Zoltar is my NFL football machine learning prediction system. Here are Zoltar’s predictions for week #7 of the 2017 NFL season:
Zoltar: chiefs by 1 dog = raiders Vegas: chiefs by 2.5 Zoltar: bills by 4 dog = buccaneers Vegas: bills by 3 Zoltar: steelers by 8 dog = bengals Vegas: steelers by 6 Zoltar: titans by 7 dog = browns Vegas: titans by 6.5 Zoltar: panthers by 0 dog = bears Vegas: panthers by 3.5 Zoltar: dolphins by 6 dog = jets Vegas: dolphins by 3 Zoltar: cardinals by 0 dog = rams Vegas: rams by 3.5 Zoltar: vikings by 6 dog = ravens Vegas: vikings by 4.5 Zoltar: packers by 6 dog = saints Vegas: saints by 4 Zoltar: colts by 6 dog = jaguars Vegas: jaguars by 3 Zoltar: cowboys by 11 dog = fortyniners Vegas: cowboys by 6 Zoltar: seahawks by 0 dog = giants Vegas: seahawks by 6 Zoltar: broncos by 0 dog = chargers Vegas: broncos by 2 Zoltar: patriots by 6 dog = falcons Vegas: patriots by 3.5 Zoltar: eagles by 6 dog = redskins Vegas: eagles by 5
Zoltar theoretically suggests betting when the Vegas line is more than 3.0 points different from Zoltar’s prediction. For week #7 Zoltar has five hypothetical suggestions. As in recent weeks, most (four) of the five are on underdogs, and just one suggestion is for a favorite. It seems like Zoltar-2017 has a bias for underdogs, but his accuracy is good, so I’m not sure if Zoltar has the bias or if Vegas has a bias towards favorites.
1. Zoltar likes the Vegas underdog Bears against the Panthers. Vegas believes the Panthers are 3.5 points better than the Bears, but Zoltar thinks the two teams are evenly matched. A bet on the Bears would pay you if the Bears win (by any score), or the Panthers win, but by 3 points or fewer.
2. Zoltar likes the Vegas underdog Packers against the Saints. Vegas has the Saints as 4.0 point favorites but Zoltar thinks the Packers are 6 points better than the Saints. This difference must be due to the injury to the Packers quarterback. Often, Vegas overreacts to a key injury, but in this case the pessimism is probably justified.
3. Zoltar likes the Vegas underdog Colts against the Jaguars. Vegas believes the Jaguars are 3.0 points better than the Colts, but Zoltar thinks the Colts are 6 points better than the Jaguars. This big difference in opinion is rather unusual.
4. Zoltar likes the Vegas favorite Cowboys against the 49ers. Vegas has the Cowboys as 6.0 points better than the 49ers but Zoltar thinks the Cowboys are 11 points better. So, a bet on the Cowboys will only pay you if the Cowboys win by 7 points or more. (If the Cowboys win by exactly 6 points, the bet is called off).
5. Zoltar likes the Vegas underdog Giants against the Seahawks. Vegas likes the Seahawks as 6.0 point favorites, but Zoltar believes that the two teams are evenly matched.
Note: There’s some weirdness with the Rams vs. Cardinals game — my output says the Vegas line is a pick ’em but my data file says Vegas favors the Rams by 3.5 points. I’ll have to walk through Zoltar’s code to find the problem — almost certainly related to the fact that the game is being played at a neutral site (London).
Week #6 was very unusual. Against the Vegas spread, which is what Zoltar is designed to do, Zoltar went a nice 4-2. Zoltar correctly liked Vegas underdogs Dolphins, Jets, Giants, and Vikings, but incorrectly liked Vegas underdog Lions and Vegas favorite Raiders.
For the 2017 season so far, against the Vegas point spread, Zoltar is a pretty good 17-8 (68% accuracy). If you must bet $110 to win $100 (typical in Vegas) then you must theoretically predict with 53% or better accuracy to make money, but realistically you must predict at 60% or better accuracy.
Just for fun, I also track how well Zoltar does when only predicting which team will win. This isn’t really useful except for parlay betting. For week #6, Zoltar was a hideously bad 4-10 just predicting winners.
For comparison purposes, I also track how well Bing and the Vegas line do when just predicting who will win. In week #6, Bing was slightly better (but still weak) than Zoltar at 5-9, and Vegas was also bad at 5-9 just predicting winners.
For the 2017 season so far, just predicting the winning team, Zoltar is 54-37 (59% accuracy), Bing is 50-41 (55% accuracy), and Vegas is 47-42 (53% accuracy). The Vegas accuracy number is the lowest I’ve seen in the last 10 years suggesting that most NFL teams are just about equal in strength.