Andrew Mooney is the Co-President of the Harvard Sports Analysis Collective. He is a senior majoring in Social Studies, which is another way of saying he's an economics major. Andrew has worked as an analytics intern in the NFL for about two years, and previously wrote for the Stats Driven blog at Boston.com. He's a big fan of all Detroit sports, and he'll throw an octopus on your ice if you're not watching.
Though I’m sure it provides players some much needed rest, it is not immediately clear what effect this time off has on performance. The qualitative cases for each side are pretty straightforward, and your grandfather used each of them liberally in instructing you in the wonders of sporting conventional wisdom. “Ah, they had an extra week to prepare AND get healthy,” he said knowingly after Washington’s 31-6 thrashing of the Eagles last season. “They just got rusty,” he told you after the Vikings fell to Chicago, 28-10, the following week. “What in the Sam Hill…” he muttered after the 49ers and Rams battled to a 24-24 tie.
Quantitatively, though, the extra time off hasn’t proven to be particularly beneficial in the case of bye weeks. In terms of winning percentage, teams generally perform right in line with expectations, without experiencing a significant positive or negative effect on performance after their week off.
But what about games scheduled on Thursdays? It’s possible that an extra two to three days off could indeed provide some much needed time to rest up and get some extra preparation without taking away from the weekly rhythm of the season.
Using Pro Football Reference, I collected the result of all 148 games featuring a team that had played the previous Thursday dating back to 2006, when Thursday Night Football as we now know it first debuted. As a proxy for how that team was expected to perform, I recorded the Vegas line for each game, also from PFR. If no such Thursday effect exists, we would expect the average difference between the expected outcomes (the point spread) and the observed outcomes (the actual margin of victory/defeat) to hover right around zero.
A quick glance at the density plot above suggests that there might be something to this hypothesis, as the distribution looks skewed ever so slightly to the right. The numbers underlying the graph back that up, showing that, on average, a team coming off a Thursday game actually beats the spread by 1.32 points. This is not an immaterial amount, especially considering that home field advantage is worth about 2.5 points, according to Jeff Sagarin. Put another way, we might say that a Thursday night game is worth an additional half-home game on the schedule (sort of).
A t-test reveals that the overall post-Thursday advantage is not statistically significant (p-value = 0.22), but with a sample of only around 150 games, we shouldn’t take that to immediately render the analysis invalid (a mistake I made in a post on Thursday nights last year, this one with injury rates). Expanding the study to include more years might reveal that the effect is in fact quite robust, and our initial suspicions were accurate.
I found another interesting item in the data when I broke down teams by whether their post-Thursday games were played at home or away. Assuming a uniform home field advantage factor is incorporated into the point spread, we would expect no difference in the performance of post-Thursday home and away teams in this sample against the spread. However, home teams enjoy an even larger post-Thursday advantage, beating the spread by an average of 2.24 points, as compared to 0.82 points for post-Thursday away teams. It could be the case that this difference would go away with a larger sample, which included 52 home games and 96 away games, but again, we’d need more information to get a clearer picture.
Obviously, NFL games were played on Thursdays well before 2006—the first Thanksgiving games were played in 1978—so this analysis could be bolstered by a larger sample of games, as alluded to before. I could also have used some other proxy for expected performance, such as the Simple Rating for a team and its opponent, to see whether my findings held up.
In looking at the games I did have in my sample, I thought that the point spread might be an inadequate way of measuring relative team quality for the Week 2 post-Thursday teams—the ones coming off the season-opening Thursday night game. At that point, even Vegas doesn’t have enough information about teams yet; something tells me that, by Week 5, the Broncos would be favored by slightly more than 4.5 points over the Giants, as they were in Week 2. Finally, it’s possible Vegas has already taken note of this phenomenon and adjusted its spreads accordingly, but that doesn’t mesh with these results.
Still, I am inclined to say that, in general, the extra days of rest following a Thursday game do afford a team some tangible advantage—though tread cautiously if you try to take advantage of your newfound knowledge in Week 6. Don’t blame me when the bank forecloses on your house; you bet on the Bills, and that’s asking for a lifetime of misfortune.