## Yards Per Attempt, Winning Percentage, and Outlier Teams

Yards per Attempt is not as good as Net Yards per Attempt, which accounts for sacks, and it’s not as good as Adjusted Net Yards per Attempt when it comes to predicting wins, since that metric includes touchdowns and interceptions. But still, vanilla Yards per Attempt usually correlates decently well with winning teams. The emphasis here is on the word usually.

There were four teams that stood out from the pack in yards per attempt last year: while 28 teams averaged less than 8.0 Y/A, four team averaged 8.2, 8.3, or 8.4 yards per attempt. Those teams were Dallas, Green Bay, Pittsburgh, and…

Why don’t you try to guess the 4th team.

[Come on, give it a good try.]

[Wrong. Guess again.]

[Nope. One more guess.]

The fourth team, believe it or not, was Washington. RG3 averaged 7.9 yards per attempt last year, which ranked 5th in the entire NFL among players with at least 214 pass attempts. Yet Griffin ranked 3rd on his own team in Y/A! On 10 fewer attempts, Kirk Cousins actually averaged 8.4 Y/A, while Colt McCoy averaged 8.3 Y/A on his 128 passes. As a result, Washington was outstanding in yards per attempt last year, despite being really bad.

The other big outlier last year was Tennessee. You may remember the 2014 Titans as co-starring in such hits as “Worst Record in the NFL” and “Bottom 5 in Points, Yards, and First Downs.” But the Titans actually finished the season above-average in yards per attempt, led by Zach Mettenberger and his 7.9 Y/A average.

Washington and Tennessee were big outliers last year. How big? If you perform a simple linear regression using Relative Y/A (i.e., the difference between a team’s Y/A average and the league average Y/A) as your input and winning percentage as your output, you get the following best-fit equation:

Win % = 0.1705 * Relative Y/A + 0.4994

Washington averaged 8.16 Y/A, Tennessee averaged 7.29, and the league average was 7.21. That means we would expect Washington, with a Relative Y/A of +0.95 to have a .661 winning percentage, far greater than the team’s 0.250 winning percentage. The Titans had a RY/A of +0.08 which would translate to an expected winning percentage of 0.513, far greater than the team’s 0.125 winning percentage. In the graph below, I plotted Relative Y/A on the X-Axis, winning percentage on the Y-Axis, and each team’s placement on the graph according to those two variables. I’ve also included the linear best-fit line and labeled Washington’s and Tennessee’s dots, so you can see just how far from expectation those teams truly were.1

That leaves me with a couple of questions. The first is how extreme are these outliers over the course of a larger period? I re-ran a regression using each team season since 1990. The best-fit formula was:

Win % = 0.1554 * Relative Y/A + 0.500

Based on those numbers, Washington’s winning percentage was 0.397 points below expectation, while Tennessee’s was 0.387 below expectation. Those were the second and third largest gaps since 1990, trailing only the 0-16 2008 Lions. Detroit had a Relative Y/A that year of -0.46 — you know, not too bad! — so the Lions were expected to have a 0.428 winning percentage. The other largest underachievers: the ’96 Ravens, ’97 Raiders, ’96 Jets, ’06 Lions, ’00 Chargers, ’01 Lions, and ’97 Colts.

The biggest overachiever was the ’01 Bears, who went 13-3 despite posting a Relative Y/A of -0.96. That Chicago team was followed by the ’95 Chiefs, ’05 Bears, ’10 Falcons, ’08 Titans, ’97 Chiefs, ’03 Patriots, ’99 jaguars, ’00 Ravens, ’06 Bears, and ’11 Ravens.

The other interesting question here is why did Washington and Tennessee struggle so much? You might think it was due to their defense, and while both teams were bad defensively, both were just as bad on offense. The real issue is that Y/A was very misleading for both teams.

Washington had a miserable sack rate, a terrible interception rate, and a terrible touchdown rate. Sure, the team ranked 4th in Y/A, but they ranked just 18th in ANY/A. Washington wasn’t a good passing team last year but a below-average one that just happened to excel in Y/A.

Tennessee wasn’t quite as bad in sack rate, interception rate, or touchdown rate, although the Titans do drop from 14th in Y/A to 25th in ANY/A. The Titans were also a bit unlucky, going 1-4 in games decided by 7 or fewer points. And it’s not as though Tennessee had a strong defense or running game to pick things up even if the passing attack was decent.

1. The next biggest “underachievers” under this methodology would be Tampa Bay, but after that, no other team is even in the same zip code of underachieving (the Falcons, Jaguars, and Giants are next). As for as “overachievers”, the Patriots, Cardinals, Lions, Bengals, and Seahawks lead the way. []
• Bryan Frye

I imagine if we looked at YPA differential we’d see fewer outliers. Then again, I haven’t looked into it and could be super wrong. I often am.

Y/A is a poor metric for measuring red zone performance, and of course the red zone has an outsize effect on game outcomes. I would think that partly explains many of these outliers.

• Richie

If I did my play index correctly http://pfref.com/tiny/JMomU Tennessee ranked 12th in red zone score percentage (88.6%) and Washington ranked 17th (87.2%) in 2014. League average was 87.4%.

What about red zone TD%? That would probably give a better indication of performance.

• Richie

I forgot to save my work. I think Tennessee is about 34%, Washington is about 32% and league average is about 35%.

Tennessee’s problem was that they were 3rd from last in red zone visits (slightly ahead of Oakland and Jacksonville).

Shockingly (to me), Miami had the most red zone visits (65).