Zoltar is my NFL football prediction computer program. It uses reinforcement learning and a neural network. Here are Zoltar's predictions for week #3 of the 2021 season. These predictions are tentative, in the sense that it usually takes Zoltar about three weeks to hit his stride.

Zoltar:    panthers  by    0  dog =      texans    Vegas:    panthers  by  7.5  Zoltar:       bills  by    6  dog =    redskins    Vegas:       bills  by    9  Zoltar:      browns  by    6  dog =       bears    Vegas:      browns  by  7.5  Zoltar:      ravens  by    0  dog =       lions    Vegas:      ravens  by    6  Zoltar:   cardinals  by    3  dog =     jaguars    Vegas:   cardinals  by  7.5  Zoltar:      chiefs  by    6  dog =    chargers    Vegas:      chiefs  by  6.5  Zoltar:      saints  by    0  dog =    patriots    Vegas:    patriots  by    3  Zoltar:      giants  by    5  dog =     falcons    Vegas:      giants  by    3  Zoltar:      titans  by    5  dog =       colts    Vegas:      titans  by  5.5  Zoltar:    steelers  by    6  dog =     bengals    Vegas:    steelers  by  4.5  Zoltar:     broncos  by    6  dog =        jets    Vegas:     broncos  by 10.5  Zoltar:     raiders  by    4  dog =    dolphins    Vegas:     raiders  by   10  Zoltar:    seahawks  by    0  dog =     vikings    Vegas:    seahawks  by    1  Zoltar:        rams  by    2  dog =  buccaneers    Vegas:  buccaneers  by    1  Zoltar:     packers  by    0  dog = fortyniners    Vegas: fortyniners  by    1  Zoltar:     cowboys  by    5  dog =      eagles    Vegas:     cowboys  by    4

Zoltar theoretically suggests betting when the Vegas line is "significantly" different from Zoltar's prediction. In mid-season I use 3.0 points difference but for the first few weeks of the season I go a bit more conservative and use 4.0 points difference as the advice threshold criterion.

At the beginning of the season, because of Zoltar's initialization (all teams regress to an average power rating) and other algorithms, Zoltar is very strongly biased towards Vegas underdogs. I need to fix this.

1. Zoltar likes Vegs underdog Texans against the Panthers.
2. Zoltar likes Vegs underdog Lions against the Ravens.
3. Zoltar likes Vegs underdog Jaguars against the Cardinals.
4. Zoltar likes Vegs underdog Jets against the Broncos.
5. Zoltar likes Vegs underdog Dolphins against the Raiders.

For example, a bet on the underdog Texans against the Panthers will pay off if the Texans win by any score, or if the favored Panthers win but by less than 7.5 points (in other words, win by 7 points or fewer).

Theoretically, if you must bet $110 to win $100 (typical in Vegas) then you'll make money if you predict at 53% accuracy or better. But realistically, you need to predict at 60% accuracy or better.

In week #2, against the Vegas point spread, Zoltar went 4-6 (using 3.0 points as the advice threshold).

Just for fun, I track how well Zoltar does when just trying to predict just which team will win a game. This isn't useful except for parlay betting. In week #2, just predicting the winning team, Zoltar went 8-8 which isn't very good but is typical of the first few weeks of the season. In week #2, just predicting the winning team, Vegas -- "the wisdom of the crowd" -- went 11-5.

Zoltar sometimes predicts a 0-point margin of victory. There are five such games in week #3. In those situations, to pick a winner (only so I can track raw number of correct predictions) in the first few weeks of the season, Zoltar picks the home team to win. After that, Zoltar uses his algorithms to pick a winner.


My prediction system is named after the Zoltar fortune teller machine you can find in arcades. Debugging software is kind of like being a detective. Unfortunately there are no crystal balls available to help.