This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.

Crazy power rankings

Posted by Chase Stuart on October 10, 2007

On Monday Night, the Dallas Cowboys won a game despite never running a play with the lead. Believe it or not, it's not the first such game the Bills have lost this year -- in week one, the Broncos also took their first lead on a game winning field goal as time expired.

While I'm certainly no Bills apologist, an 0-2 record from those two games doesn't (in my humble opinion) accurately reflect how they played in those two games. Of course, if either field goal missed, Buffalo would have gone 2-0. But what about the fact that they ran zero plays from scrimmage when trailing the opponent?

It got me to thinking of one of my favorite stats that no one uses -- time of possession when winning, tied, or trailing. Unfortunately, I don't have TOP data easily accessible, but I do have data from the Data Dominator on the number of pass attempts and rush attempts (sack data not included, unfortunately) in each of those three situations.

This led me to the idea of a different type of power rankings. I ranked each team by the number of plays run when winning, losing, or tied in the game. I then calculate their percentage the same way the NFL calculates winning percentage: (Wins + Ties * 0.5) / Total Games. Take the Patriots, for example. They've run 246 plays while in the lead, and only 19 when trailing. Another 60 came when tied, with a lot of those at the starts of games. Their winning percentage is therefore 0.849 (276/325). Here's the list for all 32 teams.

	 Pct	 W	 L	 T
NE	0.849	246	 19	 60
PIT	0.841	221	 13	 71
IND	0.762	230	 61	 31
JAX	0.731	144	 35	 57
WAS	0.635	108	 42	 95
GB	0.629	144	 64	101
TB	0.629	134	 67	 59
CHI	0.515	119	110	 70
BAL	0.514	140	130	 78
DAL	0.503	112	110	 94
BUF	0.498	 86	 87	 78
OAK	0.492	 95	 99	 47
SEA	0.490	101	107	 83
SD	0.475	120	135	 47
ATL	0.462	 63	 85	139
DET	0.451	 92	120	 74
PHI	0.371	 51	113	 77
TEN	0.367	 51	118	 82
MIN	0.364	 46	110	 80
HOU	0.360	 64	145	 81
NYJ	0.356	 57	141	 94
CAR	0.345	 58	148	 85
NYG	0.344	 58	153	 94
STL	0.340	 40	137	127
DEN	0.326	 68	170	 55
CIN	0.285	 59	168	 27
CLE	0.281	 70	195	 20
ARI	0.274	 50	191	 71
MIA	0.232	 31	178	 65
KC	0.213	 37	195	 43
SF	0.197	 24	172	 48
NO	0.188	 28	197	 46

That list bears some resemblance to most people's power rankings, although it's important to remember that there is no strength of schedule adjustment here. According to Sagarin, Dallas, Pittsburgh and Carolina have faced the easiest schedules. Meanwhile, Buffalo, New Orleans and Cincinnati have faced the toughest. Buffalo ranks higher on this list than nearly every power rankings list you'll find, despite not even factoring in that they've played the toughest schedule in the league. A couple of kicks go the other way, and Buffalo would easily be 3-2, with forgivable losses at Pittsburgh and New England (the top two teams on this list).

Margin of victory isn't factored into this analysis explicitly, but it's not ignored, either. In a blowout, you're simply not going to run many plays from scrimmage with the lead. Conversely, if you're up by 21 early and win by 31, you'll get lots of credit in this ranking system for running nearly all your plays with the lead.

Do I really think that the Giants are worse than the Vikings? No. I don't think the Cowboys are worse than the Bears or Bucs, and I'd take the Bengals over the Rams when they play each other. Ideally, these numbers would be adjusted for strength of schedule, which would knock Dallas down even further, but "feel" more accurate.

Are these useful? I don't really know. But it helps to get an idea of how the games have gone this year, and that's something that I like to know. The Titans are 3-1 with a two point loss to the Colts, so you could argue they're easily a top five team.

On the other hand, the 2 point loss to the Colts underscores the fact that Tennessee never held the lead in the game. They were down 14-3 pretty early on, and the Colts never let the Titans take over the game. The blowout win over the Saints makes it easy to forget that New Orleans held a lead for about five minutes in the third quarter. In the season opener, Jacksonville was winning for almost two full quarters. Tennessee won by three, and Jacksonville was without their field goal kicker for the day (and as a result, eschewed a 36 yard field goal attempt in the second half). That game could have easily have gone either way. Against Atlanta, the Titans were tied or trailing until one minute to go in the third quarter, and Atlanta had 1st and goal from the 1 down by a TD with two minutes to go in the game.

Yes, Tennessee deserves a load of credit for winning three games. When ranking what they've accomplished to date, they might deserve to be in your top five. But the point is they could be ever so slightly less talented (or lucky) and be 1-3 instead of 3-1 right now. For predictive purposes, I think that's something you shouldn't overlook.