Monday, January 12, 2015

NFL Picks of the Week Final results

As I blogged earlier I wrote a series of scripts to grab NFL game results data, points for/against, and the various records of NFL teams in an attempt to predict the winners of the weekly NFL matchups. This was a straight head to head pickem (no spread picks) to determine the winner of the matchup. The projections were handled using Log5 and the Pythagorean Wins Expectations calculations.

The projections were started week 11. So how did I do?


Overall Record: 73-35
Win Pct: 67.5%

118 - 6
1213 - 2
1310 - 6
1411 - 5
1512 - 4
168 - 8
1712 - 4

Week 11 and Week 16 were the worst, however I did not have a single losing week. What teams were the most difficult to project?

TeamLog 5 Wrong
Arizona Cardinals4
Philadelphia Eagles4
San Francisco 49ers3
Cleveland Browns3

TeamsPyth Wrong
Philadelphia Eagles4
Cleveland Browns3
New Orleans Saints3

I used the Pythagorean Wins Expectations calculations for my actual matchup picks. If I used Log 5 to project the winners I would have missed 3 more picks on the year. Therefore the Log 5 win percentage was 64.2%.

The Eagles

The Philadelphia Eagles caused the projections the most grief showing up on both the Log 5 and Pyth top Wrong list. What does that mean? It basically means the Eagles lost about 4 games they should not have lost. However if we look closer only 2 of those games really stand out as terrible.

The week 16 debacle against the Redskins is Exhibit A in the Eagles bad second half of the season. According to the Pythagorean Projection the Redskins had a 26.91% chance of winning that game.

The week 11 showdown with the Packers was absolutely ugly with a final score of 53-20. The Packers were given a slight edge in the projection 54.85%. So the game should have been a lot closer than the score.

The Cowboys loss in week 15 was a lot closer. Pythagorean projection stated the Eagles had a 62.95% chance of winning the game and the Cowboys pulled ahead late in the game. Same story with the Week 14 Seahawks game. The projection had the Eagles with a 55.87% chance to win against Seattle.

All of these teams are playoff teams. Looking at the Eagles entire schedule you see one loss from a non-playoff team and that is the Week 4 loss at San Francisco by less than one score.

The Eagles did not make the playoffs, but if you throw out the Redskins game, they had a great season considering their opponents and outcomes.

Future Plans

I have a few things to work on for next year. I need to solve some current week logic. I had some issues with the system recognizing the current NFL week. I may create a week table so it will just pick from a list.

Also I would like to automate it from end to end. I have been executing the scripts and such manually to make sure everything was working.


Post a Comment