We are two weeks into the NFL season, and everyone is trying to figure out which teams are for real and which have gotten lucky. With only two games, it’s incredibly hard to know a team’s true talent level. And even after the full season, it’s not always clear how good a team is, especially the average ones that make up the NFL’s middle class.
What makes inference on the strength of teams particularly difficult, besides the extraordinarily small sample size, is that there’s also the confounding variable of home-field advantage that one has to control for. Although home-field advantage has been estimated to be worth three points to the spread, that is 1. Only an estimate, and 2. Undoubtedly variable from year to year and team to team.
Towards the end of the season, controlling for the effect of home-field in past games when trying to estimate a team’s strength for future games may not be such a large issue. By then, teams have more or less played the same number of home and away games. However, early on in the season, particularly in Week 3 games, there are cases where a team may have played all of its games at home - or equally, all on the road.
This unequal distribution of home and away games could cause issues for predicting outcomes in two ways. First, it may directly affect teams – if you open up the season with two straight home games you may be more rested, as you have yet to travel and sleep on the road. Alternatively, it may affect bettors, as they are unable to properly gauge a team’s strength without having seen them play both at home and away.
Looking at the data (going back to 1985), we can see that there’s a clear and marked effect. Below is a table showing how teams that have played only home or only away games in the first two weeks perform against the spread in Week 3.
In both instances (only home and only away games) we have found statistically significant results. Teams that have only played home games are a good bet against the spread, while teams that have only played away games are a bad bet.
This fits into our theory that perhaps not having to travel during the first two weeks allows teams to be more rested and prepared, therefore they perform better. However, it also fits into the theory that perhaps bettors cannot properly evaluate teams with watching them play a home and away game.
Specifically, if we were to buy into this theory, that would suggest that they have an exaggerated view of home-field advantage early on, and that if a team performs well at home it may attribute its success to the fact that it was playing at home and not that it’s actually a very good team, thus not giving it fair lines in the future. Likewise, if a team struggles on the road it might blame those struggles on the fact that it was playing away in the first two weeks, glossing over the fact that the team may not be that good at all.
Can we attempt to figure out which of these two theories is correct? No analysis is ever going to truly answer this, but we can try.
Theoretically, if it was the two weeks of playing at home that make teams perform better, we can look at how teams that have two straight home games at any point in the season perform the next week.
If there’s a similar edge, that would suggest that players getting rest may be the primary cause. If there’s no edge, it would suggest that bettors not being able to properly evaluate teams without a home and away game may be the primary cause.
Of course, this is not a concrete analysis – it’s probably a combination of the two, in the first place, and second of all there are other factors (distance traveled, injuries, etc.) that we are not even considering. Still, it will be interesting to look at. As a reminder, these are all instances AFTER Week 3 where a team is playing after two straight home (or away) games.
--
Harrison Chase is the Co-President of the Harvard Sports Analysis Collective, a student-run organization at Harvard College dedicated to the quantitative analysis of sports strategy and management.
