Hit the Mark
The NFL may be a league of parity but in any given game two teams are rarely evenly matched. Even if it’s only an edge through home field advantage, sharp minds can almost always agree on which team should be considered the favorite. Opinions quickly differ, however, when we try to determine the degree to which one team is superior to another. This is where Vegas comes into play; oddsmakers produce their own projection to settle such disputes. And the business responsible for taking all of our action is pretty good at what they do.
After all, if Sportsbooks were inept at posting lines (and also adjusting for the money-backed opinions they receive), far greater than just a select minority would be able to consistently find success betting professional football. This led me to thinking and subsequently asking one question: exactly how accurate is Vegas at predicting the outcomes of NFL games? Dating back to 1989 (a sample size of 5,996 games), the final score has landed within +/- 3 points of the point spread 21.2% (1,272 games) of the time. Below is a chart that illustrates the relationship between the point spread and the final score of games over the past 23 seasons.
|
Final score margin from point spread |
Percentage of games |
|
+/- 1 |
9.9% |
|
2 |
15.4% |
|
3 |
21.2% |
|
4 |
27.4% |
|
5 |
32.7% |
|
6 |
38.9% |
|
7 |
44.6% |
While these results should curb the belief that oddsmakers are omniscient and “always get it right,” the fact that the final score falls within a field goal of the consensus line in 1 out of every 5 games is still beyond impressive. To put it into context, in a given week when there are 16 games on the board, the spread will be just about dead on in roughly three of those matchups. That leaves us with around 13 games that we know will deviate from the spread. Value is undoubtedly present, but identifying which select sides’ present opportunities in an already limited collection of matchups can be consistently challenging.
So now you’re led to a question of your own: what’s the takeaway here? The data supports that the spread is far from perfectly accurate at predicting the final score of games. Again, it’s important to remember here the spread isn’t designed to predict the future but rather force bettors into making a difficult decision. There are good bets to be made week in and week out but uncovering that value is easier said than done. Bettors should be selective in the games they play, trying to isolate wagering opportunities whenever possible. If you are one that bets five games each week you are essentially conceding the fate of one of your wagers to a coin flip, think about that for a second. Also, always strive to get the best of the number; in the long run, every half point counts especially around your key numbers. For example, savvy NFL investors should know that the final spread lands on 3 15.8% of the time and on 7 in 8.8% of games.
Do point spreads get more accurate as the season goes on?
Many feel that as the NFL season progresses, lines get tighter and value diminishes. The data, however, tells a different story. I analyzed both the final 8 weeks and final 4 weeks of the season and the variability of outcomes relative to the spread stays constant as we approach Christmas time and the playoffs. Whether it is week 2 or week 15, around 20% of games fall within a field goal of the line. While we may perceive our opinions of teams and evaluations of matchups as more concrete, in reality, the accuracy of our collective predictions (the consensus line) does not improve as the season goes on. The point is, wagering opportunities remain uniformly present throughout the season and it is incumbent on the individual handicapper to identify the performance factors that do not. Fluctuations in effort/motivation, matchup advantages, revenge, and scheduling angles still remain the handicapper’s best friend and should be used to gain an edge over the book whenever possible.
