Win Probabilities and Expected Wins 2007 - Week 10
Tables updated for all games and teams.
The following table lists the probabilities that the home team would win each game given the rushing averages, passing averages (includes sack yards lost and sacks as pass attempts), sack rates, third down conversion rates, penalty yards, and interception and fumble rates (by number of completions and rushing plays) for both teams. A logistic regression model trained on 1996-2006 data computes the probabilities. The final score margin is Home team points - Away team points. The sum of these probabilities of victory form the expected win totals in the second table after the jump. The tables will be updated as more box scores come in.
| Home Team | Away Team | P(Home Team Won) (%) | Final Score Margin |
| CAR | ATL | 7.3776 | -7 |
| MIA | BUF | 92.4880 | -3 |
| BAL | CIN | 3.7658 | -14 |
| NYG | DAL | 18.7485 | -11 |
| KC | DEN | 25.1351 | -16 |
| GB | MIN | 99.5719 | 34 |
| TEN | JAX | 14.3756 | -15 |
| WAS | PHI | 6.2245 | -8 |
| PIT | CLE | 97.6274 | 3 |
| NO | STL | 10.4471 | -8 |
| ARI | DET | 97.8358 | 10 |
| OAK | CHI | 0.3758 | -11 |
| SD | IND | 3.8619 | 2 |
| SEA | SF | 96.7554 | 24 |
Teams that should've won but didn't
Miami Dolphins (vs. Bills), Win Prob = 92.4880% Buffalo fumbled twice without losing the ball. They threw an interception (but they were 1.8 yards more efficient in passing). Buffalo converted 20% of third downs, Miami 47%. Miami outrushed them 1.3 yards per rush and gave up one fewer sack. The sack they gave up, though, was a safety. Special teams also made a large difference in this game. This is the fifth game Miami has lost by a field goal.
Indianapolis Colts (@ Chargers), Win Prob = 96.1381% Manning threw 6 INTs. The Colts lost their first and second string LTs, their RT, their #1 and #3 WRs, their best TE, their best DE, and they were a 29 yard field goal away from winning. Yet Indy managed 2 yards more per pass, a better sack rate, a much better 3rd down conversion rate, and fewer penalty yards. Plus, San Diego had 2 INTs and 2 fumbles of their own. This number really took me by surprise. I wasn't surprised the Colts "should have" won, but a 96% win probability? Damn.
| Team | Actual Wins | Expected Wins | Projected Win Total | |
| AFC East | ||||
| BUF | 5.0 | 3.9003 | 6.9338 | |
| MIA | 0.0 | 3.4284 | 6.0950 | |
| NE | 9.0 | 7.8664 | 13.9848 | |
| NYJ | 1.0 | 2.3378 | 4.1562 | |
| AFC North | ||||
| BAL | 4.0 | 3.6118 | 6.4210 | |
| CIN | 3.0 | 3.6679 | 6.5207 | |
| CLE | 5.0 | 3.9322 | 6.9905 | |
| PIT | 7.0 | 7.3802 | 13.1204 | |
| AFC South | ||||
| HOU | 4.0 | 3.8518 | 6.8477 | |
| IND | 7.0 | 8.0477 | 14.3070 | |
| JAX | 6.0 | 5.5259 | 9.8238 | |
| TEN | 6.0 | 4.8002 | 8.5337 | |
| AFC West | ||||
| DEN | 4.0 | 3.5485 | 6.3084 | |
| KC | 4.0 | 4.4609 | 7.9304 | |
| OAK | 2.0 | 2.3187 | 4.1222 | |
| SD | 5.0 | 4.2462 | 7.5488 | |
| NFC East | ||||
| DAL | 8.0 | 7.4061 | 13.1665 | |
| NYG | 6.0 | 4.6066 | 8.1895 | |
| PHI | 4.0 | 5.3629 | 9.5341 | |
| WAS | 5.0 | 3.8150 | 6.7822 | |
| NFC North | ||||
| CHI | 4.0 | 2.4318 | 4.3231 | |
| DET | 6.0 | 5.3984 | 9.5972 | |
| GB | 8.0 | 6.2483 | 11.1081 | |
| MIN | 3.0 | 3.6055 | 6.4099 | |
| NFC South | ||||
| ATL | 3.0 | 4.9963 | 8.8824 | |
| CAR | 4.0 | 3.5816 | 6.3672 | |
| NO | 4.0 | 3.7958 | 6.7481 | |
| TB | 5.0 | 5.9208 | 10.5259 | |
| NFC West | ||||
| ARI | 4.0 | 3.5776 | 6.3601 | |
| STL | 1.0 | 3.0571 | 5.4349 | |
| SF | 2.0 | 1.1720 | 2.0835 | |
| SEA | 5.0 | 6.0991 | 10.8428 | |
5 comments:
Derek, I have enjoyed reading your blog and about the work youre doing . In reveieing each weeks data, I am having trouble finding the two prediction tables for Week 8, both the adjusted and non-adjusted tables. Am I just not looking hard enough??!! Thanks.
--Mark, a fan.
Derek, I truly need to learn more about linear regression and such. Your new table for week 10 notes the win probability for MIA was 96.488%, but the previous tables that predicted margin of win had their probability as only 64+%. Why am I confused???
marko, I'll check the week 8 tables later today. I believe I deleted the posts because of errors in the calculations that I didn't bother correcting.
The 64% probability was the predictive probability that the Dolphins would win the game given their stats in the season up to that point. The 96% probability is the retrodictive probability that the Dolphins had won the game given the actual box score.
In other words, there was a 64% chance that the Dolphins were going to outplay the Bills on Sunday. They did that to the point, where 9.6 times out of 10, they win that game.
Thanks Derek. Then I wasnt too far off base in interpreting your information. Im sure to be back with more questions!
Reposted the predictions here.
I remember deleting one week's predictions post because of some error in the calculations. These are the correct ones.
Post a Comment