Saturday, December 24, 2016

Guess the Lines - Final Results

Here are the week 16 results for the final installment of "Guess the Lines". This is a weekly feature in which I put my Vegas NFL rankings to the test by predicting the point spread of upcoming NFL games. I compare the accuracy of my prediction against those of Bill Simmons and Cousin Sal, as featured on their weekly NFL podcast. In addition, I also calculate an implied point spread from ESPN Chalk's weekly Vegas NFL rankings.

I'm not sure if Bill and Sal intend to do a Week 17 podcast, but I've decided to halt the prediction contest at Week 16. Week 17 lines can be a bit wonky due to teams that don't have anything to play for.

Here are the week 16 results. Bill Simmons wins this week with a mean absolute prediction error of 0.84 points.

Week 16 Resultspredicted point spread:
gameaGPFhGPFmeespnbillsalactual
CIN @ HOU 1.0 -2.3 +1.0 -2.0 -3.5 -1.5 -2.5
DEN @ KC 2.0 3.1 -3.5 -2.5 -5.0 -4.0 -4.5
TB @ NO -0.1 0.4 -3.0 -1.0 -1.0 -3.5 -3.0
ARZ @ SEA 0.8 4.5 -6.0 -5.0 -8.0 -7.5 -9.0
SD @ CLE -0.7 -8.2 +5.0 +6.0 +8.0 +8.0 +6.5
IND @ OAK -0.8 3.2 -6.5 -4.5 -6.0 -6.0 -4.0
MIA @ BUF -2.1 -0.1 -4.5 -4.5 -2.5 -1.0 -3.5
ATL @ CAR 3.8 0.0 +1.5 +3.0 +4.0 +3.5 +4.0
SF @ LA -7.7 -6.6 -3.5 -4.5 -3.0 -4.0 -3.5
NYG @ PHI 0.8 -0.3 -1.5 -1.5 +3.5 +6.0 +3.0
NYJ @ NE -6.2 6.9 -15.5 -14.5 -14.5 -13.0 -16.0
DET @ DAL -0.5 5.0 -8.0 -7.5 -6.5 -6.0 -7.0
BAL @ PIT 2.4 5.9 -6.0 -6.0 -4.0 -4.0 -4.5
WAS @ CHI 2.2 -5.3 +5.0 +3.5 +3.5 +5.0 +3.5
TEN @ JAC 0.2 -5.2 +3.0 +2.0 +4.0 +6.5 +4.5
MIN @ GB 1.0 3.0 -4.5 -3.5 -7.5 -6.5 -7.0
Mean Absolute Error 1.75 1.64 0.84 1.38

The closest prediction for each game is shown in bold. As a reminder, my predictions and the ESPN predictions were made a week ago Tuesday, so as not to be influenced by the week 16 opening lines.

My rankings finished last this week, although I will point out that my biggest miss was on the NYG @ PHI line. At the time of the pod, the Giants were favored by 3, which was 4.5 points off from my prediction of Philly as a 1.5 point favorite. However, the line appeared to have closed much closer to that prediction (with some variation by sportsbook). To keep things simple though, the official score is based on the line as reported by Sal on the podcast.

And here is the final tally over seven weeks of predictions. A lower score is better (we are measuring average prediction error).

Mean Absolute Prediction Error:
week me espn bill sal
10 1.29 1.86 1.89 1.68
11 1.11 1.46 1.29 1.07
12 1.38 1.75 1.28 1.22
13 1.13 1.70 0.87 1.10
14 1.53 1.91 1.00 1.66
15 1.34 1.59 1.13 0.94
16 1.75 1.66 0.84 1.38
average 1.36 1.70 1.19 1.29

Bill Simmons finished with the overall best accuracy, with a mean absolute prediction error of just 1.19 points. Cousin Sal finished second at 1.29. My rankings were not too far behind at 1.36, and the ESPN Chalk rankings finished last at 1.70.

Despite finishing third, I am pretty happy with these results. My rankings, which are based on a simple linear regression model, performed consistently better than the ESPN rankings, which are based on the judgment of expert handicappers. In addition, they were almost as accurate as Bill and Sal, despite having to make predictions nearly a week earlier (e.g. the Week 16 predictions above were made the Tuesday prior to the Week 15 games).

3 comments: