Zoltar is my NFL football prediction computer program. It uses a neural network and a form of quasi-reinforcement learning. Here are Zoltar’s predictions for week #3 of the 2026 season. These predictions are fuzzy, in the sense that it usually takes Zoltar about four weeks to hit his stride.

Zoltar:     packers  by    6  opp =     falcons    | Vegas:     packers  by  7.5
Zoltar:       bills  by    6  opp =    chargers    | Vegas:       bills  by  7.5
Zoltar:    panthers  by    0  opp =      browns    | Vegas:    panthers  by  2.5
Zoltar:      texans  by    0  opp =       colts    | Vegas:      texans  by  2.5
Zoltar:       lions  by    8  opp =        jets    | Vegas:       lions  by  6.5
Zoltar:     jaguars  by    1  opp =    patriots    | Vegas:     jaguars  by  2.5
Zoltar:      chiefs  by    0  opp =    dolphins    | Vegas:      chiefs  by 11.5
Zoltar:      giants  by    6  opp =      titans    | Vegas:      giants  by  2.5
Zoltar:    steelers  by    6  opp =     bengals    | Vegas:     bengals  by  3.5
Zoltar:    seahawks  by    7  opp =  commanders    | Vegas:    seahawks  by  6.5
Zoltar: fortyniners  by   11  opp =   cardinals    | Vegas: fortyniners  by  8.5
Zoltar:     vikings  by    0  opp =  buccaneers    | Vegas:     vikings  by  1.5
Zoltar:      ravens  by    1  opp =     cowboys    | Vegas:      ravens  by  2.5
Zoltar:      saints  by    6  opp =     raiders    | Vegas:      saints  by  3.5
Zoltar:     broncos  by    6  opp =        rams    | Vegas:        rams  by  2.5
Zoltar:       bears  by    1  opp =      eagles    | Vegas:      eagles  by  4.5

Zoltar theoretically suggests betting when the Vegas line is “significantly” different from Zoltar’s prediction. In early-season I usually use 3.5 points difference as the threshold.

For week #3, using a 3.5-point threshold (suggest if difference between Vegas and Zoltar is strictly greater than 3.5 points) Zoltar likes two Vegas favorites and one underdog:

chiefs       at     dolphins: Bet on Vegas underdog dolphins
bengals      at     steelers: Bet on Vegas underdog steelers
rams         at      broncos: Bet on Vegas underdog broncos
eagles       at        bears: Bet on Vegas underdog bears

For example, a bet on the Vegas underdog Dolphins against the Chiefs will pay off if the Dolphins win by any score, or if the favored Chiefs win but by less than the point spread of 11.5 points (11 points or fewer — the “.5” prevents pushes).

Zoltar’s initialization algorithm is strongly biased towards underdogs, especially early in the season. As a human, of Zoltar’s four recommendations, the only one I would consider is taking underdog Broncos.

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 to account for overhead and mistakes.

In week #2, using a 3.5-point threshold, Zoltar went 1-2 against the spread. Zoltar was correct in picking the underdog Colts against the Chiefs. The Chiefs won, but only by 3 points in overtime.

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 10-6 which is OK but not great. Vegas was 11-5 at just picking winners.

My football prediction program is named after the Zoltar fortune teller machine you can find in arcades. Fortune teller machines have been around for over 100 years. The two machines on the right are from about 1926.