Football is unpredictable–even at the NFL level–for a lot of reasons. There’s all your cliches: it’s a game of inches, games aren’t played on paper, etc. There are match-up peculiarities, small sample sizes and just plain luck. Yet, sometimes the ‘unpredictability’ is because, especially early in the season, a lot of what we predict is based on projections tied to previous years. Now, every team has at least a quarter of its season played and several teams are 5 games in. So the 2017 season can really start to mean something in and of itself for projecting ahead.
One way to capitalize on this is to look at the schedule ahead for each team. Before the season, most strength of schedule metrics are measured by opponents’ prior year win percentage. That yielded this list:
5 weeks in, we can better use this season’s performances to judge remaining schedules. Football outsiders does this by averaging the DVOA (Defense-adjusted Value over Average) of teams remaining opponents. That gives these rankings:
Stats-based projections are nice, but I want to bring in speculation as well. I played out the rest of the season with this predictor (https://playoffpredictors.com/Football/NFL?L=Aw18ZXTt-DFOS1b0ckA). I used teams’ record in remaining games to build my own projections. This is not truly a strength of schedule metric because a good team might get wins against a hard schedule and vice versa, but its still an insight into what teams are facing. Here’s how it comes out:
As the season crystallizes you can see some teams face a much harder slate than expected (e.g. Colts), while some face an easier road than projected (e.g. Broncos). I also want to note that when I simulated the season for the chart above, I received the following playoff projections. I’d be shocked if they are exactly right, but they make sense in a lot of places based on early season returns.