×

Quantitative models pick a Super Bowl winner

There's no such thing as a "sure thing" when it comes to market returns, or sports betting, but one money manager is trying to develop a system to help beat the odds … at least at picking a Super Bowl winner.

Analytic Investors is a Los Angeles-based advisor that specializes in quantitative investing strategies. In other words, the firm looks to develop systems or techniques that attempt to generate superior returns. They've taken that same framework and philosophy and applied it to predicting the winner in NFL games.

This time around, the team's handicapping system has chosen the Denver Broncos to win the Super Bowl, and here's why. It's all about performance expectations, and which teams typically outperform or underperform the overall "market" for the NFL.

Peyton Manning of the Denver Broncos, left, and Cam Newton of the Carolina Panthers
Getty Images
Peyton Manning of the Denver Broncos, left, and Cam Newton of the Carolina Panthers

Each week, teams that play each other simply win, lose or tie their matchups depending on how many points they score. But when picking winners depends on using point spread to handicap matchups, winning and losing becomes a matter of determining whether expectations are set too low, or too high, much like picking stocks in the market.

Analytic Investors tracks how each team performs each week against the spread. To be considered a "win," a favorite must win by more than the point spread, and an underdog must lose by less than the spread. The teams that beat the spread the most throughout the season are considered to have "outperformed" the overall market, something referred to as "alpha." That same term is used in financial markets to quantify how much a money manager has outperformed a specific performance benchmark. Think of it like how much a mutual fund or hedge fund manager beats the performance of the S&P 500.

Read MoreHow to advertise during Super Bowl 50 without a TV ad

Those who used regular season performance as a way to predict who would ultimately play in the Super Bowl weren't all that surprised to see the Carolina Panthers and Denver Broncos. According to the firm's calculations, Carolina had the best outperformance score, and Denver had the third best. The Minnesota Vikings came in with the second best (they lost to the Seattle Seahawks during the first round of the playoffs 10-9 after a missed field goal in the game's final minute).

Regular season performance is used as a basis for picking winners during the postseason, but in a more contrarian twist, the firm's system looks to bet on underperformers in the postseason. According to Chris Lardieri, one of the study's authors, the thinking is that there will be a tendency to overvalue regular season outperformers, and undervalue underperformers. In financial terms, the relatively less expensive assets will outperform. Think of this as picking the "value" play.

The methodology led to a 4-4-2 record in picking teams during the NFL playoffs this year. While that's nothing to write home about, in year's past, the system has averaged an approximately 63 percent success rate in picking winners against the point spread. During the last 12 Super Bowls, picking the lower alpha team has been a winning strategy in nine games, or a 75 percent success rate. The system also correctly picked the first eight winners, starting with Super Bowl XXXVIII in February 2004. However, it's worth noting that the system has not worked last two Super Bowls, and has failed in three of the last four.

Read MoreMarket does best when this team wins Super Bowl

This year, the relative value pick happens to be the Denver Broncos. The Carolina Panthers are currently favored in Super Bowl 50 by 6 points. In order for the Analytic Investors system to log a successful observation, the Broncos will have to lose by less than that spread, or win outright. The last time the Broncos were the firm's pick to win? Super Bowl XLVIII, when they lost to the Seattle Seahawks 43-8. You can bet Broncos fans will be expecting a different outcome this time around.