## Adrian Peterson Was Declining Before His Injury

Adrian Peterson has a torn meniscus, which is expected to keep him out until at least December. Given Peterson’s age — he’s 31, with his half-birthday coming yesterday — it’s reasonable to start thinking about whether the end is near for Peterson.

In his last 8 games (excluding playoffs), he rushed 150 times for 529 yards (3.52) with 5 TDs, one fumble, and 26 first downs. Prior to this stretch, Peterson had averaged 159 carries, 786 yards (4.95), 6.5 TDs, 2.24 fumbles, and 37 first downs. Given his performance in the one playoff game during that stretch (23 for 45 with one fumble), including that would only make the numbers look worse.

I came up with a relatively simple adjusted rushing yards metric to measure running back performance (note that in the formula below, rushing TDs are actually worth 20 yards because all rushing TDs are also rushing first downs):

(rushing yards + rushing first downs * 9 + rushing TDs * 11 – fumbles *30) minus (rush attempts * 5)

This is basically a mix of rush efficiency and rush quantity: we take rushing yards, add some other information, and then provide a penalty for each attempt used. The graph below shows Peterson’s game-by-game results for his career, along with, in black, a trailing-8-game average:

Peterson led the NFL in rushing last year, but he did not finish the season strong and played poorly in the playoffs. Then, his first two games this year had been disappointing, too. Here’s a look at his last 18 regular season games — i.e., all of his games since the start of the 2015 season:

Player performance is volatile, as you can see in Peterson’s first chart. So it’s always tough to know whether an older player in a slump is declining due to age or just experiencing random variance (to say nothing of a player’s statistical decline due to a weaker supporting cast or a tougher slate of opponents). Still, Peterson has finished with a sub-2.00 yards per carry average just 8 times in his 127-game career, and three of those games were his last three games (including playoffs). Had Peterson not been injured, we would have been hearing questions about his productivity soon, and as early as this week. And while losing Peterson is a big blow on paper, it’s far from clear whether the Minnesota offense will actually be worse without the version of Peterson that was likely to show up over the rest of the season, anyway.

• sacramento gold miners

100 yard rushing games are a good indicator of RB effectiveness, and Peterson has just one in his last eight(counting his bad playoff game).

Really neat way to look at RB performance. Do you plan to do this with other backs? Or show comparisons between seasons/players/eras/etc?

• I probably should. Thanks, Brad.

• Richie

Would have been nice if you published this 2 weeks ago, before I took him with my first round pick in my fantasy league! I figured he might be able to squeeze out one more solid season this year.

FWIW, Football Outsiders looked at players with similarly bad starts to Peterson. One of the best comps was Frank Gore in 2011, and he went on to have 4 more 1,000+ yard seasons (though 3 years younger than Peterson).

• Sorry, Richie. Next time.

• Joaquin A

Your formula demonstrates the shortcomings of statistical analysis when applied to football. Peterson’s performances and decline are impossible to determine thanks to absolutely uninspired playcalling by Turner and absymal run blocking by the Vikings offensive line. If you actually watch the games you might see that Peterson is routinely met in the backfield by one or more defenders.

• True. However, for years Peterson’s performances and high status were impossible to verify thanks to the inspired play calling of his offensive coordinators and the great run blocking by the Vikings offensive line.

• Joaquin A

That’s weak. We have a huge sample size of greatness and a much smaller sample size that corresponds with his line (which was never great) being significantly worse.

Why don’t you use this awesome formula of yours to look at Todd Gurley’s last 10 games. That guy is finished, right?

You upvoted yourself for both comments? Someone must be insecure.

• Are you using 5.0 ARY/A as a proxy for average or replacement, or is it more of an arbitrary baseline?

• Little bit of both. The two bonuses are significant. I looked at one of Peterson’s games that was like 22 for 65 with 3 first downs with 1 TD, and that came out to 4.7 and didn’t seem too good to me.

• Now if you multiply by attempts you basically have the rushing version of RANY.

• I did, no?

• We’ll pretend I never asked that question. But we’ll leave the evidence up in case I get cocky in the future.

• I used 60% of average as a baseline (more fair to have a lower baseline for rushing than for passing, IMO), and came up with basically the same graph.

https://pbs.twimg.com/media/CtDyJ6eXEAQTJTy.jpg