Do NFL teams perform differently after a bye-week?
As the Randy Moss trade reverberated through Twitter and the Blogosphere, the Fantasy Football owner in me couldn’t have been more excited. Why? Not because Randy Moss will actually put up more points in Minnesota (in fact, he probably won’t) but because he avoided the dreaded “Bye-week.”
Minnesota took its bye week in Week 4, and New England is scheduled for its this weekend. The trade causes Randy Moss to avoid sitting out a week and thus will play all 17 weeks this season, a boon to fantasy owners. No, Kenny Britt will not make his fantasy debut for my team.
If I had a good data set on trades, I would definitely want to know how many other players has this happened to. However, I don’t.
Instead, I wondered if there is any performance difference after a bye week for NFL teams? Anecdotally, I haven’t witnessed any impact. I remember old Redskins and Eagles seasons where I thought they were stone-cold lead-pipe locks the week after the bye, only to see the teams go down in flames.
I looked at data over the last four years to see if I could detect a difference. Below is a graph of win percentage for all NFL teams the week before and the week after their bye week. Notice that the pre-week is never above the post-week.
But is the difference statistically significant or is it just random chance? I looked at the post-bye-week average and added 95% confidence intervals. Indeed, the pre-bye week is below the 95% confidence interval.
So, there is a difference, but it’s pretty small.
Complicating the data is that sometimes teams played each other before and after the bye-week, causing the sample set to regress towards 50%. I didn’t remove those instances.
Another interesting observation is that in 2006, there were two instances of a team going into a bye week playing a team coming off a bye week. In both instances, the team coming off the bye-week won.
Follow-ups: This analysis leaves me with plenty of follow-up questions, which I hope to examine in the future:
- Is there a difference in scoring margin? It appeared that pre-bye-week-losses had either very tight margins or were blow-outs.
- Will the conclusion change with more data?
- What’s the impact in other sports? I remember reading an article examining NBA win percentage by previous days off (I’ll have to find the link to that one)? What about College Football & MLB?
- What will happen when the NFL adopts a 2-bye-week-per-season model next year? Will the findings hold?