876 |
1.463 |
|||
|
8 |
72.4 |
7-2-0 |
22.0 |
16.3 |
0.521 |
0.928 |
1.448 |
|||
|
9 |
72.1 |
5-4-0 |
25.8 |
19.9 |
0.754 |
0.688 |
1.442 |
|||
|
10 |
72.1 |
5-4-0 |
26.9 |
20.4 |
0.809 |
0.633 |
1.442 |
|||
|
11 |
69.9 |
4-5-0 |
22.1 |
17.6 |
0.528 |
0.870 |
1.398 |
|||
|
12 |
68.4 |
6-3-0 |
25.4 |
20.4 |
0.736 |
0.633 |
1.369 |
|||
|
13 |
61.5 |
5-4-0 |
23.9 |
20.9 |
0.644 |
0.586 |
1.231 |
|||
|
14 |
60.4 |
6-3-0 |
18.6 |
16.8 |
0.299 |
0.910 |
1.209 |
|||
|
15 |
60.1 |
6-3-0 |
26.3 |
22.4 |
0.783 |
0.419 |
1.201 |
|||
|
16 |
60.0 |
5-4-0 |
24.6 |
21.6 |
0.685 |
0.515 |
1.200 |
|||
|
17 |
57.5 |
5-4-0 |
25.8 |
22.7 |
0.754 |
0.395 |
1.149 |
|||
|
18 |
54.7 |
4-5-0 |
20.4 |
20.0 |
0.417 |
0.677 |
1.094 |
|||
|
19 |
46.0 |
3-6-0 |
15.6 |
19.0 |
0.151 |
0.768 |
0.919 |
|||
|
20 |
45.3 |
3-6-0 |
20.8 |
22.0 |
0.439 |
0.466 |
0.905 |
|||
|
21 |
43.1 |
4-5-0 |
20.7 |
22.3 |
0.432 |
0.431 |
0.862 |
|||
|
22 |
41.6 |
4-5-0 |
24.2 |
25.2 |
0.665 |
0.168 |
0.833 |
|||
|
23 |
31.7 |
Panthers |
4-5-0 |
19.6 |
23.9 |
0.360 |
0.275 |
0.634 |
||
|
24 |
31.1 |
5-4-0 |
20.1 |
24.4 |
0.395 |
0.227 |
0.622 |
|||
|
25 |
24.4 |
3-6-0 |
21.0 |
28.3 |
0.454 |
0.035 |
0.489 |
|||
|
26 |
23.9 |
3-6-0 |
15.6 |
23.3 |
0.151 |
0.327 |
0.478 |
|||
|
27 |
21.7 |
2-7-0 |
15.8 |
23.9 |
0.160 |
0.275 |
0.435 |
|||
|
28 |
13.9 |
2-7-0 |
9.8 |
24.1 |
0.023 |
0.255 |
0.277 |
|||
|
29 |
13.5 |
1-8-0 |
17.4 |
28.4 |
0.238 |
0.033 |
0.270 |
|||
|
30 |
9.9 |
1-8-0 |
8.7 |
25.0 |
0.014 |
0.184 |
0.198 |
|||
|
31 |
9.2 |
1-8-0 |
15.9 |
29.3 |
0.165 |
0.019 |
0.183 |
|||
|
32 |
4.5 |
1-8-0 |
11.1 |
27.7 |
0.038 |
0.052 |
0.089 |
|||
|
|
|
means |
21.7 |
21.7 |
|
The Ravens' shutout of the lowly Browns worked wonders for them here, who knew they were "better" than the Saints? What a revelation! Colts were (fittingly?) back at #1 this week. Jets appear to be the best darn mediocre team around, these ratings absolutely love them this year. Only a handful of teams moved in the rankings much this week. Our Falcons fell 4 spots, exceeded only by the Eagles, down 7. Climbers were led by the Vikings, up 5 slots, Ravens up 4.
There's our Fun With Numbers for Week 10. I hope you reaped a nugget or two of insight, and take something good away from an otherwise forgettable Falcons weekend. Download the spreadsheet here, if you wish. Salud!
This FanPost was written by one of The Falcoholic's talented readers. It does not necessarily reflect the views of The Falcoholic.
24 comments
|
3 recs |
Do you like this story?
Comments
Wrongs have been righted!
The Eagles are now solidly below the Cowboys. No need to change the algorithm anymore :-)
by One.Cool.Customer on Nov 17, 2009 2:23 PM EST reply actions
Bengals fans,
if they blog, would probably be none too happy seeing the golden bottoms of the Steelers above them after beating them twice this year.
Speaking of the AFC North. I propose a “Browns deduction” be introduced in ranking teams fortunate enough to play this hapless, leaderless squad of misfits. The Ravens, getting to play Cleveland twice, beating them collectively 50 to 3 in the process, unfairly skews up the Raven’s defensive prowess. Perhaps throwing out the one “outlier” on every team’s schedule (rather it be a blowout win or loss) would bring more clarity as well.
All in all I like these rankings, but would like to see a “rolling” 3 or 5 week breakout (i suppose i can do this myself if i download the spreadsheet?) vs. using the whole season. Teams like the Panthers, Titans, and Chargers are playing much better than their cumulative ranking indicates in this poll.
Stats-based ratings don't care who beat who
Some people expect these rankings to look like the “composite poll” or “best guess” types, and they’re not. There may even have been a complaint or two from Saints fans here, I don’t quite remember! Not everyone actually reads how the ratings are calculated, I guess…
The Ravens outscored Cleveland 50-3? I kind of feel bad for that wretched franchise. It wasn’t terribly long ago that the Falcons were way down there, too. I see the point in throwing away outliers, a lot of knowledgeable people do that. Unfortunately, that’s too much work and complexity for this old goat. Copying and pasting totals into a spreadsheet is about enough for me!
Your mention of rolling averages brings up the larger issue of measuring “momentum”. Some people question whether it even exists. If it does exist, how in the world do you define it? That has to be one of the most elusive, ethereal “stats” out there. If one could actually quantify such a thing, I bet you could make a freakin’ fortune in Vegas with it. I do hear you about teams that are playing better than their cumulative records indicate, though.
You may enjoy the fine Power Ratings section of sister site Blogging The Boys, link here. Our friend One.Cool.Customer has compiled rankings and links from many different sources there. Thanks for sharing your perspective.
This is interesting
And I wish I had thought about this when we were talking about adding something to it to make it more real-world like (which I still disagree with). Throwing out a team’s best win and worst loss is a common practice used to remove any fluke games, along the theme that a team is never as good as their best nor as bad as their worst throughout the year. This could be used here to remove any statistical anomalies and possibly give us a more true representation of where the teams should be ranked.
I’m in favor of it if for no other reason than curiosity sake. Downloading recent data now! Yay for numbers.
Corey Zickefoose: I think they should still be able to play football, regardless. Tennessee is my place. You know, it's my football team.
Reporter: Even after they put a gun in your face, you say let them play football?
Corey Zickefoose: Yeah. It's Tennessee. That's the way it is sometimes.
Welcome to the SEC!
Or not
Hmmm, the link keeps redirecting me back to the main page for FileDropper. Did you make your file public?
Corey Zickefoose: I think they should still be able to play football, regardless. Tennessee is my place. You know, it's my football team.
Reporter: Even after they put a gun in your face, you say let them play football?
Corey Zickefoose: Yeah. It's Tennessee. That's the way it is sometimes.
Welcome to the SEC!
Agreed
It’s a valid concept. The problem for me is that this raises the complexity a couple of orders of magnitude. Instead of simply copying and pasting cumulative data each week, one has to maintain stats for each team’s O and D, for each week. Considering all the extra work, I’m not sure that the somewhat improved results would be worth it (for me, at least).
Filedropper has been almost as inconsistent as the Falcons lately! Keep trying the D/L, I’ve always been able to get it to work eventually.
And that's cool
I still feel there is no need to change what you have as I believe it provides a very simple and quick way to compare teams. All I’m going to do is take your spreadsheet, adjust some values, and let the formulas do the rest. It will be something quite easy and quick for me to keep up with, much like point differential.
Corey Zickefoose: I think they should still be able to play football, regardless. Tennessee is my place. You know, it's my football team.
Reporter: Even after they put a gun in your face, you say let them play football?
Corey Zickefoose: Yeah. It's Tennessee. That's the way it is sometimes.
Welcome to the SEC!
Ok
So I finally got the file downloaded and got to work trying to figure out the best method to attack this. At first I thought about only removing the highest PF points and the highest PA points. I got about halfway in then realized that I really didn’t like that method because it wasn’t completely removing the game, just half of the score.
I started over and am now removing the best and worst point differential game. Since the original is based on avg PF & PA, I figured this would be the most logical approach. I’m not finished yet because I haven’t found a good way to single out these games quickly and easily for all 32 teams at once, but an early glance is revealing not much significant change. I think it probably has a lot to do with the fact that I’m essentially removing two games played and with only seven games worth of data it doesn’t really show much change.
Doing it one game at a time does take some time though so it might not be something I’d really want to do on a weekly basis unless I was accounting for it from the beginning.
Corey Zickefoose: I think they should still be able to play football, regardless. Tennessee is my place. You know, it's my football team.
Reporter: Even after they put a gun in your face, you say let them play football?
Corey Zickefoose: Yeah. It's Tennessee. That's the way it is sometimes.
Welcome to the SEC!
Lots of work, indeed
Sounds like a royal pain all right, especially to pick up in the middle of a season. Not enough bang for the buck IMO.
If I was going to do that large a volume of work, I’d start over and code every team’s schedule, for starters. Once you have that, you can calculate a whole world of good stuff like strength of schedule and any number of opponents’ stats. You’d also be in prime position to adjust each team’s trends based on strength of opponents, etc.
Look at those averages
The average amount of offense and defense score it the same
I don’t know why, but for some reason that amazes me.
21.7 average points scored, 21.7 average points given up. How often does that happen?
The Cowboys are the France of the NFL
-thwalls
On 3rd down, throw it to Smith.
by Willgfass on Nov 17, 2009 10:57 PM EST reply actions 2 recs
Wait I'm retarded
I can’t delete it, no!!!
Of course the average would be the same
Too much math, too much physics. Please ignore
The Cowboys are the France of the NFL
-thwalls
On 3rd down, throw it to Smith.
I
Can except some of these ratings.Atleast we arent ranked behind a team we totally dominated again like the eagles in philly.But i’m still not buying us ranked behind 6-3 dallas and 5-4 ravens.Even if i can except us being ranked behind 6-3 pats
Adjust your mindset
By now you certainly must know how the numbers are calculated, and that Ws and Ls have nothing to do with the ratings. Correct, my punctuatin’ young man?
Here’s a fun fact for you. You know the Saints’ 7 defensive TDs? If you take those 49 points away from their offense, and credit them toward their defense (where they belong), the Saints jump into first place in my ratings, and by a comfortable margin to boot. Perhaps that’s more exceptable for you personally!
That's a good idea
All in favor, say aye.
Arrr, matey!
The Saints be scurvy dogs!!! 6th place is too good for the likes of them wharf rats!!!
Yep, I remember 'hemorrhage', and 'embolism', and even 'detached retina'...
It sounded scary, but when you smell the goal line, they're only words. Just words.
by Bill Dauterive on Nov 20, 2009 1:37 AM EST up reply actions
Make the Saints walk the plank, me bucko!
Robert Newton (in your picture) practically defined the way pirates were portrayed in the movies and popular culture. Interesting, if off-topic, Wiki article here.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
