Ryan Early's Power Points

I've calculated my own NFL power ratings since I was in high school and read an article about it in Sport magazine. When I became a full-time football writer in 1997, I revamped my formula into its present form. That was when I went to work for NFLtalk.com, later relaunched as sportsTALK.com. It was published as "Ryan Early's Scientific Power Ratings." Three years later when ESPN bought out the site, they rebranded them as "Ryan Early's Power Points." Granted, that's probably an improvement over the scientific  name, but I was always a little worried about borrowing the name of a popular presentation program name. Alas, I figure the few people who remember my power ratings do so under that name and so I'm keeping it. At least, for now.

What I do is analyze each game's drive statistics and attribute each team's offensive, defensive, and special teams performances towards the game's point margin. How this (generally) works is that in each game a team is expected to, on average, score about 20 points from 12 offensive drives. That works out to approximately 1.6 points every time the offense takes the field. If the defense forces the offense to punt, then the defense outperformed the offense on that series and is credited with performing above expectations.

The cool part of my formula is that the expected points is adjusted based on the starting field position. So a defender intercepts a ball and returns it to the opponents' 1 yard line. The offense takes the field and their expected points on that drive is pretty high. Almost 7 points. So when they punch it into the end zone on the next play, they are credited with exceeding expectations, yes. But not by much. The bulk of the points for that touchdown went to the defense for setting up the offense with such great field position.

As you can (hopefully) see. This appropriately credits each unit of a team with their contribution to the final game result. For example, in week 4 of this season the Ravens beat the Jets 34 - 17. That's a pretty high scoring game. And many weeks from now, some analyst on ESPN is going to mention the Jets are averaging some number of points allowed per game as some kind of negative comment on how their defense is playing. The only problem with that is the fact that the Ravens scored 3 touchdowns when the Jets defense wasn't even on the field. Yes, the Jets offense gave up 2 fumble return touchdowns and 1 interception return touchdown. Those points should be credited against the Jets offense and not the defense, yes? I certainly think so. And more astonishingly, of the 17 points the Jets did score against the Ravens included a kickoff return touchdown and an interception return touchdown. In other words, the Jets offense was only responsible for scoring 3 points for their team that night while directly giving up 3 touchdowns.

Never in my many years of calculating these power ratings have I see a game result like this:
The Jets defense outperformed the Ravens offense by 15 points.
The Jets special teams outperformed the Ravens special teams by almost 5 points.
And yet the Jets offense were completely dominated by the Ravens defense, being outperformed by an enormous 37 points.

So how do I translate that into a team's Power Points? I'm not going to give away my secret formula here, but it involves running the season over and over again, making adjustments to each team's ratings until they accurately measure the season's performance to date. So as an example, last Sunday night the Packers beat the Falcons 25 - 14. That may have been a good win by the Packers, especially since they overcame some uneven play in the first half and an injury to their starting left tackle to dominate the second half. Yet, based on the two teams' play during the first four weeks, the Packers performed worse than the Power Points expected. It thought the Packers offense would outplay the Falcons defense by 22 points when they only outperformed them by 9. That 13 point difference means that either the Packers offense isn't as good as the Power Points thought, the Falcons defense is better than thought, or the two teams home field advantage is greater than thought.

(My power points attempt to measure how much better a team plays at home than on the road. Unlike all other power rating systems, I don't think the home field advantage is a league-wide constant but instead varies for each team).

My formula then adjusts the Packers offense rating down a little, the Falcons defense rating up a little, and both teams' home field advantages up a little. It then re-runs the season. If the Packers offense isn't as good as thought, then all of the Packers' opponents' defenses must be worse than thought. And adjustments are again made. These adjustments reverberate through all 32 teams, even the ones on a bye week. So the Cowboys had their Power Point rating increase this week while they sat at home, because several of their previous opponents showed themselves to be better than originally thought and thus the Cowboys performance against them was better than originally thought.

And the season is run again, with more adjustments made, and then run again, and again, and again, until there's no change in the ratings.

Now, early in the season, like where we are now, there can be some large swings from week to week in the ratings. That's because there isn't a lot of data available right now and thus a big blow-out win makes a big impact. I've tried to adjust for this by putting a cap on how much a blowout can change the ratings on each running of the season, but those many iterations add up. And so the 49ers jumped all the way from 14th to 3rd this week. I seriously doubt they are the 3rd best team in the NFL this season (though you never know, maybe they are).

Week 5 Power Points

Rk	Team	HfAdv	Off.	Rk	Def.	Rk	Sp.Tm	Rk	Pow.Pt
1	Green Bay Packers	2.40	13.51	3	-0.61	14	1.47	8	14.37
2	New England Patriots	0.64	15.08	1	-3.61	21	1.35	9	12.81
3	San Francisco 49ers	2.98	3.96	12	4.34	3	3.72	1	12.03
4	Baltimore Ravens	4.48	0.42	16	11.96	1	-1.75	26	10.63
5	Detroit Lions	3.97	6.41	7	4.01	4	-0.24	15	10.18
6	New Orleans Saints	2.27	14.77	2	-6.28	25	1.17	11	9.66
7	Buffalo Bills	-0.27	5.51	8	2.02	9	2.00	6	9.53
8	New York Jets	4.09	0.55	15	5.42	2	2.90	3	8.87
9	Oakland Raiders	0.27	8.10	4	-3.94	23	3.17	2	7.32
10	Dallas Cowboys	-1.35	5.35	10	1.62	11	-0.29	16	6.68
11	Pittsburgh Steelers	6.95	6.64	5	0.51	12	-0.89	18	6.26
12	Cincinnati Bengals	2.04	-0.21	18	2.77	6	2.33	5	4.89
13	Houston Texans	0.99	0.98	14	3.56	5	-0.04	14	4.50
14	Tennessee Titans	3.65	6.48	6	-2.29	18	-1.09	20	3.09
15	Chicago Bears	2.07	-2.68	27	1.99	10	2.45	4	1.76
16	Carolina Panthers	2.34	5.43	9	-4.38	24	-2.66	32	-1.61
17	San Diego Chargers	-0.43	1.68	13	-1.63	16	-1.84	28	-1.79
18	Washington Redskins	1.23	-2.52	25	2.59	7	-2.15	29	-2.08
19	Denver Broncos	1.69	-0.03	17	-1.08	15	-1.11	21	-2.22
20	Philadelphia Eagles	-0.41	4.90	11	-9.08	30	1.83	7	-2.34
21	Atlanta Falcons	2.27	-1.90	24	-2.81	19	0.90	13	-3.82
22	Seattle Seahawks	0.26	-3.10	28	-0.35	13	-0.48	17	-3.93
23	Minnesota Vikings	3.93	-0.53	21	-2.24	17	-2.65	31	-5.42
24	New York Giants	-1.63	-0.52	20	-3.49	20	-1.65	25	-5.66
25	Jacksonville Jaguars	2.42	-11.09	32	2.42	8	1.24	10	-7.43
26	Arizona Cardinals	2.63	-0.77	22	-6.38	26	-1.29	24	-8.43
27	Miami Dolphins	1.52	-0.23	19	-7.02	28	-1.18	22	-8.44
28	Cleveland Browns	-0.51	-5.49	29	-3.71	22	-1.21	23	-10.41
29	Tampa Bay Buccaneers	2.51	-2.55	26	-9.60	31	1.00	12	-11.15
30	Kansas City Chiefs	0.63	-0.82	23	-10.65	32	-2.27	30	-13.74
31	Indianapolis Colts	1.43	-6.39	30	-8.83	29	-0.99	19	-16.20
32	St. Louis Rams	0.34	-9.57	31	-6.59	27	-1.76	27	-17.93
	Average:	1.73	1.61		-1.61		0.00		0.00