The 2015 NFL season now has 9 weeks under its belt and it’s time to see how Amos has done this season against a broader view of prediction models, as well as how Amos has performed compared to the previous season’s Bischoff model.
For those just joining, Amos is a statistical model created to predict the outcomes of each NFL game. Amos currently predicts games straight-up. Additionally, Amos then predicts the remainder of the season to compute the ending record of each team.
Amos also has peers within the NFL prediction field. While there are a number of sources for game predictions, I have identified Microsoft’s Bing Prediction (Cortana), ESPN’s Football Power Index (FPI) and Nate Silver’s Elo as benchmarks for comparison to Amos.
First, a look at how the Amos model compares with last year’s model.
The Bischoff model contained much more volatility up to this week last year. Specifically, having two breakout performances within Week 5 (93%) and 6 (87%), but also deepening to a 44% performance in Week 2. In comparison, Amos has performed much more consistently, not having a performance under 50%. Amos still has ground to cover to prove superiority over the Bischoff model.
Let’s look at how Amos’ performance over the season has matched up with its selected peers.
Amos hasn’t had any standout weeks within its peer model group, but that’s also due to Amos’ peers being high performers. Amos can’t boast a week where it has been the top performer, yet its consistency is what has allowed it to remain a top contender. Bing has struggled to having any form of consistency, with an explosive Week 8, reaching 93%, Bing has also stooped to a 38% Week 2. Elo has been a dominant predictive model for the majority of the season, boasting two consecutive 86% performances over Week 5 and Week 6, and is currently in the lead. Finally, FPI has performed consistently as well, but has not had any extraordinary weeks yet.
How about a broader scope?
Amos has proven to be a top contender within the prediction arena, however not strong enough to gain a top 3 stop. Currently sitting in 7th place. The top performing at mid-season is Nate Silver’ Elo.
Finally, probably the most frequent question I get regarding Amos: how does Amos perform if you take it to Vegas? I’ve been pulling moneyline odds throughout the season as Amos’ algorithm finishes computing the probable winners each week. I’ve staged a scenario of a starting account with $1,000 and each of Amos’ predictions are placed with a $100 bet on the moneyline.
Amos has done very well with its picks. What’s important in betting is not only getting as many as you can correct, but also predicting key upsets to cash in on the odds. With three weeks of having losses, this account would be up to $1,667.03 on the season.
There’s still a whole next half to the season, and it’ll be exciting to see how things develop moving forward. I’ll continue to compile these data points and publish them at the end of the season.
Now for the familiar stats and Week 10 picks.
Amos, not having a good week, went 7-6, having misses of Green Bay, New Orleans, Atlanta, Denver, Dallas and San Diego.
ESPN’s FPI went 9-4, having misses of Green Bay, New Orleans, Dallas and San Diego.
Microsoft’s Bing went 9-4, only having misses of New Orleans, Atlanta, Denver and San Diego.
Nate Silver’s Elo went 8-5, having misses of New Orleans, Atlanta, Denver, Dallas and San Diego.
Moving onto Week 10 picks, only three games having differentiating picks – two of which are Bing’s.
Amos has the same picks as everyone, besides Washington vs New Orleans, where Amos is favoring Washington (61%).
Microsoft’s Bing has two picks differing from the pack, which are Tampa Bay over Dallas (51%) and Oakland over Minnesota (55%). Both being interpreted as very close games.
Nate Silver’s Elo and ESPN’s FPI have no differentiating picks this week.
Finally, from the performances of each of the teams, Amos has simulated the rest of the season. The biggest gainer this week was Baltimore, who was able to pick up 2 more wins over the season as its remaining opponents have been devalued. The biggest losers this week are NY Jets, Denver, Green Bay and Tampa Bay, all of which picked up two more loses throughout the remaining season.
However, this week something very interesting happened. Denver’s devaluation has now left New England as going the remainder of the season without being beat. Up to this point, Amos only gave Denver the ability to beat New England. With Denver’s performance at Indianapolis, Amos currently does not see Denver winning against New England November 29th, leaving New England with a perfect season.
See the results below and each team’s change from the last week. Next week.
Week 1: 11-5
Week 2: 9-7
Week 3: 11-5
Week 4: 9-6
Week 5: 10-4
Week 6: 10-4
Week 7: 9-5
Week 8: 10-4
Week 9: 7-6
Week 11: 10-4
Week 12: 9-7
Week 13: 11-5
Week 14: 9-7
Week 15: 14-2
Week 16: 7-9
Week 17: 9-7