Regression to the Mean

15 Oct


Indeed, Fake Jeff Duncan. Indeed.

In fact, I’m having trouble finding a statistic that doesn’t point to a regression. [1]

Another game in the books, and the Falcons continue to be outgained on a yards-per-play basis. It gets even worse when we focus on the last three games (which you’d expect when you’re talking regression, of course), and most of this post will, indeed, focus on the Falcons’ last three games, versus Carolina, Washington and Oakland.

First, just to get it out of the way, those teams are a combined 5-11 on the season. Not exactly a gauntlet of juggernauts right there.

Back to Fake Jeff’s point: on the year, the Falcons have been outgained on a per-play basis 5.9 to 5.6 yards. The regression is clear when this stat is broken down to a per-game basis: +1.14, -0.78, +0.64, +0.04, -1.39, -1.57. [2]  That’s ugly, and with one notable exception (week 2 against San Diego), it’s a steady turn for the worse in a pretty good predictor of wins and losses. Excluding the AFC East, which is currently locked in a four-way tie (what bullshit that is), the Falcons are the only team in the NFL with even a share of a division lead that has a negative overall yards per play differential. With the best record in the league, they are tied at 12th with 3 other teams in yards per play, and are tied with four other teams at eighth most yards per play allowed.

The combined records of the five unique teams with whom they are tied in those two categories? 10-18.

Now, anyone who reads this blog is probably aware that around here we believe in the Scoreboard as the ultimate predictor of efficiency. And yes, the Falcons have won that statistical category in all 6 games. So why don’t we see if there’s a way to detect regression in PTOMAC™ as well?

I would think that would be easy. We can compare the scoreboard with the overall trend in scoreboard numbers for each individual team. We can look at the points the Falcons scored compared to that defense’s usual allowance, and we can look at points the Falcons allowed compared to the points that team usually scores, and subtract! We’ll call it…point differential!

“They already have that, dumbass.”

Oh. Well then. Let’s do it. Subtracting the opponent’s average point differential from the differential of the games they lost to the Falcons, we get the following six numbers:

+2.8, +10.2, +28.4, -4.6, +7.8, -9.2

If you’d like to total it, that’s +41.4 in weeks one to three, and -6.0 in weeks four to six.

Yikes, Falcons.

Probably the most glaring issue with the Falcons right now (at least statistically) is the problem with the running game.

“Do you mean Offensively or Defensively?”

Yes.

Defensively, the Falcons have allowed 863 yards rushing on 165 carries in six games, for 5.2 yards per attempt. That total rushing yards statistic is fourth worst in the league. [3]  The yards per attempt number is second worst in the league.

“Maybe they’ve played a bunch of great running teams.”

The Falcons could only wish. Counting the big days against the Falcons, opponents have gained 3965 yards on 888 carries. That’s an average per team of 148 for 661, at 4.47 yards per carry. This hypothetical anti-Falcons team would rank 13th in total rushing yards, and 10th in yards per carry. Including their games against one of the league’s worst rushing defenses in the Falcons.

Their offensive rushing attack isn’t much better. 139 attempts for 519 yards at 3.73 yards per carry lands them at 18th overall in rushing yards and 27th overall in yards per carry. Against opponents who average (per team) 575 yards allowed on 139 carries, at 4.13 yards per carry.

Ultimately what this all means is that the Falcons have relied for three straight weeks on luck (opponents managing to meet the Falcons’ increasing turnover rate – 1, 2, and 3 respectively – by making mistakes, few of which were “forced” by defense) and Matt Ryan’s ability to throw it all over prevent defenses with a minute left in the game.

Matt Ryan’s passer ratings in the first six games: 136.4, 101.5, 107.8, 107.2, 89.0, and 59.4.

If history has taught us anything, luck doesn’t last. Matt Ryan will make mistakes. And teams who get outplayed for three hours a week eventually start losing the battle on the scoreboard.  The schedule doesn’t get any easier, either: @Eagles, Cowboys, @Saints, Cardinals, @Bucs, Saints coming up.  I see two clear losses right there, am I right? 

Wishful thinking aside, this will catch up with the Falcons, and they’ll come back down to Earth faster than Felix Baumgartner. Felix landed more gracefully than the average human lands on the first step of a moving escalator. Not the Falcons. The Fall of the Falcons will involve fire, and a sickening thud.

I hope.

I hate them so much.


 [1]  All stats were taken and/or calculated from various pages on Pro-Football-Reference.com.  Of course, any errors probably result from my own dumbassery.

[2]  Overall yards per play calculated by cumulative (total yards) / (pass att. + rushes + sacks).

[3]  Ten teams have only played 5 games to this date, so some of the “total” stats are skewed.  This is why per-play or per-game numbers are also noted (although, some of the rankings are counterintuitively worse than the totals for Atlanta’s defense, which has played 6 games).

Tags: