NFL bettors have a myriad of options to choose from, as the number of prop bets is massive. Still, the main wagers revolve around the outcome of the game itself: who will win, who will cover, and will the game go Over the total?
If you are reading this, you’re probably going to put some money on one of those bets, and thus it might help to have a little statistical briefing on the history of those kinds of wagers in past Super Bowls.
A technique called smoothing splines will be used to show trends in the data more clearly. This technique creates a smooth line through a noisy dataset, which is especially helpful when looking at sports data, as there’s often a ton of variability and it can be hard to see any trends without a smoother. For an example, below is a plot of the number of points scored in the Super Bowl over the years.
There appears to be an upward trend, but there’s an incredible amount of variation. We can add a smooth curve through the data to see the trend more clearly.
Now the trend is much more evident, and you can clearly see an increase in the 1980’s and then a slight downturn around 2000.
Of course, visualizing total points scored isn’t the most helpful thing to do if you are trying to win some money. What is of more concern would be whether the game goes Over the total or not.
We can do the same thing, now trying to visualize this outcome, which is a binary result (either it goes Over or it doesn’t – all pushes were discarded) and therefor is encoded as 0s and 1s. In this case, 1 represents a game that goes Over the total, and thus a higher probability indicates the game is more likely to go Over. We’ve also included a blue line that represents the 50 percent line or the place where the game going over and under would be equally likely.
As you can see, except for a brief period right around 2010, it has been favorable for the last 30 or so years to take the Over, although not incredibly so.
So there isn’t an incredible amount of edge available on betting either the Over or the Under. But what about betting on the actual game?
We can look at how often the underdog covers against the spread by doing the same thing but instead with a binary indicator for whether the underdog covered ATS.
There is a clear tread here. Ever since 2000, it has been profitable to bet on the underdog to cover. Only four times since 2000 has the Super Bowl underdog not cashed in, and in recent years the dog has been especially good. What is even more remarkable is that they have also been a good bet to win the game straight up.
Doing the same thing but looking at when the underdog has won the game, we can see that the last time a betting favorite won the Super Bowl was 2011.
Why has this been the case? One might suspect that the lines have just been getting more and more off, but we can see that is not the case. Looking at the difference between the line and the actual outcome, there is no noticeable increase in how much the line is off by.
There were games in the 1990’s that were just as far off as the line from two Super Bowls ago, where Denver was favored by three points but ended up losing by 35. Those games were just massively off in both directions. For example, the 49ers crushed the Broncos by 45 points in 1990, so even though they were favored by 12, the line was still off by 33 points.
The fact the lines do not appear to be getting much, if any, worse should dissuade you a little from betting on the underdog to win outright.
It appears that these big misses have been going in underdogs favor more than not recently, but the lines themselves aren’t getting any worse. What is actually happening is that the lines are just getting tighter. Here is a plot of how much the favorites have been favored by (not including this year):
As you can see, the lines have been as tight (close to zero) recently as they have ever been. This has allowed for an unprecedented number of underdogs to win outright despite the lines being as accurate as ever. This trend of close lines continues into this year, with the Patriots being 3-point favorites against the Falcons.
Hopefully this information not only helps with your betting this weekend (betting the underdog to cover seems to be the profitable trend) but also helps to caution you on why betting the underdog to win straight up might not be the best idea (no evidence the lines are getting worse).
Harrison Chase is the Co-President of the Harvard Sports Analysis Collective, a student-run organization at Harvard College dedicated to the quantitative analysis of sports strategy and management.