NFL Week 10 Has Been Strangely Unpredictable For 30+ Seasons

November 5, 2025

I've been working on NFL Elo models for the past few years, and noticed week 10 seems to have an awful lot of upsets.

 

For some background, I calculate week-by-week Elo ratings for each team based on only a few data points including game score, game venue (home/away/neutral field), rest days, etc. Based on team Elo ratings and those factors, my models are used to calculate each team's expected performance in the next game. After each game, team Elo ratings are updated based on actual vs. expected team performance. My theory is that only a few data points are needed to produce decent Elo ratings (no advanced metrics or stats of any kind) and I compare my models' pick rates (percentage of game winners correctly picked) against Vegas's straight up pick rates as a benchmark.

As part of my work on this project this past offseason, I calibrated my Elo models against NFL games from 1994 and later. I picked 1994 because it featured the addition of the two-point conversion, but that was also near the introduction of free agency (1993) and the salary cap (1994).

Here, I've plotted my models' pick rates over the entire Super Bowl era. I've included Vegas's straight-up pick rates as well, from sportsoddshistory.com.

Week 10 looks pretty inaccurate for my models, but I'm not too surprised because, again, they're calibrated against all games from 1994-2024, and none from 1966-1993. We see that Vegas however doesn't have much of a dip in their average pick rate for week 10.

If we look at only the past 31 seasons from 1994-2024, we see my models have an even worse accuracy in week 10. That's a bit surprising because my models were calibrated against these seasons, but what's more surprising is that Vegas also has a poor pick rate for week 10:

What's been going on with week 10 for the past 31 seasons? Vegas averages an overall pick rate of 66.63%, but week 10 is Vegas's worst week by far over that span with a pick rate of only 60.37%. Week 11's pick rate is high, especially for Vegas, but doesn't seem like as much of an outlier as week 10.

For fun, here are some recent week 10 upsets:

I assume there's nothing intrinsically different about week 10 itself, and this is just a statistical anomaly. The trade deadline may have something to do with it, but it hasn't always been after week 9. I'd be interested to hear if anyone else has any idea what could be causing this.

There was a little discussion on this in my accompanying post on reddit.