Zoltar is my NFL prediction computer program. It uses a deep neural network and reinforcement learning. The results for the first seven weeks were completely crazy. The covid pandemic has had a massive effect on the NFL.

I have switched Zoltar into experimental mode last week. I fiddled with five of Zoltar’s parameters:

Base rating pts for a win: 16 to 14
Rout definition: win by 20+ to win by 22+
Rout rating pts: 4 to 2
Underdog away win bonus pts: 4 to 2
Home field advantage: 3 pts to 2 pts

These changes made Zoltar more conservative, and surprisingly (to me) Zoltar’s predictions were extremely close to the Las Vegas point spread data.

The most interesting parameter is the home field advantage. For decades, both humans and Zoltar have given 3 points to the home team. But with covid, you can argue that the home field advantage is dramatically greater (say 5 or 6 points or more), or dramatically smaller (say 0 or 1 point), or even negative. I won’t know until after the disastrous 2020 season ends and I have enough data.

Using the experimental parameters, here are Zoltar’s predictions for week #9 of the 2020 NFL season:

Zoltar:     packers  by    1  dog = fortyniners    Vegas:     packers  by  2.5
Zoltar:    redskins  by    4  dog =      giants    Vegas:    redskins  by  3.5
Zoltar:     broncos  by    1  dog =     falcons    Vegas:     falcons  by    4
Zoltar:      titans  by    1  dog =       bears    Vegas:      titans  by  5.5
Zoltar:      texans  by    1  dog =     jaguars    Vegas:      texans  by  6.5
Zoltar:      ravens  by    1  dog =       colts    Vegas:      ravens  by  3.5
Zoltar:       bills  by    3  dog =    seahawks    Vegas:    seahawks  by    3
Zoltar:     vikings  by    5  dog =       lions    Vegas:     vikings  by    4
Zoltar:      chiefs  by    9  dog =    panthers    Vegas:      chiefs  by 10.5
Zoltar:     raiders  by    1  dog =    chargers    Vegas:    chargers  by  1.5
Zoltar:    steelers  by    5  dog =     cowboys    Vegas:    steelers  by   14
Zoltar:   cardinals  by    3  dog =    dolphins    Vegas:   cardinals  by  4.5
Zoltar:      saints  by    1  dog =  buccaneers    Vegas:  buccaneers  by  4.5
Zoltar:    patriots  by    1  dog =        jets    Vegas:    patriots  by  7.5

Old Zoltar theoretically used to suggest betting when the Vegas line is more than 3.0 points different from Zoltar’s prediction. It’s not clear how that advice parameter will change.

When you bet on an underdog your, bet pays off if the underdog wins by any score, or if the game is a tie, or if the favorite team wins but by more than the Vegas point spread. If the favorite team wins by exactly the point spread, the bet is a push. You lose your bet if the favorite wins by more than the Vegas point spread.

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.

Experimental Zoltar likes five underdogs in week #9:

1. Underdog Broncos vs. Falcons
2. Underdog Bears vs. Titans
3. Underdog Jaguars vs. Texans
4. Underdog Bills vs. Seahawks
5. Underdog Saints vs. Buccaneers

Zoltar also likes underdogs Cowboys vs. Steelers, and underdog Jets vs. Patriots but those games are off Zoltar’s scale.

Last week, experimental Zoltar’s predictions were good and he went 2-0 against the spreads, correctly liking underdogs Denver Broncos and NY Giants. (Zoltar also liked the underdog NY Jets but only because the game was literally off the scale.)

Just predicting the winning team, Zoltar was a terrible 8-6, but the Vegas point spread was even worse at 7-7. The result for Vegas is one of the worst I’ve ever seen and indicates how severely impacted the 2020 season is.

Left: “The Oracle” (1880), Camillo Miola. Center: “Question to the Cards” (1889), Edouard Bisson. Right: “The Crystal Ball” (1902), John William Waterhouse.