NFL Betting Angles Fact or Fiction #7: Motivation

Many bettors like to take motivation into account when making their NFL and college football plays. The most common angle involves betting on a team that is still in playoff contention against a team who has been eliminated from playoff contention. The theory says that the team who still has the possibility of making the playoffs will be more motivated to play well, and may outperform the point spread which is based solely on past stats and injuries.

Before we get into the numbers, I think it is always worth a subjective discussion about whether this type of angle is likely to be real. Any statistical test we can do on these angles comes from some form of data analysis versus some form of prior expectation. In the most common test we build our test against the null prior, that our angle has no football impact. But if we know that our angle should have an impact beforehand through some sort of prior, football-knowledge based assumption, we may be able to make conclusions with slightly less data. However, the only way to really do this is to decide our prior before we see any results; otherwise we will be biased by our knowledge of the numbers.

Personally, I am on the fence about this angle. My default view is that players at the NFL level almost always play at or near their top performance level, barring injury. It is extremely hard to make it to the NFL, and players who make it to such a high level are already likely to be incredibly motivated individuals. While the playoffs and chance to play for a title probably does lead to some added motivation, I think most players are motivated primarily by money and their desire to be the best, neither of which really involve titles. Even if the games no longer matter there is always the motivation of knowing that their play will likely be on film and their statistics counted, both of which could lead to a bigger contract.

Still, one area where motivation could play a major role is injuries. With the season over, players with moderate injuries are probably more likely to sit out than they otherwise would be, especially players in a contract year. But even here, I am not completely convinced that this will actually detract from team performance. My theory, which has served me well in sports betting, is that if a player is questionable to start, it doesn't really matter much if they play anyway and one is best off either passing on the game or only betting the game if you are happy betting his team even if he doesn't play. Especially at the NFL level, I think most players are so close in value that it is extremely unlikely an injured player is really better than the alternative, barring "safety" injuries like concussions.

In any case, my view is that this angle is probably over-rated, and I expect little to no impact, but I would still be open to a slight impact if that is what the data shows. Either way, it is worth digging into the numbers to see if I am right, and if the market values it. If motivation really is over-rated, it could be an angle worth fading.

In this study I have gone back and manually looked at the standings in the last three weeks of the season and assigned a status of "Out" to all teams who were out of playoff contention at the time the game was played, and a status of "Locked" to teams who had made the playoffs and had nothing to play for. I have only focused on the last three weeks of NFL seasons from 2012 to 2017; I chose the last three weeks because it is very rare for a team to be mathematically out of contention before week 15. As a service to sports bettors I have attached a table of which teams had which status at the end of the article.

We start by looking at the change in performance of teams who are out of playoff contention. We use our favorite points for/against + quarterback yards per attempt model to test for this, with an added variable being the difference between the "out of contention" status. We will assign a 1 if the home team is out of contention, a 0 if both or neither teams are out of contention, and a -1 if the road team only is out of contention. Finally, we will look at both yards per play and scoring margin as part of this model, to try and filter out some of the variance in scoring margin for what is a small sample of relevant games.

Playoff contention status causes no significant difference in yards per play - the coefficient of -0.06 suggests that a home team out of the playoffs will only average -0.06 yards per play worse than they would if they were in the playoffs. Since one yard per play over the course of a typical game is worth about five points, this translates to only about -0.30 points per game. But these same teams wind up being outscored by an additional -1.49 points per game, an impact that is not statistically significant but only had about a 21% probability of occurring by chance. What makes up the difference is that teams that were out of playoff contention averaged an additional 0.35 turnovers per game, which believe it or not, only had about a 4% chance of occurring due to chance alone in these games.

We can run the same model as above on the market's point spread, rather than the actual result of the game. We find that the market does value motivation - after adjusting for the same variables above, teams that are still in playoff contention are favored by an additional 0.9 points on average compared to those that are not. Given that the 1.49 point impact we modeled over these same games is higher than the 0.9 point market estimate we would expect the motivation angle to be a (small) winner, and it was. Taking teams who remained in playoff contention versus those who were out of playoff contention resulted in a small win of 2.1 units in 124 games versus Pinnacle closing lines.

Overall, while yards per play suggests there is not much behind the playoff motivation angle, the fact that the actual results of games across this sample are pretty much in-line with the market means we should probably include it as part of our handicapping. The handicapping impact is probably near the market estimation of 0.9 points, but may be slightly higher. Rather than being a dominant factor, motivation is the type of factor that might swing us for or against those marginal games where we think we may have a small edge.

One problem with the basic model above is that it doesn't account for injuries. One might expect that players with minor injuries will sit out the last few games of the season, which, if we are accounting for injuries already, might lead to us double-counting the influence of motivation. A decent way to test for this is to measure the average Madden ratings of players who played in the game.

Surprisingly, teams who are out of the playoff hunt actually have higher average Madden ratings than their opposition after adjusting for past team strength. In addition, if one includes Madden ratings in our simple model, the model suggests an even higher impact of playoff-eliminated status, closer to -2 points. If one makes no adjustments, playoff-eliminated teams have lower Madden ratings, but we would expect that either way as playoff-eliminated teams tend to be worse teams. We can't be completely sure, but it doesn't appear that teams that are out of the hunt are resting marginally injured players just because the season is over. This would suggest that teams who are out of contention really do play slightly worse due to lack of motivation and not other factors. The difference resulting from turnover margin, a notoriously hard variable to predict, may indicate a lack of focus in these games.

Verdict: Fact (motivation is real), Fiction (betting motivation blind can win money)

As a bonus, we can also take a look at those teams who have clinched home-field advantage or otherwise have nothing to play for. These situations are a little different from those teams who have simply fallen out of contention, since teams who have clinched almost always rest players and often will also tend to hold back on their game plans. Games with teams that have truly nothing to play for are very rare, with only 11 such games in my estimation since 2012. Using the same approach we used for teams who were out of playoff contention, these teams averaged 0.71 yards per play (a statistically significant impact even in just 11 games) / 2.9 points per game worse than that what we'd expect, and the market priced these teams as 3 points per game worse. So there is no edge to be had, but it is important to avoid betting on these teams unless the odds are extremely favorable.