Mathematical System of Power Rankings

The “Efficiency Rating” stat we will discuss in this article is an attempt to measure the performance of each NFL team by measuring their "efficiency" for every offensive and defensive possession. It is our belief, through hours of study and research, that it is efficiency, not volume, that wins more NFL games.

First off, we went through numerous formulas that attempted to measure the efficiency of NFL teams. Most of them were a failure, some of them average, but none of them were a “true” success. We feel that this one is. A human being can never remain unbiased when creating a subjective list or ranking. This is why we feel the need to create a formula – to eliminate the subjectivity and human error. However, we do feel that it is a bad idea to rely solely on a formula, because there is nothing like watching an NFL game and making observations. The problem is, unless your job deals with the NFL (and even still), you likely are not watching even 2/3 of the NFL games throughout the season. So if you rely solely on your observations, you will miss out big time. There needs to be some kind of balance between stats, both advanced and traditional, and observations. So I advise, when we post these power rankings during the season, for you to take a strong look at them as well as other various stats, along with your own general belief to make a judgment about a certain team or game/bet.

To create these power rankings, we needed to complete a study where we would find which stats correlated best to wins in the NFL and the exact correlation between a certain stat and wins. This is why we haven’t posted an article in a few days (we were working on this). We went back to 2000, crunched tons of numbers, and came up with the stats we would end up using for these power rankings. Obviously, we’re using the stats that have the highest correlation to wins.

We ended up using 15 stats in our formula, not all weighted equally, but weighted as to how the stat correlates to wins. The 15 stats we used were all team stats, not individual, and all delta, or differential. For example, point differential. Because of this, we won’t come up with offensive and defensive totals. We will only come up with 1 total rating at the end, but it will take into account offense and defense because the stats are differential stats. A team can be positive or negative in any of the 15 stats. Here are the specific 15 stats we used:

1. Point Differential

2. ANY/A Differential

3. QB Rtg. Differential

4. Pass YPA Differential

5. Team Total Yards per Play Differential

6. Pass TD% Differential

7. Rushing TD Differential

8. Total 1^st Down Differential

9. Turnovers Differential

10. Passing TD Differential

11. Rushing Yard Differential

12. Sack Yard Differential

13. Fumble Rate Differential

14. Total Yards per Drive Differential

15. Defensive + Special Teams TD Differential

The first 14 stats had the highest correlation to wins of the 30 or so stats we crunched. Our original calculations only included those 14 stats, but we eventually added defensive + special teams TD differential to penalize bad special teams units like the Chargers and reward good ones like the Bears, as well as scoring defenses.

I’m not going to bore you with the formula or too many numbers. It gets complicated after this. All you really need to know is that those are the 14 stats that went into the calculations, and from there you can make an opinion on whether you like this stat or not.

Of course though, we wouldn’t consider such a rating a success if the Bills and Cardinals occupied the top 2 spots. As expected, the rankings came out very well. Using this rating would have predicted the Super Bowl 2 years in a row at the start of the playoffs. That’s more than most people could say, as everyone was climbing on the New England bandwagon and not many people had the 6 seeded Packers reaching the big game. So I’m going to let you see for yourself the Efficiency Rating of all 32 teams. The table is sorted by “Efficiency Rating”. To the right of that, we have “Total Rating of Opponents”. For that column, we just added the rating of each team’s 16 opponents so we could adjust for strength of schedule. To the right of this column, we just came up with the average rating by dividing that number by 16. From there, we were able to come up with our “Adjusted Efficiency Rating”. This is just the Efficiency Rating stat adjusted for strength of schedule. We have their rankings in Adjusted… to the right of that.

Team	Efficiency Rating	Total Rating of Opponents	Average Rating of Opponents	Adjusted Efficiency Rating	RK
Steelers	147.27	-65.42	-4.09	143.18	1
Patriots	135.50	45.67	2.85	138.35	2
Packers	123.14	-60.80	-3.80	119.34	3
Chiefs	115.47	-619.66	-38.73	76.74	8
Giants	110.93	-380.75	-23.80	87.13	5
Ravens	101.38	-136.70	-8.54	92.84	4
Chargers	93.77	-254.03	-15.88	77.89	7
Falcons	93.00	-172.68	-10.79	82.21	6
Colts	68.27	32.85	2.05	70.32	9
Saints	56.95	-373.17	-23.32	33.63	14
Bears	53.19	-188.04	-11.75	41.44	12
Eagles	47.75	29.08	1.82	49.57	10
Jets	45.39	-77.00	-4.81	40.58	13
Titans	35.77	108.89	6.81	42.58	11
Buccaneers	32.16	-382.33	-23.90	8.26	15
Raiders	-13.56	-28.03	-1.75	-15.31	17
Texans	-19.39	271.64	16.98	-2.41	16
Rams	-37.25	-596.54	-37.28	-74.53	24
Lions	-42.95	222.53	13.91	-29.04	19
Cowboys	-45.10	81.45	5.09	-40.01	21
49ers	-45.27	-249.17	-15.57	-60.84	23
Bengals	-52.20	473.35	29.58	-22.62	18
Dolphins	-52.63	282.10	17.63	-35.00	20
Vikings	-56.60	169.32	10.58	-46.02	22
Jaguars	-76.76	0.68	0.043	-76.72	26
Seahawks	-81.05	-139.20	-8.7	-89.75	27
Browns	-102.74	451.10	28.19	-74.55	25
Redskins	-117.49	306.74	19.17	-98.32	28
Broncos	-121.46	239.90	14.99	-106.47	29
Bills	-136.19	465.72	29.11	-107.08	30
Cardinals	-142.55	-327.90	-20.49	-163.04	32
Panthers	-154.69	315.93	19.75	-134.94	31

The rankings came out pretty well. Personally, I think you want to look at both the Efficiency Rating and Adjusted Efficiency Rating stats rather equally. You may think that because it is adjusted for strength of schedule it is automatically better, which it actually is, but this will not affect a team’s record. For example, we could see the Chiefs benefited from an easy schedule since they fell from 4^th to 8^th when the Rating was adjusted for schedule. We can use that in their playoff matchup against the Ravens to tell that Baltimore has the advantage in that game, but we can’t use this to predict their regular season record as well as the und-adjusted one. Get it?

I don’t know what to say about the 6-10 Titans. Obviously, we know they got unlucky and had a tough schedule, but their rating jumps out at you as being too high. As for other observations, the Giants seem a little high, though people forget they did have a great year and were on their way to winning the division before the infamous “Miracle at the Meadowlands II” against the Eagles. The Jets may also be a little low, but they’ve never been a team to dominate statistically, so this is where the eye test comes in to play.

Overall, we love this rating. Look for us to continue to use this throughout next season and keep your eyes peeled for an article where we will come out with “Strength of Victory”. Basically, we’ll add up the total rating of the opponents that each team defeated so we can see which teams defeated the best teams, rather than just which teams played the best teams. Like for the Panthers, they had a very tough schedule, but that’s still no excuse for winning only 2 games.

There’s really not many flaws with this stat. You can argue that we should have used one stat over another in this formula, but our studies basically told us to use the stats that we did. I think the fact that we used 15 stats really tells the whole story, as it takes into account nearly every facet in the game, and the (mathematically) most important stat in that facet of the game. We’ve got pretty much every base covered. I guess the only flaw is like we said, stats can lie. Usually with 15 stats it will normalize, which it did, but there will still be noticeable errors like the Titans over the Jets. But overall, this rating is definitely a keeper. If you’re into gambling, which I know some of us are, go ahead and check back here when the 2011 football season rolls around. I know you won’t be disappointed.