NFL Power Index: Final 2016 Rankings
The Seattle Seahawks finished 9th overall in the final NFL Power Index rankings for 2016. Their playoff opponent for this week finished 24th.
The 2016 regular season is complete, which means it is time for the final NFL Power Index rankings of the year. Overall there wasn’t much change from last week, which should be expected in a stats-based metric. Large sample sizes means it is tough to move averages.
The Falcon’s wrapped up a great season as the No. 1 team. The Patriots closed the gap over the final couple of weeks, but Atlanta would have needed to collapse in Week 17 to be caught. That simply didn’t happen.
Cleveland put together a strong final 2 weeks to avoid being the first time to ever finish the season in single digits. They’ll still have the first overall pick in the draft though.
The Seattle Seahawks finished 9th overall. They held that same ranking for the final three weeks of the season. Seattle’s offensive metrics never improved enough for them to be ranked any higher.
The NFL Power Index never liked Seattle’s wild-card playoff opponent. The Lions spent the entire season ranked in the low 20’s, well outside the range of a typical playoff team. They finished 24th, suggesting the Seahawks should win easily on Saturday.
The reasons that the NFL Power Index doesn’t like the Lions has a lot to do with their lack of balance on offense. They cannot run the ball, and rely very heavily on Matt Stafford to carry the team. They also aren’t good on 3rd down on either side of the ball.
Here are the final NFL Power Index rankings for 2016:
Over 30 NFL teams stats are combined, first into sub-metrics and then together, to form the NFL Power Index. Strength of Schedule is provided by TeamRankings.
As I said last week, this will be the final edition of the NFL Power Index in its current form. It’ll be getting an overhaul in the offseason.
The game is evolving before our eyes. Running no longer correlated to winning at a high rate like it used to. The opposite is true for third down conversion rates. The correlation coefficients have shifted dramatically over the past few years, and my current model has reached its limited to account for that.
Luckily, thanks to the awesomeness of math, I’ll be able to normalize the new results so they look like the current ones. 100 will still be a theoretical maximum and so forth. Good times till be had by all.