Week 1 Football Betting: Exploiting the Biggest Mispricing of the Week
Minnesota Vikings @ San Francisco 49ers +3.0 (-125)
Levi's Stadium, Santa Clara, CA
September 14, 2015, 7:20pm PST
Our work highlights the Minnesota Vikings at the San Francisco 49ers +3 as the marquee matchup of the NFL week 1 schedule. Yes, you read that right, we are referring to a contest between the #23 and #24 teams as marquee. To clarify, as we are disciplined bettors first and sports fans second, rather than by excitement or star appeal, we define marquee in this context on the basis of probability of win and/or expected return. By these measures, marquee is an absolutely fitting descriptor for the second game of Monday's scheduled two-fer. We expound below.
After calibrating for San Francisco's home field advantage, our power rankings methodology, powered by ESPN's NFL FPI statistic, has the Niners winning outright by 2 points, suggesting that sportsbooks are off by a massive five points (for comparative purposes, for all week 1 games, the median difference between the betting line and our FPI-based spread estimate is -0.3; the average is -0.6).
Our two independent simulation routines likewise forecast a straight-up San Francisco victory (by 1-3 points), signaling 4-6 points of value with the home dog.
Further (as if huge, across the board differences in our estimation of the line and the actual spread were not enough), public opinion is extreme in favoring Minnesota to cover. As of the time of this writing, SportsInsights reports that 74% of the wagers they monitor back the Vikings. Cappers Mall observes similar herding, with 72% of tracked bets following Minnesota.
In a prior note, we referenced a report from OddsShark in which the authors review eight NFL seasons and conclude that "NUMBERS DON’T LIE – FADING THE PUBLIC IS PROFITABLE". SportsInsights goes a step further, and not only identifies the profitability of fading the public, but explores the psychology that underlies this phenomenon. The writers there hypothesize that human beings love to root for winners, which translates to irrationally betting on favorites -- despite the influence of the spread as a leveling factor.
More robust empirical studies too support the idea that when public opinion is extremely synchronized, it is wrong more often than not. Paul and Weinbach (2010), for example, find that betting the spread against public sentiment is statistically profitable in the NFL. As such, we regard the lopsided Vikings bias prevalent among bettors as validation of our contrarian view.
Given that closing spreads have been demonstrated to estimate actual outcomes (we reference several academic studies here that support this conclusion), we could point to the colossal move in the opening spread (from San Francisco -3.5 to +3.0) as casting doubt on our call. To explore this possibility, we perused TeamRankings database of opening and closing lines for 1,865 NFL games since 2008. Interestingly, there are only four observed instances of moves in the betting line of 6 points or more . The opening favorite is 2-2 from this spot. Thus while the sample size is too small to lend itself to definitive conclusion, historically, this set up has not been damning. Further, given that the opening spread is five months old (many Las Vegas sportsbooks began introduced opening lines for 2015 NFL games back in April), and in light of the unanimous support for the 49ers to cover from the analytical factors in our stable, we take the warning from the changing line with a grain of salt.
Bottom Line: The spread change might be interpreted as a yellow flag, but it might also be nitpicking. Given the unanimous and very strong indications from all of the quantitative factors we analyze, we are quite comfortable betting San Francisco +3.0. While our work does uniformly call for the Niners to pull off the upset from this position, at +115, we don't think the moneyline offers adequate compensation to warrant giving up the 3-point buffer. We recommend the home, dog versus the number.