Zoltar is my NFL football prediction computer program. It uses a neural network and a type of (quasi) reinforcement learning. Here are Zoltar’s predictions for week #2 of the 2026 season. These predictions are fuzzy, in the sense that it usually takes Zoltar about four weeks to hit his stride.

Zoltar:       bills  by    6  opp =       lions    | Vegas:       bills  by  4.5
Zoltar:     falcons  by    4  opp =    panthers    | Vegas:    panthers  by  1.5
Zoltar:       bears  by    6  opp =     vikings    | Vegas:       bears  by  5.5
Zoltar:      texans  by    6  opp =     bengals    | Vegas:      texans  by  2.5
Zoltar:    patriots  by    6  opp =    steelers    | Vegas:    patriots  by  5.5
Zoltar:     packers  by    1  opp =        jets    | Vegas:     packers  by  3.5
Zoltar:      eagles  by    5  opp =      titans    | Vegas:      eagles  by  6.5
Zoltar:      ravens  by    6  opp =      saints    | Vegas:      ravens  by  8.5
Zoltar:  buccaneers  by    6  opp =      browns    | Vegas:  buccaneers  by  9.5
Zoltar:     broncos  by    2  opp =     jaguars    | Vegas:     broncos  by  2.5
Zoltar:    chargers  by    9  opp =     raiders    | Vegas:    chargers  by  6.5
Zoltar:    seahawks  by    5  opp =   cardinals    | Vegas:    seahawks  by  4.5
Zoltar:     cowboys  by    6  opp =  commanders    | Vegas:     cowboys  by  3.5
Zoltar: fortyniners  by    9  opp =    dolphins    | Vegas: fortyniners  by 13.5
Zoltar:      chiefs  by    2  opp =       colts    | Vegas:      chiefs  by  6.5
Zoltar:        rams  by    9  opp =      giants    | Vegas:        rams  by  7.5

Zoltar theoretically suggests betting when the Vegas line is “significantly” different from Zoltar’s prediction. For the first few weeks of the season I usually use a mildly aggressive 3.5 points difference as the advice threshold criterion.

For week #2 using 3.5 points as threshold:

panthers     at      falcons: Bet on Vegas underdog falcons
dolphins     at  fortyniners: Bet on Vegas underdog dolphins
colts        at       chiefs: Bet on Vegas underdog colts

For example, a bet on the Vegas underdog Falcons against the Panthers will pay off if the Falcons win by any score, or if the favored Panthers win but by less than the 1.5 point spread (in other words, by 1 point).

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 take into account things like random mistakes and overhead.

In week #1, against the Vegas point spread, using a 3.5-point threshold, Zoltar went 0-3. Zoltar picked Vegas underdogs Colts, Texans, and Broncos, but all three of those underdogs lost convincingly.

But Zoltar usually takes three weeks to catch up to what humans know at the beginning of the season (new players, new coaches).

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 #1, just predicting the winning team, Zoltar went 6-5 (with 5 games predicted even) which is pretty bad. For Las Vegas, just predicting which team will win, Vegas went 11-5, which is good but not great.

My NFL prediction system is named after the Zoltar fortune teller machine you can find in arcades.

Right: This is an old (from 2018) sheet where you can see the point spread for a given week. You would place a bet by going to a person at a desk in the sports book. This old style, of using paper and pencil and attendant, is rarely used anymore – replaced by online applications and kiosks.