Game Predictions Week 10

Game probabilities for week 10 NFL games are listed below. The probabilities are based on an efficiency win model explained here and here. The model considers offensive and defensive efficiency stats including running, passing, sacks, turnover rates, and penalty rates. Team stats are adjusted for previous opponent strength. Games in which the model disagrees with consensus favorites are highlighted in red.

Vprob	Visitor	Home	Hprob
0.38	ATL	CAR	0.62
0.61	BUF	MIA	0.39
0.14	CLE	PIT	0.86
0.47	DEN	KC	0.53
0.39	JAX	TEN	0.61
0.34	MIN	GB	0.66
0.42	PHI	WAS	0.58
0.19	STL	NO	0.81
0.46	CIN	BAL	0.54
0.33	CHI	OAK	0.67
0.71	DAL	NYG	0.29
0.40	DET	ARI	0.60
0.91	IND	SD	0.09
0.06	SF	SEA	0.94

Luckiest NFL Teams Week 9

Based on opponent-adjusted generic win probability (GWP), the number of expected wins can be estimated for each team. Teams that have won more games than expected can be considered lucky, while teams with fewer wins than expected can be considered unlucky.

The list of NFL teams sorted from luckiest (positive numbers) to unluckiest is posted below. We would expect most teams to be within +/- 1.0 wins. So teams outside that margin can be deemed significantly lucky or unlucky. (The list accounts for the fact that four teams have not yet had their bye.)

Team	GWP	Act Wins	Exp Wins	Luck
DET	0.41	5	3.3	1.7
GB	0.54	6	4.3	1.7
CHI	0.18	3	1.5	1.5
CLE	0.45	5	3.6	1.4
TEN	0.60	6	4.8	1.2
KC	0.35	4	2.8	1.2
NYG	0.62	6	5.0	1.0
HOU	0.46	5	4.1	0.9
NE	0.91	9	8.2	0.8
SF	0.16	2	1.3	0.7
BAL	0.43	4	3.5	0.5
CAR	0.43	4	3.5	0.5
WAS	0.56	5	4.5	0.5
JAX	0.57	5	4.6	0.4
SD	0.45	4	3.6	0.4
NO	0.45	4	3.6	0.4
DAL	0.85	7	6.8	0.2
BUF	0.48	4	3.9	0.1
OAK	0.24	2	1.9	0.1
DEN	0.40	3	3.2	-0.2
PIT	0.79	6	6.3	-0.3
IND	0.92	7	7.3	-0.3
ARI	0.42	3	3.4	-0.4
MIN	0.46	3	3.7	-0.7
ATL	0.40	2	3.2	-1.2
SEA	0.69	4	5.5	-1.5
PHI	0.57	3	4.6	-1.6
STL	0.21	0	1.7	-1.7
NYJ	0.30	1	2.7	-1.7
CIN	0.48	2	3.8	-1.8
TB	0.80	5	7.2	-2.2
MIA	0.30	0	2.4	-2.4

Week 9 Efficiency Rankings

NFL team efficiency rankings are listed below in terms of generic winning probability. The GWP is the probability a team would beat the league average team at a neutral site. Each team's opponent's average GWP is also listed, which can be considered to-date strength of schedule. GWP modifies the generic win probability to reflect the strength of past opponents. Offensive ranking (O Rank) is based on each team's offensive GWP, i.e. it's the team's GWP assuming it had a league-average defense. D Rank is vice-versa. Rankings are based on a logistic regression model applied to data through week 9. A full explanation of the methodology can be found here.

Rank	Team	Last Wk	GWP	Opp GWP	O Rank	D Rank
1	IND	2	0.92	0.58	2	2
2	NE	1	0.91	0.53	1	3
3	DAL	3	0.85	0.47	3	5
4	TB	4	0.80	0.52	4	8
5	PIT	5	0.79	0.44	9	4
6	SEA	6	0.69	0.47	12	12
7	NYG	11	0.62	0.46	16	7
8	TEN	9	0.60	0.53	27	1
9	JAX	7	0.57	0.55	8	23
10	PHI	10	0.57	0.49	5	21
11	WAS	12	0.56	0.51	24	6
12	GB	14	0.54	0.45	13	14
13	BUF	19	0.48	0.56	19	11
14	CIN	15	0.48	0.55	7	26
15	MIN	23	0.46	0.47	14	15
16	HOU	17	0.46	0.47	18	24
17	CLE	20	0.45	0.50	6	28
18	NO	25	0.45	0.57	11	30
19	SD	8	0.45	0.44	15	19
20	CAR	24	0.43	0.53	22	18
21	BAL	16	0.43	0.41	28	9
22	ARI	21	0.42	0.51	23	17
23	DET	22	0.41	0.43	25	10
24	DEN	13	0.40	0.55	10	29
25	ATL	26	0.40	0.47	20	20
26	KC	18	0.35	0.42	26	13
27	NYJ	29	0.30	0.54	21	32
28	MIA	28	0.30	0.55	17	31
29	OAK	27	0.24	0.43	31	16
30	STL	31	0.21	0.53	29	27
31	CHI	30	0.18	0.50	30	25
32	SF	32	0.16	0.50	32	22

Kitna's 10+ Wins

Before the 2007 NFL season Detroit quarterback Jon Kitna predicted the Lions would win more than 10 games. Surprising many fans this year, DET has managed to put together a 6-2 record and the Lion's prospect of winning 10 or more games is realistic.

Based on their offensive and defensive efficiency stats to-date, they are ranked 15th in the NFL with a 0.47 generic win probability (GWP). Although they are 6-2, they have faced the 3rd easiest schedule so far, with their opponents averaging a 0.41 GWP.

And although Martz's offensive scheme and the Lions' receiving corps receives most of the attention, their offense has been below average (0.40 GWP) while their defense has been the stronger squad (0.54 GWP). Although DET's running and passing efficiencies are roughly average across the board, the offense has suffered from an above average fumble rate and the defense has benefited from an above average interception rate.

TEAM	OPass	ORun	OIntRate	OFumRate	DPass	DRun	DIntRate	PenRate
DET	6.23	4.28	0.029	0.045	6.02	3.82	0.044	0.30
AVG	6.17	4.08	0.032	0.025	6.22	4.06	0.032	0.38

Detroit's upcoming opponents and the probabilities of winning each game are listed in the table below. Their prospective strength of schedule is exactly average, with a GWP of 0.50.

Vprob	Visitor	Home	Hprob
0.54	DET	ARI	0.46
0.54	NYG	DET	0.46
0.43	GB	DET	0.57
0.42	DET	MIN	0.58
0.78	DAL	DET	0.22
0.46	DET	SD	0.54
0.31	KC	DET	0.69
0.39	DET	GB	0.61

After crunching the numbers, the likelihood of finishing the season at each possible record is illustrated in the table below by the bar graph (the probability of winning exactly that many games). The cumulative probability is the curved line with its scale on the right (the probability of winning at least x games.) [Note: the GB game could be much easier assuming GB has clinched a playoff birth.]


It looks just as likely that DET would finish with 9 or fewer wins as 10 or more. It's very close. If they continue to play at their current level, there is even an outside chance they can win 12 games, but 7 wins is equally as likely as 12. The table below summarizes the probabilities of finishing with each possible number of wins and the cumulative probability of winning at least that many wins.

Wins	Prob	At least
14	0.00	0.00
13	0.01	0.01
12	0.06	0.07
11	0.16	0.23
10	0.26	0.50
9	0.27	0.76
8	0.17	0.93
7	0.06	0.99
6	0.01	1.00

Patriots and Colts

The Patriots came into Indianapolis and beat the Colts in one of the most hyped mid-season games in NFL history. NE came into the game ranked #1 in efficiency rankings, accounting for strength of schedule. IND came into the game ranked just behind at #2.

This week the Colts will be ranked #1 ahead of the Pats. IND's generic win probability (GWP) -- the probability they would win a game against a league average team at a neutral site-- is 0.92, while NE's is 0.91.

Except for the score, the Colts actually outplayed the Patriots. Given their in-game statistics, it was surprising that NE won the game. One model says IND would have had a 77% chance of winning the game. The 4-point victory by the Patriots came down to one or two critical, high-leverage plays.

Most fans would say, so what? Who cares who was more efficient. The only thing that matters is the score. "You are what your record says you are," says Bill Parcells. That's true, but in-game efficiency stats are very often more indicative of future performance.

Unfortunately for IND, their hopes of an undefeated season are now dashed while NE's are still intact. Updating a previous post, here is NE's probability of finishing the regular season 16-0 accounting for their performance in yesterday's win.

Here are NE's upcoming games and the associated probabilities of winning each one.

Vprob	Visitor	Home	Hprob
0.85	NE	BUF	0.15
0.08	PHI	NE	0.92
0.89	NE	BAL	0.11
0.16	PIT	NE	0.84
0.03	NYJ	NE	0.97
0.02	MIA	NE	0.98
0.82	NE	NYG	0.18

So assuming they don't rest starters or suffer a critical injury, NE's probability of finishing the season 16-0 is now:

0.85 * 0.92 * 0.89 * 0.84 * 0.97 * 0.98 * 0.82 = 0.45.

Advance Game Predictions

Here is a summary of predicted probabilities for the outcomes of NFL games in future weeks.

Week	Visitor	Home	Vprob	Hprob
12	GB	DET	0.68	0.32
12	NYJ	DAL	0.05	0.95
12	IND	ATL	0.88	0.12
12	BUF	JAX	0.30	0.70
12	DEN	CHI	0.70	0.30
12	HOU	CLE	0.44	0.56
12	MIN	NYG	0.29	0.71
12	NO	CAR	0.42	0.58
12	OAK	KC	0.26	0.74
12	SS	STL	0.82	0.18
12	TEN	CIN	0.51	0.49
12	WAS	TB	0.18	0.82
12	SF	ARI	0.17	0.83
12	BAL	SD	0.29	0.71
12	PHI	NE	0.09	0.91
12	MIA	PIT	0.08	0.92
13	GB	DAL	0.16	0.84
13	ATL	STL	0.64	0.36
13	BUF	WAS	0.36	0.64
13	DET	MIN	0.31	0.69
13	HOU	TEN	0.31	0.69
13	JAX	IND	0.12	0.88
13	NYJ	MIA	0.40	0.60
13	SD	KC	0.61	0.39
13	SS	PHI	0.48	0.52
13	SF	CAR	0.19	0.81
13	TB	NO	0.81	0.19
13	CLE	ARI	0.43	0.57
13	DEN	OAK	0.73	0.27
13	NYG	CHI	0.80	0.20
13	CIN	PIT	0.15	0.85
13	NE	BAL	0.92	0.08
14	CHI	WAS	0.15	0.85
14	CAR	JAX	0.22	0.78
14	DAL	DET	0.89	0.11
14	MIA	BUF	0.27	0.73
14	NYG	PHI	0.44	0.56
14	OAK	GB	0.09	0.91
14	PIT	NE	0.20	0.80
14	SD	TEN	0.38	0.62
14	STL	CIN	0.19	0.81
14	TB	HOU	0.76	0.24
14	ARI	SS	0.20	0.80
14	MIN	SF	0.77	0.23
14	CLE	NYJ	0.55	0.45
14	KC	DEN	0.28	0.72
14	IND	BAL	0.90	0.10
14	NO	ATL	0.37	0.63
15	DEN	HOU	0.45	0.55
15	CIN	SF	0.76	0.24
15	ARI	NO	0.44	0.56
15	ATL	TB	0.13	0.87
15	BAL	MIA	0.50	0.50
15	BUF	CLE	0.46	0.54
15	GB	STL	0.81	0.19
15	JAX	PIT	0.23	0.77
15	NYJ	NE	0.03	0.97
15	SS	CAR	0.69	0.31
15	TEN	KC	0.65	0.35
15	IND	OAK	0.96	0.04
15	DET	SD	0.27	0.73
15	PHI	DAL	0.14	0.86
15	WAS	NYG	0.33	0.67
15	CHI	MIN	0.18	0.82
16	PIT	STL	0.90	0.10
16	DAL	CAR	0.88	0.12
16	CLE	CIN	0.37	0.63
16	GB	CHI	0.81	0.19
16	HOU	IND	0.07	0.93
16	KC	DET	0.38	0.62
16	MIA	NE	0.03	0.97
16	NYG	BUF	0.57	0.43
16	OAK	JAX	0.10	0.90
16	PHI	NO	0.63	0.37
16	WAS	MIN	0.47	0.53
16	ATL	ARI	0.44	0.56
16	BAL	SS	0.19	0.81
16	NYJ	TEN	0.19	0.81
16	TB	SF	0.93	0.07
16	DEN	SD	0.37	0.63
17	NE	NYG	0.82	0.18
17	BUF	PHI	0.30	0.70
17	CAR	TB	0.10	0.90
17	CIN	MIA	0.58	0.42
17	DAL	WAS	0.80	0.20
17	DET	GB	0.19	0.81
17	JAX	HOU	0.57	0.43
17	NO	CHI	0.60	0.40
17	PIT	BAL	0.80	0.20
17	SS	ATL	0.65	0.35
17	SF	CLE	0.16	0.84
17	TEN	IND	0.10	0.90
17	MIN	DEN	0.42	0.58
17	SD	OAK	0.76	0.24
17	STL	ARI	0.23	0.77
17	KC	NYJ	0.45	0.55

NFL Prediction Tracking

Here is a great site that tracks the season-long performance of the various NFL prediction models. It also summarizes their weekly predictions. (Thanks to Andy for the link). Model accuracy records from previous years are also available. I will add its link to the right so it can be referred to as the season progresses.

There are many competing statistical models, but it looks like the vast majority don't do very well. Fortunately, it appears the efficiency model used on this site ranks among the very best of the models. So far for 2007, the NFL Stats efficiency model would rank 3rd out of 61. (And if you nitpick, you would notice that the only two models with better records have either attempted predictions for fewer weeks or have possibly cherry-picked games to predict.)

The records for 2006 were strikingly lower. It appears to confirm it was a difficult year to predict with many upsets and with a historically weak average home field advantage. With its 62.56% accuracy rate, the NFL Stats efficiency model would have ranked #1 out of 58 (by a whisker).

Interestingly, there appears to be a lot of churn among the best prediction models. The best models one year aren't usually among the best in other years. This may indicate many of them are benefiting from a significant amount of luck. With so many different models, there are bound to be some that do well not because of predictive insight, but due to randomness. The models with true predictive ability would be the ones that rank near the top from year-to-year.

You can show your support for NFL Stats by sending an email recommending its inclusion in the rankings. [Edit: Never mind. He can't include me unless I predict scores as well.]

Season Win Projections Week 8

Season win totals and division standing projections are listed below. As before, projections are based on each team's opponent-adjusted generic win probability (GWP). The projections account for future opponent strength, and total wins account for current and projected wins. Methodology can be found here.

Team	Rank	Proj GWP	Fut Opp	Proj W
AFC E
NE	1	0.90	0.55	15.2
BUF	19	0.37	0.51	6.3
NYJ	29	0.19	0.57	2.5
MIA	28	0.27	0.54	2.2
AFC N
PIT	5	0.76	0.47	11.9
CLE	20	0.45	0.42	8.1
CIN	15	0.60	0.43	7.4
BAL	16	0.28	0.63	6.6
AFC S
IND	2	0.90	0.52	15.1
JAX	7	0.60	0.52	10.4
TEN	9	0.58	0.52	10.2
HOU	17	0.35	0.55	6.8
AFC W
SD	8	0.61	0.53	9.5
DEN	13	0.61	0.44	8.5
KC	18	0.39	0.52	7.5
OAK	27	0.28	0.52	4.6
NFC E
DAL	3	0.80	0.54	13.2
NYG	11	0.55	0.51	10.4
WAS	12	0.51	0.58	8.6
PHI	10	0.49	0.57	7.4
NFC N
GB	14	0.64	0.43	11.7
DET	22	0.35	0.51	8.2
MIN	23	0.45	0.47	6.0
CHI	30	0.21	0.50	4.7
NFC S
TB	4	0.86	0.40	10.9
CAR	24	0.32	0.55	6.9
NO	25	0.41	0.42	6.7
ATL	26	0.29	0.51	3.6
NFC W
SEA	6	0.80	0.38	11.2
ARI	21	0.47	0.41	7.2
SF	32	0.19	0.44	3.7
STL	31	0.22	0.44	2.0