Zoltar is my NFL football prediction computer program. It uses a neural network and a form of reinforcement learning. Here are Zoltar’s predictions for week #10 of the 2024 season.

Zoltar:     broncos  by   12  opp =     raiders    | Vegas:     broncos  by  6.5
Zoltar:       colts  by    3  opp =     falcons    | Vegas:     falcons  by  1.5
Zoltar:       bears  by    7  opp =      giants    | Vegas:       bears  by  5.5
Zoltar:      texans  by    6  opp =     jaguars    | Vegas:      texans  by  3.5
Zoltar:     vikings  by    6  opp =      ravens    | Vegas:      ravens  by  3.5
Zoltar:    panthers  by    6  opp =      saints    | Vegas:    panthers  by  4.5
Zoltar:       bills  by    5  opp =    dolphins    | Vegas:       bills  by  5.5
Zoltar:        jets  by    4  opp =      browns    | Vegas:        jets  by  2.5
Zoltar:  buccaneers  by    6  opp =    patriots    | Vegas:  buccaneers  by  4.5
Zoltar:    seahawks  by    8  opp =   cardinals    | Vegas:    seahawks  by  1.5
Zoltar:        rams  by    0  opp = fortyniners    | Vegas: fortyniners  by  1.5
Zoltar:       lions  by    3  opp =  commanders    | Vegas:  commanders  by  1.5
Zoltar:    chargers  by    5  opp =    steelers    | Vegas:    chargers  by  4.5
Zoltar:      eagles  by    0  opp =     packers    | Vegas:      eagles  by  1.5

Zoltar theoretically suggests betting when the Vegas line is “significantly” different from Zoltar’s prediction. In mid-season I typically use a threshold of 3 points difference but in early and late parts of the season I use a more conservative threshold of 4 points difference.

For week #10, using 3 points as the threshold, Zoltar likes three Vegas underdogs and two Vegas favorites:

raiders      at      broncos: Bet on Vegas favorite broncos
falcons      at        colts: Bet on Vegas underdog colts
ravens       at      vikings: Bet on Vegas underdog vikings
cardinals    at     seahawks: Bet on Vegas favorite seahawks
lions        at   commanders: Bet on Vegas underdog lions

A bet on the Vegas underdog Colts against the Falcons will pay off if the Colts win by any score, or if the favored Falcons win but by less than 1.5 points (i.e., 1 point). If the favored team wins by exactly the point spread, the wager is a push. This is why point spreads often have a 0.5 added — called “the hook” — to eliminate pushes.

I use the early Vegas point spreads, usually posted on late Monday night, right after the Monday Night Football game. By the time you read this, the point spreads will likely have changed. I’ve noticed that, compared to last year, point spreads are changing more dramatically by late Tuesdays than they used to. A swing of 7 points is not uncommon. I speculate that this is due to a huge increase in betting. When a lot of money is bet on one team in a matchup, the bookmakers must make a huge change in the point spread to encourage betting on the other team. Bookmakers only make money when the betting amounts are close to equal on both teams.

This is the part of the season where injuries start having a big effect on the point spread. Weeks 10 through 14 are usually Zoltar’s weakest.

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 and miscellaneous mistakes.

In week #9, against the Vegas point spread, Zoltar went a terrible 1-3 using 3.0 points as the advice threshold because of the large point spread shifts. Alas, if Zoltar had been more cautious and used 4.0 points as the threshold, he would have been 1-1 against the spread.

For the season, against the spread, Zoltar is 22-16 (~57% accuracy).

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 #9, just predicting the winning team, Zoltar went only 3-4 with 7 games to close to call, which is terrible. Vegas was much better, going 8-6 at just predicting the winning team.

My prediction system is math-based in the sense that it computes a numeric rating for each team and then uses ratings to compute the predicted margin of victory of the better team. An entirely different approach for predicting NFL football scores is to use a sophisticated simulation program and then simulate a game many thousands of times.

Right: “Half-Time Football” by Lakeside was produced in the late 1970s and early 80s. It’s strictly a dice game and is medium complexity.