The Vikings' and Lions' Paths to the Playoffs

I recently noted that according to NFL Forecast's computation of playoff odds for each team, MIN at 5-6 appears to have a far better shot at a wildcard spot than 6-5 Detroit. In fact, there are five more teams at 5-6, but none with as a good a shot as MIN. But since DET plays at MIN this Sunday I'll focus on those two teams.

First, MIN is the better team according to their efficiency stats. Although their passing game is below average, they don't throw as many interceptions as DET. Their running game is spectacular, on both sides of the ball. Usually, the running game is not as important as passing, except when it is exceptionally good as it is with MIN.








TeamO Pass
O Run
O Int Rate
O Fum Rate
D Pass
D Run
D Int Rate
Pen Rate
MIN5.65.60.0290.0316.43.00.0280.38
DET6.14.00.0350.0476.43.70.0340.50
NFL Avg6.14.00.0320.0266.14.00.0320.37


But the real difference between the two teams lies in some stats that usually go overlooked. One thing that really hurts DET is their fumble rate. They lead the NFL in fumbles per play. Almost 5% of all offensive snaps result in a fumble. Their penalty rate is also the worst in the league at half a yard per snap. Accordingly, DET's generic win probability (GWP-the chance they'll beat a league-average team at a neutral site) ranks 27th at 0.33. In contrast, MIN's GWP ranks 13th at 0.56.

This Sunday MIN hosts DET in a crucial game. The prediction model gives MIN a solid advantage with a 0.78 probability of winning at home. That would split the season series and even up their records at 6-6. Now let's look at their remaining schedule.

DET's remaining schedule is listed in the table below. Their opponent's efficiency rank and the probability DET will win the game are also listed. DET has a tough schedule ahead. Their opponents' average GWP is 0.64. Plus, only two of their five remaining games are at home. The GB odds are relatively uncertain because by week 17, the Packers might be resting their starters if they've locked up their playoff seed.











OpponentOpp RankWin Prob.
at Min130.22
DAL30.11
at SD100.81
KC210.51
at GB60.13


Here is MIN's remaining schedule. They have a much easier schedule with an average opponent GWP of only 0.37, with three home games.









OpponentOpp RankWin Prob.
DET270.78
at SF320.80
CHI310.86
WAS120.57
at DEN140.57


Anything can happen, but it looks like MIN has a much better shot at a wildcard than DET despite being a game behind in the standings. That's one reason why this weekend's game is so important. If DET can upset MIN, they'd own the tiebreaker and a two game lead, possibly eliminating MIN from contention.

But let's not forget about the other teams in contention. Here are the other 5-6 teams scrapping for a wildcard berth, their efficiency ranking, and their average future opponent GWP.









TeamRankOpp GWP
PHI90.57
WAS120.52
ARI230.42
NO250.48
CHI310.55


I'm not saying MIN will definitely or even probably make the playoffs, but they do have the best shot of all the teams behind the Giants (7-4). Of the other teams, PHI appears strongest, but they also have the toughest schedule. WAS is about even with MIN in terms of team efficiency but it has a much tougher schedule. The loss of Sean Taylor may also affect the team in unpredictable ways. ARI has a relatively weak schedule, but has not been playing well themselves and their secondary is banged-up. NO has been playing well lately but their defense is practically non-existent. They're probably the hardest team to predict this year because of their slow start. CHI is simply bad, even on defense. It's surprising they have 5 wins at this point.

It's possible and perhaps likely we'll see an 8-8 team make the playoffs out of the NFC again this year. I think MIN has the best shot to be that team.

Winning on the Road

Teams with good running offenses do not win any more road games than other teams, all things being equal. That's right--being good at running the ball does not help teams win on the road. Contrary to what we've been told for years, having strong passing game is far more important to visiting teams than having a good running game.

Let me make a clarification up front. I'm not suggesting that having a great running performance in an individual road game doesn't help win that game. I'm saying that being a "running team" doesn't help win on the road. Teams that are "built" to win on the road are those that pass well and don't fumble.

I stumbled on this somewhat by accident when I was studying the effect of climate on home field advantage. My usual game model is a logistic regression that estimates the probability of winning based on team efficiency stats and home field. A simplified version looks like this:

Team A season efficiency stats
Team B season efficiency stats
Team A at home [1 if true, 0 if false]

This method resulted in balanced weights for the coefficients for Team A and Team B, and the importance of home field was captured in the coefficient for 'Team A at home.' But during my research into climate effects, I decided to try an alternate model without the home field variable. It looked like this:

Visiting team season efficiency stats
Home team season efficiency stats

With this method we would expect unbalanced coefficients. Because home teams win more often, the coefficients for the home team were generally stronger than for the visiting team. This means that if two theoretically equal teams played, the home team would have a higher probability of winning, which is exactly what we observe.

By examining the imbalance of each stat we can see what kind of teams tend to win on the road. Here are the regression coefficients for each efficiency stat for both home and visiting teams. (Logisitic regression is more difficult to interpret than linear regression. The coefficients indicate the change in the log of the odds ratio of the outcome. But we are comparing the relative strength of each coefficient, so don't worry about the "log odds ratio" for now. Just pay attention to the relative size of the coefficient between home and away team stats.)












Team StatHome CoeffVisitor Coeff
O Pass0.40-0.49
O Run0.48-0.04*
D Pass-0.600.48
D Run-0.270.16
O Int Rate-16.4015.61
D Int Rate19.30-17.63
O Fum Rate-11.5730.77
Pen Rate-1.651.30


The model has very solid goodness-of-fit stats, and it is 69.9% accurate (retrodictively) in predicting game winners. The regression is based on all regular season game outcomes in the past five years (n=1280).

First, compare the coefficients for offensive passing efficiency. The coefficient for home teams is 0.40, and for visiting teams is 0.49. We can interpret these numbers by saying "having a good passing game is slightly more important in winning for the visiting team than for the home team." But the difference is slight, and may not be significant.

Next, compare the coefficients for offensive running efficiency. The coefficient for home teams is 0.48, but for visiting teams it is only 0.04--practically zero! (The difference of 0.44 is strongly significant.) The near-zero coefficient for visiting teams' offensive running efficiency is what tells us that running well simply doesn't matter on the road.

Also, for some reason, teams that tend to fumble more often than others are at a greater disadvantage as visiting teams. And conversely, teams that don't fumble tend to have a greater advantage as road teams.

There are imbalances between nearly all of the team efficiency stats, but none as stark as that for offensive running. This was such an unexpected result and I had no prior theoretical basis for the observation, so I confirmed the results with a simpler analysis using correlation coefficients.

For each team over the past 5 seasons (n=160), I added up their road and home wins. The correlations of each efficiency stat with home and road wins were calculated. If different stats affect a team's ability to win at home and on the road differently, as we saw in the regression, then we should see different correlation coefficients. For example, the correlation between offensive running efficiency and home wins should be much stronger than running and road wins.















O PassO RunO Int RateO Fum RateD PassD RunD Int RatePen Rate
Away Wins0.550.10-0.40-0.42-0.340.000.31-0.16
Home Wins0.480.22-0.36-0.36-0.45-0.070.36-0.17


The correlations confirm the results from the regression. Being a good running team is more important to winning when at home, and not nearly as important to the visiting team. (Also note that stopping the run is just not important whether on the road or at home, as we've seen in previous research.)

I also did an even simpler analysis by comparing the average season running efficiency stats of all road winners and for all home winners. Again, the data was from every regular season game over the past five years (n=1280). The average offensive running efficiency for road winners was 4.13 yds/rush and for road losers was 4.10 yds/rush--a difference of only 0.03 yds/rush.

We see the opposite with passing efficiency. Road winners average 6.18 yds/att and road losers average 5.90 yds/att--a difference of 0.28 yds/att which is about 9 times larger than the difference we found for running. Again we see indications that passing efficiency tends to be the more important stat for the road team.

The obvious question is "why?" Why doesn't being a capable running team help win road games? My only theory is that passing well helps come from behind far more than running well. If road teams tend to find themselves behind more often than home teams, then unless they can pass, they wouldn't be able to score quickly and come back from a deficit. But home teams need to come back from deficits too, so the reason why running teams don't win on the road remains puzzling.

Playoff Race Predictions Wk 12

Beginning this week, NFL Stats and NFL Forecast are teaming up to predict the playoff races for the 2007 NFL season. NFL Forecast offers a free a web/java-based program that calculates playoff probabilities for each team and for each possible playoff seeding based on individual game probabilities. Check out the java program here. For the rest of the season, NFL Forecast will be using the game outcome probabilities published here at NFL Stats as its official baseline. The resulting playoff chances will be published weekly at the NFL Forecast blog. (By the way, some of the java code is being updated, so be patient if some of the functions are temporarily disabled. But the blog site has all the final output based on the NFL Stats game probabilities.)

NFL Forecast's playoff prediction engine goes beyond wins and losses, and accounts for the complex NFL tiebreaker and seeding procedures that baffle many fans.

My personal observations of the current playoff predictions focus on the second NFL wildcard spot. (Scroll down past the division ranking probabilities to the AFC/NFC Playoff Seeding Probability tables). It appears totally up for grabs. DET is currently holding the spot at 6-5, but there are 7 teams right behind at 5-6. I was very surprised to see 5-6 MIN with a 46% chance to grab a playoff spot, while DET has only a 4% chance. More on the reasons for this later.

Season Win Projections Week 12

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 (Fut Opp), and projected wins (proj W) is a total of current and estimated future wins. The methodology is described more fully here.













































TeamRankProj GWPFut OppProj W
AFC E
NE10.910.4815.6
BUF190.390.536.9
NYJ240.340.543.7
MIA180.430.492.1
AFC N
PIT50.700.5111.5
CLE170.610.3510.1
CIN200.500.406.5
BAL260.180.704.9
AFC S
IND20.880.4513.4
JAX70.650.5211.3
TEN150.450.548.2
HOU160.310.676.6
AFC W
SD100.690.409.5
DEN140.580.467.9
KC210.430.476.2
OAK290.140.673.7
TeamRankProj GWPFut OppProj W
NFC E
DAL30.850.5014.2
NYG110.530.549.7
WAS120.550.527.7
PHI90.540.577.7
NFC N
GB60.790.3813.9
MIN130.680.378.4
DET270.230.627.2
CHI310.190.556.0
NFC S
TB40.870.3411.3
NO250.390.486.9
ATL220.410.485.1
CAR280.210.625.0
NFC W
SEA80.710.4110.5
ARI230.470.427.4
SF320.180.503.9
STL300.230.523.2

Week 12 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 12. A full explanation of the methodology can be found here.





































RankTEAMLast WkGWPOpp GWPO RankD Rank
1NE10.910.5413
2IND30.860.5461
3DAL20.850.5025
4TB40.770.4847
5PIT50.710.42152
6GB70.690.48511
7JAX80.670.53319
8SEA60.630.381112
9PHI120.600.55720
10SD100.600.54139
11NYG90.570.53178
12WAS110.570.59206
13MIN150.560.511213
14DEN130.540.51827
15TEN140.490.52274
16HOU160.480.471624
17CLE170.460.49929
18MIA230.420.561921
19BUF190.410.591818
20CIN200.400.501031
21KC210.400.502610
22ATL220.390.502217
23ARI180.390.422115
24NYJ270.370.562330
25NO280.370.481432
26BAL240.350.422814
27DET260.330.492416
28CAR250.300.502523
29OAK310.250.452925
30STL290.250.463122
31CHI300.230.523028
32SF320.180.453226

Game Predictions Week 13

Game probabilities for week 13 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.




















VprobVisitorHomeHprob
0.23GBDAL0.77
0.59ATLSTL0.41
0.27BUFWAS*0.73
0.22DETMIN0.78
0.40HOUTEN0.60
0.19JAXIND0.81
0.37NYJMIA0.63
0.61SDKC0.39
0.44SEAPHI0.56
0.27SFCAR0.73
0.81TBNO0.19
0.49CLEARI0.51
0.72DENOAK0.28
0.77NYGCHI0.23
0.16CINPIT0.84
0.93NEBAL0.07

NFL Home Field Advantage by Climate 2

In a previous post, I looked at how home field advantage is affected by weather. Each NFL city was categorized by its December climate--dome, warm, moderate, and cold. By comparing the winning percentage of home teams in games in the early season (weeks 1-12) and the late season (13-17) the effect of cold weather was estimated.

In this post, the effect of cold weather on home field advantage will be measured more precisely. By using logistic regression, relative team strength is accounted for. In addition, the statistical significance of the observed weather effect indicates if the effect is real and systematic, or just a result of luck and small sample size.

The last post left off with this table. The left most column describes the visiting team's climate, and the top row describes the home team's climate. For each combination of visiting and home climate, the change in home team winning percentage from early to late in the season is listed. Positive numbers indicate that cold weather may favor the home team. For example, when dome teams play at warm cities, the home team winning percentage was 20% higher late in the season than early in the season. But when dome teams play at moderate cities, the home team winning percentage appeared to be 9% lower late in the season.


Difference
Dome
Warm
Mod
Cold
Dome
0%
+20
-9
+35
Warm
-5
+14
-7
-3
Mod
-8
-24
-6
-13
Cold
-8
+2
-15
+8


Several match-ups indicate that the cold and wind of late-season outdoor football has an effect on HFA. Consistent with other research, it appears that dome teams suffer when playing in cold climates.

Another remarkable combination is moderate weather teams playing in warm cities (-24%). But it's not clear why moderate teams would have an easier time playing in the balmy breezes of Florida or San Diego in December. Other notable match-ups are dome teams playing at warm cities (+20%), and cold teams playing at moderate cities (-15%).

To determine if the observed differences in HFA between early and late season games is really due to the change of weather, and not due to relative team strength or luck, several logistic regression models were run. The models were based on every regular season game from the 2002 through 2006 seasons (n=1280). For each game, each team was designated either Team A or Team B. The general model specification was the following:

Dependent variable:
Team A won

Independent variables:
Team A season efficiency stats
Team B season efficiency stats
AHome
[Weather dummy variable]

The team efficiency stats include offensive and defensive passing and running efficiency, turnover rates, and penalty rates. AHome is a dummy variable that is 0 when Team A is away and 1 when Team A is home. The [Weather dummy variable] is 1 when the particular climate match-up of interest is present for the game, and 0 when otherwise.

A general HFA variable (AHome) was included to isolate the effect of weather from the general home field advantage due to travel, psychology, officiating, or other effects.

Several models were run for each climate match-up of interest. The table below lists the statistical significance of the 'weather variable' and the resulting home field advantage calculated from the regression results. Note that the overall HFA rate is 57.5%.



Visitor
Home
p-value
HFA (%)

Dome
Cold
0.03*
88
Dome
Warm
0.28
75
Warm
Warm
0.58
71
Warm
Cold
0.74
76
Mod
Warm
0.11
40
Mod
Cold
0.95
62
Cold
Mod
0.13
49


Accounting for relative team strength, the only truly significant result is for dome teams in cold weather (p=0.03), with an expected home team winning percentage of 88%. This result is confirmed by the actual 86% home team winning percentage when dome teams play at cold cities late in the year. The regression's estimate of 88% suggests that the cold city home teams may have been slightly weaker compared to their dome opponents over the past 5 years.

Two other types of match-ups might be considered marginally significant--moderate teams at warm cities (p=0.11), and cold teams at moderate cities (p=0.13). Both results indicate a reduced HFA later in the season for those match-up types. But since there were 16 combinations of climate match-ups, chances were that we could see one or two type-II errors, i.e. see significance when none is truly there. Because there is no a priori theoretical reason why to expect those results, we shouldn't deem them significant.

Finally, one last model specification was run to make sure no other climate match-up types were significant. The final model included all late-season match-ups types together. No additional types were significant.

It's clear that situation of a dome team playing playing in a cold city late in the season creates a much stronger HFA than normal. But a larger data set is needed to conclusively analyze other weather match-ups. Dividing 1280 games into 16 types of match-ups and 2 weather periods creates small sub-samples.

A prediction model's accuracy may benefit from enhancing the weight of HFA in certain situations. Looking ahead, there are only three 'dome at cold' match-ups in 2007. STL travels to CIN in week 14, and NO visits CHI and DET visits GB in week 17.

Season Win Projections Week 11

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 is described here.













































TeamRankProj GWPFut OppProj W
AFC E
NE10.900.5515.4
BUF190.440.517.7
NYJ270.280.573.7
MIA230.340.542.0
AFC N
PIT50.740.4711.4
CLE170.600.429.6
CIN200.510.436.1
BAL240.220.635.3
AFC S
IND30.870.5213.2
JAX80.590.5210.5
TEN140.500.529.0
HOU160.380.557.3
AFC W
SD100.650.538.9
DEN130.620.448.7
KC210.500.527.0
OAK310.150.522.9
TeamRankProj GWPFut OppProj W
NFC E
DAL20.870.5414.2
NYG90.600.5110.6
WAS110.510.588.0
PHI120.430.577.6
NFC N
GB70.730.4313.4
DET260.260.517.6
MIN150.590.477.5
CHI300.240.505.4
NFC S
TB40.870.4011.2
NO280.350.426.1
CAR250.300.555.8
ATL220.320.514.9
NFC W
SEA60.710.3810.3
ARI180.560.418.4
STL290.220.443.3
SF320.210.443.3

Luckiest NFL Teams After Week 11

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.





































TeamGWPActual WinsExpectedLuck
GB0.6496.42.6
DET0.3563.52.5
CHI0.2442.41.6
CLE0.4864.81.2
TEN0.5065.01.0
NYG0.6176.10.9
NE0.91109.10.9
JAX0.6176.10.9
NO0.3343.30.7
DAL0.8598.50.5
BUF0.4554.50.5
ARI0.4554.50.5
CAR0.3743.70.3
BAL0.3843.80.2
HOU0.4854.80.2
SF0.2022.00.0
OAK0.2022.00.0
KC0.4144.1-0.1
DEN0.5255.2-0.2
PIT0.7477.4-0.4
SS0.6566.5-0.5
IND0.8588.5-0.5
PHI0.5655.6-0.6
WAS0.5655.6-0.6
SD0.5755.7-0.7
STL0.2722.7-0.7
ATL0.3833.8-0.8
MIN0.4944.9-0.9
CIN0.4334.3-1.3
NYJ0.3523.5-1.5
TB0.8068.0-2.0
MIA0.3803.8-3.8


Poor Miami. They've played well enough to win about 4 games, but are still waiting for their first victory.

NFL Home Field Advantage by Climate

In this post I'll begin an analysis of home field advantage in the NFL and its relationship to climate. Others have examined the connection between weather and HFA previously, but here I'll attempt to present the data with a clear and novel approach. This post represents the 'clear' part. In following posts, the novel part will use logistic regression to account for team strength and determine the significance of each particular type of climate match-up.

I began by dividing each home city into four categories: dome, cold, moderate, and warm based on a combination of each city's average December high temperature and wind speed. The dome cities include STL, NO, MIN, DET, IND, and ATL. The cold cities are GB, BUF, CLE, CHI, KC, DEN, NYG, NYJ, NE, PHI, PIT, and CIN. Moderate cities include BAL, WAS, CAR, SEA, OAK, SF, TEN, and DAL. The warm cities are MIA, TB, JAX, HOU, SD, and ARI.

Based on all NFL regular season games from the 2002 through 2006 seasons, the winning percentage of the home team was calculated for each type of climate match-up. I divided the season into "early" and "late." The early season is defined as weeks 1 through 12 and the late season is defined as weeks 13 through 17. For example, the winning percentage of the home team in match-ups of cold teams at moderate cities in the early season is 57%, with n=77 examples of such cases.

The home winning percentage of all types of weather match-ups are presented below in a series of pairs of tables. Each table is presented the same way, with the visiting team climate on the left and the home climate on the top. The first table in each pair is for the early season (pre-December), and the second table is for the late season (December games).

Sample size is usually an issue when populations are divided up among several classes. For that reason, the first pair of tables lists the number of cases of each type. For example, the top right cell of the first table lists the number of games featuring dome teams playing at cold cities in the early season. The same cell in the second table lists the same type of match-up in the late season.








Wk 1-12DomeWarmModCold
Dome31294462
Warm34205062
Mod40495178
Cold596677128










Wk 13-17DomeWarmModCold
Dome15172121
Warm16111730
Mod25201740
Cold20284359


The second pair of tables simply lists the straight-up winning percentage of the home team in each type of match-up. For example, the bottom left cell of the first table lists the home team winning percentage when cold teams play at domes in the early season. The same cell in the second table lists home winning percentage of cold teams at domes in the late season.









Wk 1-12DomeWarmModCold
Dome61%456151
Warm68506666
Mod60595976
Cold53525750










Wk 13-17DomeWarmModCold
Dome61%655286
Warm63645963
Mod52355363
Cold45544258


What immediately stands out is the very high winning percentage of cold teams hosting dome teams late in the season. The most remarkable result, however, may be the 35% home winning percentage of warm teams hosting moderate teams. Note that there are only about 20 cases of each type of match-up in the past 5 years, so these results could be due to luck or due to general team strengths of the according type of teams. Perhaps a couple moderate teams have been relatively dominant over warm weather division rivals between '02 and '06.

The final table lists the difference in home winning percentage between late season and early season match-ups. Simply put, it is late season winning percentage minus early season winning percentage. A high positive number indicates cold weather may give an advantage to the home team. A negative number or near-zero number suggests otherwise. We'd expect to see a zero for dome teams at dome teams, because outdoor weather is obviously not a factor. This method begins to account for relative team strengths over the period studied.









DifferenceDomeWarmModCold
Dome0%20-935
Warm-514-7-3
Mod-8-24-6-13
Cold-82-158


Take dome teams for example. By reading across, we see that dome teams seem to have no greater HFA in late season than the early season against other dome teams--as we'd expect. We also see that they are at a 20% disadvantage playing at warm cities but, for some reason, have a 9% better advantage playing at moderate cities late rather than early. Lastly, we see that dome teams appear to be at a severe disadvantage playing in cold cities late in the season, apparently giving up 35% advantage to the cold.

As mentioned above, some of the observed differences in HFA due to weather may be due to luck and relative team strengths among the weather-classes. The final table is a simple way of accounting for team strength, but it does not address the possibility that the differences are primarily due to luck. The final part of this article will use logistic regression to more powerfully account for team strength and test for statistical significance.

Game Predictions Week 12

Game probabilities for week 11 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.





















VprobVisitorHomeHprob
0.70GBDET0.30
0.06NYJDAL0.94
0.86INDATL0.14
0.27BUFJAX0.73
0.71DENCHI0.29
0.42HOUCLE0.58
0.31MINNYG0.69
0.37NOCAR0.63
0.21OAKKC0.79
0.78SEASTL0.22
0.48TENCIN0.52
0.19WASTB0.81
0.18SFARI0.82
0.25BALSD0.75
0.09PHINE0.91
0.14MIAPIT0.86


I wouldn't put too much stock in the NO-CAR probability. The QB/WR injury situation in CAR needs to be watched, and NO has been so inconsistent.

Week 11 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 11. A full explanation of the methodology can be found here.





































RANKTEAMLast WkGWPOpp GWPO RankD Rank
1NE10.910.4912
2DAL30.850.4925
3IND20.850.5441
4TB50.800.4836
5PIT40.740.43123
6SEA60.650.41916
7GB70.640.47814
8JAX90.610.53522
9NYG80.610.51148
10SD130.570.541610
11WAS120.560.57197
12PHI100.560.49720
13DEN150.520.53629
14TEN110.500.52284
15MIN140.490.501321
16HOU180.480.451725
17CLE200.480.541027
18ARI210.450.492211
19BUF170.450.581815
20CIN160.430.531130
21KC260.410.50279
22ATL190.380.472318
23MIA270.380.532023
24BAL220.380.412913
25CAR240.370.522419
26DET250.350.462512
27NYJ280.350.532131
28NO230.330.521532
29STL290.270.482624
30CHI300.240.533026
31OAK310.200.433128
32SF320.200.513217

Updated NE Undefeated Watch

Pittsburgh and the Giants are obviously the biggest obstacles remaining for the Patriots. Here is the updated probability NE will finish the regular season undefeated.











VprobVisitorHomeHprob
0.88NEBUF0.12
0.09PHINE0.91
0.90NEBAL0.10
0.21PITNE0.79
0.03NYJNE0.97
0.03MIANE0.97
0.81NENYG0.19


0.88*0.91*.90*0.79*0.97*0.97*0.19 = 0.45

Season Win Projections Week 10

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 is described here.













































TeamRankProj GWPFut OppProj W
AFC E
NE1

0.910.5015.4
BUF17

0.400.577.8
NYJ280.220.602.6
MIA270.260.561.9
AFC N
PIT4

0.810.4612.7
CLE20

0.550.388.8
CIN160.560.416.9
BAL220.260.645.8
AFC S
IND2

0.920.4213.4
JAX9

0.580.5310.0
TEN110.570.4910.0
HOU180.360.596.5
AFC W
SD130.610.429.3
DEN15

0.580.408.1
KC26

0.360.486.5
OAK31

0.160.563.1
TeamRankProj GWPFut OppProj W
NFC E
DAL3

0.880.4814.1
NYG8

0.610.5110.3
WAS120.420.627.9
PHI100.490.607.5
NFC N
GB7

0.740.3713.2
DET

25



0.290.588.0

MIN14

0.610.377.3

CHI300.200.545.4
NFC S
TB5

0.850.3910.9
NO230.410.476.9
CAR240.300.586.1
ATL19

0.340.595.4
NFC W
SEA6

0.760.3810.3
ARI21

0.510.397.6
SF320.190.453.4
STL29

0.220.512.6

QB Rating Week 10

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 10.











































RankPlayerQBRatAttYdsAir YdsYACIntRushYdsSk YdsFum+WP16
1Brady131.8299268615591127417396243.30
2Garrard102.914811416804610301347322.57
3Roethlisberger110.22422020128173971813517752.01
4Manning P90.8315238615418451010-66332.00
5Romo103.32922555147310821115818941.97
6Anderson90.729222311393838920557531.91
7Palmer91.1328246415099551013169031.56
8Garcia95.525819679819863301117131.40
9Schaub85.021516581084574714389471.33
10Hasselbeck89.3320230112891012822478421.05
11Warner86.2178131485246269-24561.03
12Delhomme111.88662431131316264610.76
13Cutler84.8248188210418419221027660.61
14McNabb91.731522951152114342810715160.55
15Losman80.8109820449371313737330.48
16Boller74.613477245931337223710.37
17Favre96.2354275712931464817-28560.32
18Brees84.83712447133611111214384550.30
19Pennington87.519013178194987172310220.25
20Leinart61.9112647374273411422300.10
21Kitna92.92982294133595981946211110.10
22Campbell79.92611735906829727147819-0.04
23Harrington81.02641737856881512301450-0.15
24Bulger73.1242160297163186151524-0.19
25Manning E78.628318201007813111541915-0.53
26Holcomb73.183515288227100871-0.64
27Lemon69.416692348244141870862-0.69
28Green72.61419875154727732532-0.77
29Smith57.2193914595319413891216-0.79
30Young62.219211126364761055217795-0.93
31Huard74.72621766951815119-11663-1.14
32Rivers77.9255174391782610191210710-1.31
33McNair73.9205111356854541032858-1.37
34Edwards69.9121790385405547711-1.48
35Carr65.711454026028031554651-1.64
36Griese77.12171508730778109161016-1.67
37Culpepper73.312581739941841536826-1.86
38Grossman53.9103642376266659963-2.53
39McCown J59.5122760348412918105848-3.67

Advance Game Predictions Weeks 12-17

Here is a list of the win probabilities of all remaining games in the 2007 NFL schedule. These will be continue to be updated as the season progresses.

Game Predictions Week 11

Game probabilities for week 11 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.





















VprobVisitorHomeHprob
0.35ARICIN0.65
0.20CARGB0.80
0.45CLEBAL0.55
0.04KCIND0.96
0.18MIAPHI0.82
0.89NEBUF0.11
0.35NOHOU0.65
0.16OAKMIN0.84
0.86PITNYJ0.14
0.35SDJAX0.65
0.78TBATL0.22
0.11WASDAL0.89
0.67NYGDET0.33
0.51STLSF0.49
0.10CHISEA0.90
0.50TENDEN0.50


It seems as if the consensus favorites and the mathematical models have converged. Only one game difference between them the past two weeks. I would not be confident in the Baltimore prediction without knowing the injury status of their secondary. Without a healthy McAlister, Rolle, and Reed, I don't see how they could be the favorite.

Week 10 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 10. A full explanation of the methodology can be found here.





































RANKTEAMLast WkGWPOpp GWPO RankD Rank
1NE20.910.5313
2IND10.890.5641
3DAL30.860.5024
4PIT50.790.4375
5TB40.780.5038
6SEA60.660.421216
7GB120.630.48914
8NYG70.620.52166
9JAX90.600.54621
10PHI100.600.51522
11TEN80.560.52272
12WAS110.530.53227
13SD190.530.51189
14MIN150.480.511317
15DEN240.480.531025
16CIN140.480.52826
17BUF130.470.542011
18HOU160.450.471528
19ATL250.430.451919
20CLE170.430.531127
21ARI220.410.482415
22BAL210.390.422910
23NO180.380.521432
24CAR200.370.512318
25DET230.370.442512
26KC260.330.452813
27MIA280.320.531730
28NYJ270.300.532131
29STL300.230.492629
30CHI310.220.493024
31OAK290.200.413123
32SF320.170.513220

Weather and Home Field Advantage

Someone recently pointed out a study that indicated home field advantage (HFA) is not the same for every stadium. While that's certainly true, it's very hard to quantify. By definition, the same team is always the home team when measuring a particular location's HFA, so in any given year there would be a lot of team strength captured in a variable accounting for the field's HFA.

The efficiency model I've used includes a factor for HFA, but it is the same regardless of climate. This is the beginning of an effort to quantify the effect of climate on HFA and to see how much of HFA is due to climate differences and how much is due to other factors such as crowd noise, referee psychology, or travel.

The table below lists each home team along with their average December weather. Click on the table headers to sort




































TeamAvg Dec TAvg Dec WindWind Chill
GB2910.519.7
BUF3613.327.3
CLE3712.129.0
CHI3711.029.5
NE4212.235.3
KC4211.235.7
PIT4210.436.0
NYG4410.838.3
NYJ4410.838.3
CIN4410.238.5
PHI4410.138.6
DEN448.439.3
SEA469.541.3
WAS467.842.0
BAL499.345.0
TEN498.945.2
CAR547.451.9
SF567.154.4
DAL5710.854.5
OAK587.156.8
HOU658.065.0
ARI655.165.8
JAX667.866.3
SD665.666.8
TB728.473.5
MIA759.277.1
ATL700.079.2
DET700.079.2
IND700.079.2
MIN700.079.2
NO700.079.2
STL700.079.2


It's not a surprise that Green Bay is coldest by far, followed by places such as Buffalo, Cleveland, and Chicago. Green Bay would even qualify as a cold climate through November with an average 36 deg wind chill. But I was surprised by how much colder (and windier) a place like Kansas City is than Baltimore or Washington. I'm still considering how to classify each city. Domes are easy, but where is the line drawn between cold and moderate? Should there be cold, moderate, and "warm" classes? For now I'd put the line between cold and moderate at 40 deg wind chill, between DEN and SEA. I'd also define warm weather teams starting at 60 deg between OAK and HOU.

Continue reading this article here.

Season Win Projections Week 9

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 is described here.















































TeamRankProj GWPFut OppProj W
AFC E
NE20.910.5015.4
BUF130.450.537.6
NYJ270.220.612.5
MIA280.250.562.0
AFC N
PIT50.810.4612.5
CLE170.510.449.1
BAL210.320.627.5
CIN140.560.426.4
AFC S
IND10.930.4414.5
TEN80.610.4910.9
JAX90.530.559.2
HOU160.360.607.5
AFC W
SD190.460.497.7
DEN240.520.397.1
KC260.380.467.0
OAK290.260.484.0
TeamRankProj GWPFut OppProj W
NFC E
DAL30.850.5013.8
NYG70.570.5510.6
WAS110.460.608.7
PHI100.460.616.7
NFC N
GB120.640.4011.1
DET230.380.538.1
MIN150.570.397.6
CHI310.190.504.5
NFC S
TB40.850.4111.0
NO180.510.438.1
CAR200.380.567.0
ATL250.320.594.5
NFC W
SEA60.800.3510.4
ARI220.520.407.1
SF320.170.493.3
STL300.210.491.7

Advance Game Predictions Weeks 11-17

As requested, here is a list of the win probabilities of all remaining games in the 2007 NFL schedule. These will be continue to be updated as the season progresses. I posted them off the front page of the site because it's such a long and ugly table.

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.


















VprobVisitorHomeHprob
0.38ATLCAR0.62
0.61BUFMIA0.39
0.14CLEPIT0.86
0.47DENKC0.53
0.39JAXTEN0.61
0.34MINGB0.66
0.42PHIWAS0.58
0.19STLNO0.81
0.46CINBAL0.54
0.33CHIOAK0.67
0.71DALNYG0.29
0.40DETARI0.60
0.91INDSD0.09
0.06SFSEA0.94

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.)





































TeamGWPAct WinsExp WinsLuck
DET0.4153.31.7
GB0.5464.31.7
CHI0.1831.51.5
CLE0.4553.61.4
TEN0.6064.81.2
KC0.3542.81.2
NYG0.6265.01.0
HOU0.4654.10.9
NE0.9198.20.8
SF0.1621.30.7
BAL0.4343.50.5
CAR0.4343.50.5
WAS0.5654.50.5
JAX0.5754.60.4
SD0.4543.60.4
NO0.4543.60.4
DAL0.8576.80.2
BUF0.4843.90.1
OAK0.2421.90.1
DEN0.4033.2-0.2
PIT0.7966.3-0.3
IND0.9277.3-0.3
ARI0.4233.4-0.4
MIN0.4633.7-0.7
ATL0.4023.2-1.2
SEA0.6945.5-1.5
PHI0.5734.6-1.6
STL0.2101.7-1.7
NYJ0.3012.7-1.7
CIN0.4823.8-1.8
TB0.8057.2-2.2
MIA0.3002.4-2.4

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.




































RankTeamLast WkGWPOpp GWPO RankD Rank
1IND20.920.5822
2NE10.910.5313
3DAL30.850.4735
4TB40.800.5248
5PIT50.790.4494
6SEA60.690.471212
7NYG110.620.46167
8TEN90.600.53271
9JAX70.570.55823
10PHI100.570.49521
11WAS120.560.51246
12GB140.540.451314
13BUF190.480.561911
14CIN150.480.55726
15MIN230.460.471415
16HOU170.460.471824
17CLE200.450.50628
18NO250.450.571130
19SD80.450.441519
20CAR240.430.532218
21BAL160.430.41289
22ARI210.420.512317
23DET220.410.432510
24DEN130.400.551029
25ATL260.400.472020
26KC180.350.422613
27NYJ290.300.542132
28MIA280.300.551731
29OAK270.240.433116
30STL310.210.532927
31CHI300.180.503025
32SF320.160.503222

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.






TEAMOPassORunOIntRateOFumRateDPassDRunDIntRatePenRate
DET6.234.280.0290.0456.023.820.0440.30
AVG6.174.080.0320.0256.224.060.0320.38


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.












VprobVisitorHomeHprob
0.54DETARI0.46
0.54NYGDET0.46
0.43GBDET0.57
0.42DETMIN0.58
0.78DALDET0.22
0.46DETSD0.54
0.31KCDET0.69
0.39DETGB0.61


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.













WinsProbAt least
140.000.00
130.010.01
120.060.07
110.160.23
100.260.50
90.270.76
80.170.93
70.060.99
60.011.00