Tweet
First the chart. Estimated wins is from Football outsiders. Py Wins is the football variation of Bill James Pythagorian Wins, from the same source.
TM W EstW RK PyW RK
NE 14 14.1 1 12.3 1
PIT 12 12.2 2 12.1 3
BAL 12 11.4 3 10.6 6
GB 10 10.9 4 12.1 2
PHI 10 10.8 5 9.4 10
ATL 13 10.4 6 11.2 4
NYG 10 10.3 7 10.1 7
NYJ 11 9.8 8 9.8 8
NO 11 9.7 9 9.2 11
SD 9 9.6 10 10.9 5
MIA 7 8.9 11 6.2 23
IND 10 8.8 13 9.1 13
TEN 6 8.8 12 8.5 16
TB 10 8.6 14 8.7 15
HOU 6 8.2 16 7.1 18
CHI 11 8.2 15 9.5 9
KC 10 8 17 9.1 12
DET 6 7.8 18 7.8 17
CLE 5 7.5 19 6.1 24
CIN 4 7.2 20 6.1 25
SF 6 7.1 21 6.8 20
OAK 8 6.9 22 8.9 14
DAL 6 6.8 23 7 19
JAC 8 6.7 25 6.4 22
MIN 6 6.7 24 6 26
SEA 7 6.2 26 5.5 28
DEN 4 6.1 27 5.2 29
WAS 6 5.9 28 5.9 27
STL 7 5.7 29 6.8 21
BUF 4 5.6 30 4.4 31
CAR 2 3.8 31 2.4 32
ARI 5 3.7 32 4.4 30
As you can see, the predictors track pretty well overall. My Chiefs were two wins better than expected, after three straight years of being below expectations.
The biggest difference between the two is Miami, by far, followed by Oakland and New England. In all cases, the record fell between, though the FBO predictor is quite close.
The biggest differences between FBO and record are Bengals, Titans and Browns, all 2 1/2 games worse than predicted, and Bears and Falcons both more than 2 1/2 games better. Vs the Pythagorian predictor only the Titans are more than 2 1/2 off, worse than projected. More than two games off are the Bengals and Packers, which makes sense considering post season.
What to make of this? There is a strong tendancy for teams to zero over time. If they are high one year, there is another year where they are low. In cases where bother predictors are off significantly, the teams are "lucky" or "unlucky", and luck is fickle. As a practical matter, it comes from winning (or losing) blowouts and losing (or winning) close games. Much analysis has been done on records in baseball, and blowouts are a much better predictor of the strength of a team than nail biters.
Green Flags - Cincinatti and Tennessee both were better than their records by at least two games. Cleveland and Detroit are the same, to a lesser degree. All four clubs
Red Flags - Atlanta and Chicago stand well apart. Both won more than two games above expectations. No one else is close, with New Orleans, at barely 1.5 games, next in line.
Notably absent are Arizona, NE, San Diego, Dallas, NYG. All are within a game of the averaged predictors. Washington, Oakland, Philly and Pittsburgh were essntially dead on target. If I were a fan of the Redskins, that would worry me. Ditto the Cardinals, who should have won a game less than they did.
J
TM W EstW RK PyW RK
NE 14 14.1 1 12.3 1
PIT 12 12.2 2 12.1 3
BAL 12 11.4 3 10.6 6
GB 10 10.9 4 12.1 2
PHI 10 10.8 5 9.4 10
ATL 13 10.4 6 11.2 4
NYG 10 10.3 7 10.1 7
NYJ 11 9.8 8 9.8 8
NO 11 9.7 9 9.2 11
SD 9 9.6 10 10.9 5
MIA 7 8.9 11 6.2 23
IND 10 8.8 13 9.1 13
TEN 6 8.8 12 8.5 16
TB 10 8.6 14 8.7 15
HOU 6 8.2 16 7.1 18
CHI 11 8.2 15 9.5 9
KC 10 8 17 9.1 12
DET 6 7.8 18 7.8 17
CLE 5 7.5 19 6.1 24
CIN 4 7.2 20 6.1 25
SF 6 7.1 21 6.8 20
OAK 8 6.9 22 8.9 14
DAL 6 6.8 23 7 19
JAC 8 6.7 25 6.4 22
MIN 6 6.7 24 6 26
SEA 7 6.2 26 5.5 28
DEN 4 6.1 27 5.2 29
WAS 6 5.9 28 5.9 27
STL 7 5.7 29 6.8 21
BUF 4 5.6 30 4.4 31
CAR 2 3.8 31 2.4 32
ARI 5 3.7 32 4.4 30
As you can see, the predictors track pretty well overall. My Chiefs were two wins better than expected, after three straight years of being below expectations.
The biggest difference between the two is Miami, by far, followed by Oakland and New England. In all cases, the record fell between, though the FBO predictor is quite close.
The biggest differences between FBO and record are Bengals, Titans and Browns, all 2 1/2 games worse than predicted, and Bears and Falcons both more than 2 1/2 games better. Vs the Pythagorian predictor only the Titans are more than 2 1/2 off, worse than projected. More than two games off are the Bengals and Packers, which makes sense considering post season.
What to make of this? There is a strong tendancy for teams to zero over time. If they are high one year, there is another year where they are low. In cases where bother predictors are off significantly, the teams are "lucky" or "unlucky", and luck is fickle. As a practical matter, it comes from winning (or losing) blowouts and losing (or winning) close games. Much analysis has been done on records in baseball, and blowouts are a much better predictor of the strength of a team than nail biters.
Green Flags - Cincinatti and Tennessee both were better than their records by at least two games. Cleveland and Detroit are the same, to a lesser degree. All four clubs
Red Flags - Atlanta and Chicago stand well apart. Both won more than two games above expectations. No one else is close, with New Orleans, at barely 1.5 games, next in line.
Notably absent are Arizona, NE, San Diego, Dallas, NYG. All are within a game of the averaged predictors. Washington, Oakland, Philly and Pittsburgh were essntially dead on target. If I were a fan of the Redskins, that would worry me. Ditto the Cardinals, who should have won a game less than they did.
J
- Wonderboy