Zoltar is my NFL football prediction computer program. It uses custom reinforcement learning and a neural network. Here is Zoltar’s prediction for the 2021 season Super Bowl LVI (week #22) to be played on Sunday, February 13, 2022.
Zoltar: rams by 5 dog = bengals Vegas: rams by 4.5
Update: The Rams did in fact beat the Bengals by a score of 23-20 so Zoltar’s prediction was pretty good.
I posted Zoltar’s prediction last week but I did not take the home field factor into account correctly. The Bengals are the virtual home team, but the game is being played on the Rams home field in Los Angeles. I had assumed the game would be at neutral site, not in L.A.
When I ran Zoltar last week, I set the home field advantage to 0.0 points to account for what I thought was a neutral site. The normal home field advantage is 3.0 points. For this Super Bowl, I decided to use a home field advantage of 2.0 points for the Rams. I used my intuition rather than hard data.
After adjusting for home field, Zoltar thinks the Los Angeles Rams are 5.0 points better than the Cincinnati Bengals. The Las Vegas point spread says the Rams are 4.5 points better than the Bengals — up half a point from last week’s 4.0 point spread. Because Zoltar agrees very closely with the Vegas line, Zoltar does not strongly recommend a bet. But if forced, Zoltar would say bet on the Rams.
Notice that if I use a 1.0 point home field advantage instead of 2.0 points, then Zoltar would think the Rams are 4.0 points better than the Bengals, and if forced, Zoltar would say bet on the Bengals.
A bet on the Rams will pay off only if the Rams win by more than 4.5 points (in other words, 5 points or more). A bet on the Bengals will pay off if the Bengals win by any score, or if the Rams win but by les than 4.5 points (i.e., 4 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 #21, against the Vegas point spread, Zoltar went 0-0 (using the standard 3.0 points difference as the advice threshold). Zoltar agreed closely with Vegas and so did not have any theoretical advice.
Overall, for the season, Zoltar is 66-53 against the spread (~55%).
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 #21, just predicting the winning team, Zoltar went 1-1, correctly liking the Rams over the 49ers but incorrectly liking the Chiefs over the Bengals.
In week #21, just predicting the winning team, Vegas — “the wisdom of the crowd” also went 1-1, the same as Zoltar.
Left: My system is named after the Zoltar fortune teller machine you can find in arcades. Coin-operated fortune teller machines have been made for well over 100 years. Center: The Oracle fortune teller machine was manufactured about 1925 by the Exhibit Supply Company. Right: The Uncle Sam’s Philosophy machine was made about 1890 (unknown company). Note: “Uncle Sam” as a symbol of the United States government came into existence about 1812 so it’s not clear why this machine with a Negro caricature was named as it was).