Zoltar is my NFL football machine learning prediction system. It’s a hybrid system that uses a custom reinforcement learning algorithm plus a neural network. Here are Zoltar’s predictions for week #13 of the 2017 NFL season:

Zoltar:     cowboys  by    6  dog =    redskins    Vegas:     cowboys  by  1.5
Zoltar:     packers  by    6  dog =  buccaneers    Vegas:     packers  by    2
Zoltar:    dolphins  by    6  dog =     broncos    Vegas:     broncos  by    1
Zoltar:    patriots  by    5  dog =       bills    Vegas:    patriots  by    8
Zoltar:     falcons  by    2  dog =     vikings    Vegas:     falcons  by    3
Zoltar:       bears  by    6  dog = fortyniners    Vegas:       bears  by  3.5
Zoltar:     jaguars  by    6  dog =       colts    Vegas:     jaguars  by  9.5
Zoltar:      ravens  by    2  dog =       lions    Vegas:      ravens  by    3
Zoltar:      chiefs  by    4  dog =        jets    Vegas:      chiefs  by  3.5
Zoltar:      titans  by    6  dog =      texans    Vegas:      titans  by  6.5
Zoltar:    chargers  by   13  dog =      browns    Vegas:    chargers  by   13
Zoltar:     raiders  by    7  dog =      giants    Vegas:     raiders  by  7.5
Zoltar:      saints  by    6  dog =    panthers    Vegas:      saints  by    4
Zoltar:        rams  by    0  dog =   cardinals    Vegas:        rams  by    7
Zoltar:      eagles  by    0  dog =    seahawks    Vegas:      eagles  by    4
Zoltar:    steelers  by    5  dog =     bengals    Vegas:    steelers  by    6

Zoltar theoretically suggests betting when the Vegas line is more than 3.0 points different from Zoltar’s prediction. In previous years, Zoltar typically would recommend two or three games, but this season Zoltar and the Las Vegas line have differed greatly, and so Zoltar has been recommending six to eight games per week — the large number of injuries to key players is the primary factor for the difference this season. This week, Zoltar has six hypothetical suggestions.

1. Zoltar likes the Vegas favorite Cowboys over the Redskins. Vegas has the Cowboys as slim 1.5 point favorites but Zoltar thinks the Cowboys are 6 points better than the Redskins. So, a bet on the Cowboys will pay you only if the Cowboys win by more than 1.5 points (in other words, 2 points or more).

2. Zoltar likes the Vegas favorite Packers against the Buccaneers. Vegas thinks the Packers are just 2.0 points better than the Buccaneers.

3. Zoltar likes the Vegas underdog Dolphins against the Broncos. Vegas has the Broncos as 1.0 point favorites, but Zoltar thinks the Dolphins are 6 points better than the Broncos. A bet on the Dolphins will pay off if the Dolphins win by any number of points, or if the Broncos win, but by less than 1 point — which is impossible, so the game is really a toss up (if the Broncos win by exactly 1 point, the game is a push).

4. Zoltar likes the Vegas underdog Colts against the Jaguars.

5. Zoltar likes the Vegas underdog Cardinals against the Rams.

6. Zoltar likes the Vegas underdog Seahawks against the Eagles.

This NFL season has been extraordinarily more volatile than any season in recent memory. There are key injuries, literally every week, and the Vegas point spreads change rapidly and in large amounts.

==

Zoltar had a so-so week last week. Against the Vegas spread, which is what Zoltar is designed to predict, Zoltar went 4-3. (I have an advanced version of Zoltar that takes injuries into account — advanced Zoltar went 1-0 in week #13).

For the 2017 season so far, against the Vegas point spread, Zoltar is a mediocre 36-22 (62% accuracy). If you must bet $110 to win $100 (typical in Vegas) then you must theoretically predict with 53% or better accuracy to make money, but realistically you must predict at 60% or better accuracy.

Just for fun, I also track how well Zoltar does when only predicting which team will win. This isn’t really useful except for parlay betting. For week #12, Zoltar was an excellent 14-2 just predicting which team would win

For comparison purposes, I also track how well Bing and the Vegas line do when just predicting who will win. In week #12, Bing was a good 12-4, and Vegas was also good at 13-3 when just predicting winners.

For the 2017 season so far, just predicting the winning team, Zoltar is 118-58 (67% accuracy), Bing is 115-61 (65% accuracy), and Vegas is slightly behind at 109-60 (64% accuracy). The best humans are typically about 67% accurate predicting winners, so currently, both Zoltar and Bing are just about as good as the best humans when just predicting which team will win.

My Zoltar is named after the arcade fortune teller machine, which in turn is named after a machine from the 1988 movie “Big”