Season Win Projections Week 6

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). But starting this week, the projections now account for future opponent strength. Total wins account for current and projected wins. Methodology can be found here.

Team	Proj GWP	Proj Wins
AFC E
NE	0.87	14.7
BUF	0.39	5.3
NYJ	0.32	4.2
MIA	0.31	3.1
AFC N
PIT	0.70	11.7
CLE	0.59	8.9
CIN	0.65	8.1
BAL	0.30	7.0
AFC S
IND	0.85	14.4
JAX	0.64	11.1
TEN	0.57	9.2
HOU	0.50	9.0
AFC W
DEN	0.68	9.5
SD	0.55	8.5
KC	0.51	8.1
OAK	0.28	5.1
NFC E
DAL	0.73	12.3
NYG	0.72	11.2
WAS	0.63	9.9
PHI	0.63	8.9
NFC N
GB	0.61	11.1
MIN	0.41	6.5
DET	0.31	6.4
CHI	0.24	4.4
NFC S
TB	0.84	12.4
CAR	0.37	7.7
ATL	0.46	5.6
NO	0.35	4.9
NFC W
SS	0.73	10.3
ARI	0.49	7.9
SF	0.26	4.9
STL	0.29	2.9

Future Opponent Strength

Generic Win Probability (GWP) tells us the probability a team will beat a league-average opponent at a neutral site, but teams don't play theoretical average teams. Some teams have harder schedules than others. Some teams have already played their weakest opponents and are probably overrated. Some teams have weathered an onslaught from the NFL's toughest teams and are likely underrated.

The table below lists each team and its GWP adjusted for previous opponents. Also listed is past opponent average GWP and future opponent average GWP. The table is sorted in order of future opponent strength. Teams with higher opponent average GWP will have an uphill climb for the rest of the season.

Team	GWP	Past Opp GWP	Fut Opp GWP
BAL	0.41	0.36	0.61
WAS	0.70	0.51	0.59
NYJ	0.39	0.57	0.58
MIA	0.39	0.53	0.58
JAX	0.70	0.51	0.57
CAR	0.43	0.43	0.57
DET	0.36	0.49	0.57
IND	0.88	0.53	0.56
DAL	0.78	0.44	0.56
HOU	0.56	0.56	0.56
BUF	0.45	0.69	0.56
PHI	0.68	0.51	0.55
NE	0.89	0.49	0.55
OAK	0.33	0.48	0.55
SD	0.60	0.51	0.55
TEN	0.61	0.59	0.54
CHI	0.25	0.56	0.52
MIN	0.43	0.40	0.52
PIT	0.72	0.41	0.51
KC	0.52	0.50	0.51
NYG	0.72	0.57	0.51
ATL	0.45	0.57	0.49
DEN	0.67	0.55	0.49
SF	0.25	0.53	0.49
STL	0.27	0.56	0.48
NO	0.32	0.67	0.46
GB	0.56	0.56	0.45
CIN	0.60	0.61	0.45
ARI	0.44	0.49	0.45
CLE	0.54	0.57	0.45
TB	0.81	0.49	0.44
SEA	0.64	0.49	0.39

Baltimore has the toughest forthcoming schedule. In fact, the Ravens have a stretch when they play SD, NE, IND consecutively. They get a breather against Miami before facing SEA and PIT again. And the AFC North is without its traditional doormat, the Browns, who have already beat Baltimore once.

Although the Ravens are 4-2, they have not performed terribly well against very poor opponents so far this year, including the three worst teams in the worst division in the NFL--the NFC West.

The Ravens' defense will likely keep them in games, but unless there is significant offensive improvement, I expect a significant downturn. There will probably be lots of much-hyped "collapse" stories in the media. Billick will be on the hot seat again and journalists will ask if he has "lost the team." But the reality is this team isn't that good to begin with, and following their bye in week 8 their schedule becomes as brutal as it had been soft.

Game Predictions Week 7

Game probabilities for week 7 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.

Vprob	Visitor	Home	Hprob
0.19	ARI	WAS	0.81
0.54	ATL	NO	0.46
0.38	BAL	BUF	0.62
0.15	MIN	DAL	0.85
0.92	NE	MIA	0.08
0.09	SF	NYG	0.91
0.86	TB	DET	0.14
0.51	TEN	HOU	0.49
0.64	KC	OAK	0.36
0.26	NYJ	CIN	0.74
0.11	CHI	PHI	0.89
0.12	STL	SS	0.88
0.48	PIT	DEN	0.52
0.73	IND	JAX	0.27

Week 6 Efficiency Rankings

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 strength of schedule. The adjusted GWP (Adj GWP) modifies the generic win probability to reflect the strength of to-date opponents. OGWP is each team's offensive GWP, i.e. it's the team's GWP assuming it had a league-average defense. DGWP is vice-versa. To-date average opponent GWP is also listed. Rankings are based on data through week 6. A full explanation of the methodology can be found here.

Team	Rank	Last Wk	GWP	OGWP	DGWP	Opp GWP
NE	1	1	0.89	0.79	0.72	0.49
IND	2	2	0.89	0.80	0.64	0.53
TB	3	8	0.82	0.66	0.61	0.49
DAL	4	4	0.76	0.74	0.59	0.44
NYG	5	3	0.72	0.75	0.68	0.56
PIT	6	10	0.72	0.62	0.63	0.41
JAX	7	6	0.71	0.76	0.50	0.51
WAS	8	5	0.70	0.49	0.83	0.51
DEN	9	11	0.68	0.66	0.54	0.55
PHI	10	9	0.67	0.74	0.60	0.52
SS	11	7	0.66	0.66	0.52	0.49
TEN	12	12	0.61	0.34	0.78	0.59
CIN	13	16	0.60	0.65	0.56	0.62
SD	14	18	0.60	0.67	0.51	0.52
GB	15	15	0.56	0.56	0.64	0.56
HOU	16	14	0.55	0.64	0.52	0.55
CLE	17	26	0.53	0.62	0.40	0.56
KC	18	19	0.52	0.41	0.61	0.50
BUF	19	20	0.46	0.41	0.57	0.69
ATL	20	22	0.44	0.51	0.51	0.57
MIN	21	21	0.43	0.51	0.49	0.40
CAR	22	24	0.41	0.53	0.54	0.42
ARI	23	13	0.41	0.55	0.55	0.49
NYJ	24	27	0.40	0.50	0.36	0.56
BAL	25	23	0.40	0.41	0.66	0.36
MIA	26	17	0.37	0.61	0.34	0.52
DET	27	29	0.36	0.32	0.48	0.49
NO	28	31	0.34	0.49	0.34	0.68
OAK	29	25	0.32	0.28	0.52	0.48
SF	30	28	0.26	0.25	0.50	0.52
STL	31	30	0.25	0.36	0.47	0.55
CHI	32	32	0.25	0.32	0.44	0.55

QB Ranking Week 5

The QB Wins Added Per 16 Games stat (+WP16) estimates how many wins a quarterback adds to his team's record over the course of a 16-game season. +WP16 is explained here and here. Last year's ratings can be found in the second link.Here is the list of 2007 QBs and their vital stats through week 5.

Name	Att	Yds	Int	Rush	Yds	Sk Yds	Fum	YAC	+WP16
Manning P	165	1319	2	7	-7	15	0	397	4.2
Brady	158	1383	2	5	6	19	2	534	3.7
Warner	69	580	1	4	0	6	2	199	3.3
Garrard	102	848	0	25	108	59	2	292	3.1
Garcia	113	914	0	11	29	27	1	446	2.2
Schaub	157	1299	4	12	32	63	5	455	2.1
Romo	171	1508	8	12	78	46	1	629	2.0
Palmer	160	1171	6	6	9	42	0	381	1.9
Culpepper	26	193	0	7	28	16	0	92	1.8
Campbell	113	869	3	16	80	27	4	361	1.4
Roethlisberger	131	1013	3	8	57	89	3	393	1.4
Hasselbeck	155	1148	4	10	1	45	1	459	1.2
Anderson	159	1251	8	10	17	62	3	426	1.0
Delhomme	86	626	1	6	26	46	1	313	0.7
Pennington	118	811	5	8	9	50	0	304	0.3
Smith	84	461	1	8	67	74	2	158	0.3
Cutler	152	1158	6	13	33	26	4	529	0.2
Boller	51	287	1	4	9	7	0	136	0.0
Leinart	112	647	4	11	42	23	0	273	0.0
McNabb	136	943	1	13	35	85	4	454	0.0
Favre	210	1527	4	9	-2	54	2	853	-0.1
Harrington	156	1070	3	5	13	91	0	541	-0.2
McNair	157	922	2	9	26	37	4	467	-0.4
Young	100	583	5	29	123	18	2	254	-0.4
Manning E	157	1076	6	9	13	32	2	537	-0.6
Bulger	133	765	4	4	16	67	1	317	-0.6
Rivers	157	1156	6	9	1	77	5	530	-0.8
Green	141	987	7	7	32	53	2	472	-0.9
Kitna	167	1333	6	12	31	136	8	603	-0.9
Huard	149	1029	6	6	2	75	2	497	-1.1
Holcomb	67	427	1	0	0	67	1	212	-1.2
Carr	62	330	2	9	33	35	0	183	-1.6
Edwards	79	507	3	3	5	53	1	272	-2.0
Brees	177	929	9	5	23	28	2	453	-2.0
Clemens	47	295	2	3	2	30	1	150	-2.2
Losman	47	255	1	6	51	67	3	135	-2.7
Griese	77	500	4	3	7	63	2	250	-2.7
Jackson	56	329	5	7	26	0	0	188	-3.1
Grossman	89	500	6	4	9	82	1	216	-3.3
McCown J	68	494	5	9	46	55	7	247	-3.9

The +WP16 stat basically tells us how much a QB is hurting or helping his team. For example, Tom Brady's performance to date would take a team that is completely average in every other way to between an 11 and 12 win season (8 + 3.7 = 11.7).

Luckiest NFL Teams Week 5

Based on opponent-adjusted generic win probability (GWP), the number of expected wins can be estimated for each NFL 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 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 lucky or unlucky.

Team	Adj GWP	Wins	Expected	Luck
DET	0.25	3	1.3	1.7
CHI	0.16	2	0.8	1.2
CAR	0.36	3	1.8	1.2
BAL	0.37	3	1.9	1.1
GB	0.60	4	3.0	1.0
DAL	0.82	5	4.1	0.9
SF	0.26	2	1.3	0.7
OAK	0.34	2	1.3	0.7
IND	0.90	5	4.5	0.5
PIT	0.70	4	3.5	0.5
NE	0.92	5	4.6	0.4
CLE	0.34	2	1.7	0.3
TEN	0.69	3	2.8	0.2
HOU	0.61	3	3.0	0.0
JAX	0.76	3	3.1	-0.1
ARI	0.62	3	3.1	-0.1
KC	0.43	2	2.2	-0.2
WAS	0.81	3	3.2	-0.2
SD	0.47	2	2.4	-0.4
NYJ	0.28	1	1.4	-0.4
MIN	0.40	1	1.6	-0.6
TB	0.73	3	3.6	-0.6
SEA	0.73	3	3.7	-0.7
NO	0.23	0	0.9	-0.9
ATL	0.40	1	2.0	-1.0
NYG	0.82	3	4.1	-1.1
BUF	0.43	1	2.1	-1.1
STL	0.24	0	1.2	-1.2
CIN	0.60	1	2.4	-1.4
DEN	0.70	2	3.5	-1.5
PHI	0.70	1	2.8	-1.8
MIA	0.47	0	2.4	-2.4

Behind the Prediction--WAS at GB

The Packers are a 3 point favorite at home against the Redskins, but the efficiency model has Washington as the favorites with a 0.62 probability of winning. In this post I'll breakdown why my system favors the Skins.

Although Washington is 3-2 and Green Bay is 4-1, the Redskins are performing significantly better per play. Below is a table that lists each team's generic win probability (GWP) and their opponents' average GWP, each adjusted for to-date oppenent stregnth. Also listed is each team's offensive and defensive GWP.

Team	GWP	Opp GWP	Off GWP	Def GWP	Opp Def	Opp Off
WAS	0.75	0.55	0.60	0.79	0.52	0.58
GB	0.59	0.55	0.63	0.56	0.55	0.53

Washington has been playing at a higher level in terms of opponent-adjusted efficiency. The two teams have faced roughly equal opponents through week 5, with Washington having faced stronger offenses and Green Bay facing stiffer defenses.

The next table breaks down each team's (unadjusted) efficiency stats.

TEAM	O Pass	O Run	O Int Rate	O Fum	D Pass	D Run	D Int Rate	Pen Rate
WAS	7.14	3.83	0.027	0.038	4.52	3.82	0.027	0.35
GB	6.73	3.35	0.019	0.028	6.12	3.99	0.029	0.43

The Redskins' offense is both running and throwing the ball slightly better than Green Bay. Their turnovers are higher, however. The biggest difference is on defense, where the Skins' pass defense is particularly strong, giving up only 4.5 yds per pass attempt.

A lot of attention has been given to Favre and his resurgent offense this year. But, by far, the bigger turnaround story in 2007 is the Redskins pass defense. In 2006 they had the worst pass defense by far, giving up over 6.9 yds per pass (including sack yds). So far in 2007, they have been the league's best.

Green Bay's advantage lies in turnover rates. Favre and the Packers have been measurably better at preventing interceptions than the Redskins. The Packers also have home field advantage.

Running the efficiency stats through the game prediction model, Washington's pass defense trumps Green Bay's Lambeau Feild, giving them a 0.62 probability to win.

The current consensus has Green Bay as a 3 point favorite, which essentially says the teams are equal because home field is generally worth 3 points. As we can see however, the teams are not equal and Washington's pass defense should carry the day given the Packers' tendency to rely on their passing game.

Game Predictions Week 6

Game probabilities for week 6 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. The model now compensates for early-season instability in team efficiency stats to improve confidence levels.

Vprob	VISITOR	HOME	Hprob
0.58	CIN	KC	0.42
0.35	HOU	JAX	0.65
0.52	MIA	CLE	0.48
0.67	MIN	CHI	0.33
0.72	PHI	NYJ	0.28
0.34	STL	BAL	0.66
0.40	TEN	TB	0.60
0.62	WAS	GB	0.38
0.29	CAR	ARI	0.71
0.59	NE	DAL	0.41
0.36	OAK	SD	0.64
0.18	NO	SEA	0.82
0.77	NYG	ATL	0.23