Can Quantitative Analysis Pick Super Bowl Winners?

Can Quantitative Analysis Pick Super Bowl Winners?Los Angeles-based Analytic Investors began using quantitative analysis to pick Super Bowl winners in 2004. From that year through 2013, the firm racked up a record that would impress any gambler and threaten to bankrupt many a small-time bookie: 9-1. But in each of the past two years, Analytic’s picks have come up short – will this be the year the firm returns to its winning ways?

First, a few words on Analytic’s methodology: The firm assumes that the NFL teams that generate the most “alpha” in the regular season will underperform in the playoffs, and vice-versa, due to the psychological effects of raised (or lowered) expectations and general mean-reversion. Betting on the team with the lower regular-season alpha to “beat the spread” in playoff games has been the right bet 63% of the time since 2004, including a 4-4-2 (wins-losses-ties) record in the games leading up to this year’s Super Bowl. But before we go any further, we need to examine the concept of “point spreads.”

Point Spreads Explained

When football bettors place wagers, they typically do so according to a “point spread.” If the Giants and Jets are two unevenly matched teams, then odds-makers will estimate the projected margin of victory for the greater team over the lesser, establishing a baseline “point spread.” For instance, if the Giants were predicted to win the game by 7 points, then the baseline spread would be 7 points – and this would mean bettors on the Giants would need the Giants to win by more than 7 points in order to win their bets. By contrast, bettors on the Jets could win even if the Jets lost the game – so long as they did so by less than 7 points.

Point spreads move like market prices in response to demand from bettors. If too many people bet on the Giants with the 7-point handicap, then the odds-makers will gradually raise the point spread to achieve equilibrium. If too many people bet on the Jets, then they will reduce the point spread (or potentially even invert it). If circumstances change prior to kickoff – i.e., if a key player is sidelined with injury – then point spreads can change quickly.

Analytic’s NFL Alpha

So imagine a Bears fan bet $1 on his team to “beat the spread” in each of the Bears’ 16 regular season games. Although the team lost almost twice as many games (10) as it won (6) in 2015, the Bears were the point-spread underdogs in all but two of their 16 games, and they were able to beat the spread more than half the time to earn consistent bettors a 23.4% return on investment (“ROI”).

NFL Alpha Analytic Investors

Super Bowl 50

But Super Bowl 50 doesn’t involve the Bears, the Jets, or the Giants – it matches the Carolina Panthers vs. the 6-point underdog Denver Broncos. These are two of the three teams that generated the most alpha during the regular season, with the Panthers going 15-1 and earning 61.5% ROI for consistent bettors, while the Broncos went 12-4 and generated a 31.9% ROI.

According to Analytic Investors’ methodology, this means you should put your money on Denver to lose by less than six (or outright win), but a word of caution: Two years ago, Analytic also picked Peyton Manning’s Broncos, who were just two-point underdogs, and they went down to an embarrassing 43-8 defeat. A third straight Super Bowl loss for Analytic Investors would certainly be unflattering to their impressive track record, but the 63% winning percentage for the past 12 playoffs continues to make the case for the soundness of their alpha-based betting strategy.

To study up before Sunday’s big game, download Analytic’s NFL Alphas 2015-16.

Jason Seagraves contributed to this article.

Add a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.