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 #4 of the 2025 season. These predictions are a bit tentative, in the sense that it usually takes Zoltar about four weeks to hit his stride.

Zoltar:    seahawks  by    0  opp =   cardinals    | Vegas:   cardinals  by  2.5
Zoltar:     vikings  by    3  opp =    steelers    | Vegas:     vikings  by  2.5
Zoltar:  commanders  by    0  opp =     falcons    | Vegas:  commanders  by  3.5
Zoltar:       bills  by   13  opp =      saints    | Vegas:       bills  by 12.5
Zoltar:       lions  by   15  opp =      browns    | Vegas:       lions  by 10.5
Zoltar:      texans  by    9  opp =      titans    | Vegas:      texans  by  5.5
Zoltar:    patriots  by    1  opp =    panthers    | Vegas:    patriots  by  3.5
Zoltar:    chargers  by    7  opp =      giants    | Vegas:    chargers  by  3.5
Zoltar:      eagles  by    0  opp =  buccaneers    | Vegas:      eagles  by  1.5
Zoltar:        rams  by    4  opp =       colts    | Vegas:        rams  by  6.5
Zoltar: fortyniners  by    6  opp =     jaguars    | Vegas: fortyniners  by  5.5
Zoltar:      chiefs  by    6  opp =      ravens    | Vegas:      chiefs  by  1.5
Zoltar:     raiders  by    1  opp =       bears    | Vegas:       bears  by  1.5
Zoltar:     packers  by    1  opp =     cowboys    | Vegas:     packers  by  1.5
Zoltar:    dolphins  by    6  opp =        jets    | Vegas:    dolphins  by  5.5
Zoltar:     broncos  by    2  opp =     bengals    | Vegas:     broncos  by  2.5

Zoltar theoretically suggests betting when the Vegas line is “significantly” different from Zoltar’s prediction. In mid-season I typically use 3.0 points difference but for the first few weeks of the season I am usually a bit more conservative and use 4.0 or 5.0 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 strongly biased towards Vegas underdogs. The favor-underdogs effect seems less pronounced this season, compared to previous seasons.

For week #4, using a conservative 4.0-point difference threshold, Zoltar has two suggestions:

browns       at        lions: Bet on Vegas favorite lions
ravens       at       chiefs: Bet on Vegas favorite chiefs

As a human, I agree with Zoltar on the Lions, but I’m skeptical about picking the Chiefs.

A bet on the favorite Lions against the Browns will win only if the Lions win by more than 10.5 points. If the Browns win outright, or if the Lions win but by less than 10.5 points, the bet on the Lions loses. If the Lions win by exactly 10.5 points, the bet is a push — which of course is impossible and the reason why point spreads usually have a 0.5 factor.

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 overhead).

In week #3, against the Vegas point spread, using a conservative 4.0-point threshold, Zoltar went a very good 3-0. For the season so far, Zoltar is 6-3 against the spread (67% win rate).

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 #3, just predicting the winning team, Zoltar went an excellent 10-1 (with five games to close to call). Vegas was 11-5 at just predicting winners in week #3 — a typical result.

Left: Electric football was invented in the late 1940s. This is a 1960s era version. My brother Roger and I played this game a lot. The game is still very popular today.

Right: The 3M Pro Football game was introduced in 1966 but is no longer manufactured. I’ve played it a few times and enjoyed it.