Zoltar is my NFL football prediction computer program. It uses reinforcement learning and a neural network. Here are Zoltar’s predictions for week #15 of the 2021 season. It usually takes Zoltar about four weeks to hit his stride and takes humans about eight weeks to get up to speed, so weeks six through nine are usually Zoltar’s sweet spot. After week nine, injuries start having a big effect.
Zoltar: chiefs by 0 dog = chargers Vegas: chiefs by 3 Zoltar: bills by 7 dog = panthers Vegas: bills by 9.5 Zoltar: browns by 6 dog = raiders Vegas: browns by 5 Zoltar: colts by 1 dog = patriots Vegas: patriots by 2.5 Zoltar: cardinals by 5 dog = lions Vegas: cardinals by 12.5 Zoltar: jaguars by 2 dog = texans Vegas: jaguars by 3.5 Zoltar: dolphins by 9 dog = jets Vegas: dolphins by 7 Zoltar: cowboys by 0 dog = giants Vegas: cowboys by 10 Zoltar: eagles by 1 dog = redskins Vegas: eagles by 2.5 Zoltar: titans by 0 dog = steelers Vegas: titans by 2.5 Zoltar: packers by 0 dog = ravens Vegas: packers by 2 Zoltar: broncos by 4 dog = bengals Vegas: bengals by 3 Zoltar: fortyniners by 6 dog = falcons Vegas: fortyniners by 7.5 Zoltar: rams by 6 dog = seahawks Vegas: rams by 7 Zoltar: buccaneers by 6 dog = saints Vegas: buccaneers by 11.5 Zoltar: vikings by 0 dog = bears Vegas: vikings by 4
Zoltar theoretically suggests betting when the Vegas line is “significantly” different from Zoltar’s prediction. In mid-season I usually use 3.0 points difference but for the first few weeks and last few weeks of the season I go a bit more conservative and use 4.0 points difference as the advice threshold criterion. In middle weeks I sometimes go ultra-aggressive and use a 1.0-point threshold.
Note: Because of Zoltar’s initialization (all teams regress to an average power rating) and other algorithms, Zoltar is much too strongly biased towards Vegas underdogs. I need to fix this.
For week #15:
1. Zoltar likes Vegas underdog Colts against the Patriots
2. Zoltar likes Vegas underdog Lions against the Cardinals
3. Zoltar likes Vegas underdog Giants against the Cowboys
4. Zoltar likes Vegas underdog Broncos against the Bengals
5. Zoltar likes Vegas underdog Saints against the Buccaneers
6. Zoltar likes Vegas underdog Bears against the Vikings
For example, a bet on the underdog Colts against the Patriots will pay off if the Colts win by any score, or if the favored Patriots win but by less than the point spread of 2.5 points (in other words, by 2 points or less).
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 #14, against the Vegas point spread, Zoltar went 0-3 (using the conservative 4.0 points as the advice threshold. Overall, for the season, Zoltar is 48-42 against the spread (~53%).
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 #14, just predicting the winning team, Zoltar went 10-4 which is pretty good.
In week #14, just predicting the winning team, Vegas — “the wisdom of the crowd” also went 10-4, the same as Zoltar.
Zoltar sometimes predicts a 0-point margin of victory, which means the two teams are evenly matched. There are five such games in week #15. 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.
Left: This is a photo of the back of a Frosted Flakes cereal box from 1967. It had an interesting football game where you’d spin a spinner (or roll two dice) and use the chart to determine the output. An early simulation. A generation of American computer scientists were inspired by clever games such as this. Right: Cereal boxes have activities that target girls as well as boys. This is a box of Unicorn Cereal that has an art activity for girls.