While this is an Eagles site, I was looking into the Cowboys prior to the season as part of the AV-based win prediction modeling I do. Last year I was wrong on the Cowboys, having them modeled at 7.4 wins. What went wrong? Two main things. Injuries play a big part in the model and I expected Dak to struggle more coming back from his ankle injury than he did. But second, the Cowboys (and Trevon Diggs specifically) had a crazy amount of turnover value.
Last year, the Cowboys were #1 in the league and 4th highest across all teams over the past 5 seasons in amount of defensive EPA generated from turnovers, with 138% of their total defensive EPA coming from turnovers. This puts them at a 98th percentile in defensive value coming from turnovers. As a further point, Trevon Diggs (who I like more than most Philly fans) generated 15 AV last season – Jalen Ramsey has only done that twice in six reasons, Jaire Alexander has never reached 15 AV, Marshon Lattimore has peaked at 9 AV in six seasons, and our own Darius Slay hit 14 AV once in his 9 seasons.
Good defenses generate turnovers… or turnovers make good defenses…
Defenses can be good without turnovers, but typically turnovers and better defenses go hand in hand. Below shows every team over the past 5 seasons with their overall defensive value in EPA on the X axis and the value generated from turnovers on the Y axis (note that for defenses, a more negative EPA is better).
With an R2 of 0.34, about a third of defensive value is explained by turnovers which is not shocking.
But turnovers are difficult to repeat year to year
Turnovers are notoriously unstable year to year for most teams, though. Below shows teams on the X axis by their percentile of turnover value generated (higher percentile to the right means their defense generated more turnover value) vs. that team’s change in turnover value generated the following season. A positive value on the Y axis means that team improved turnover value the following season, a negative number means it dropped.
This shows a large reversion to the mean – teams that generate a high amount of turnover value (teams to the right) generally cannot repeat that the following season. Conversely, teams that generate low turnover value (teams to the left) typically have better defensive turnover performances the next season.
Another view of the same, but this time showing the amount of change by each quintile grouping of defenses. If you aren’t familiar with box plots, they show the average and range of values. The first quintile on the left (20th percentile) are the defenses that generated the least amount of turnover value. The center line inside the box is the average change, which shows these defenses add, on average, 42 EPA of turnover value the next season. 50% of all defenses are inside the box and range from +11 to +65 EPA the following season and the remaining 25% of each side go to the ends of the lines.
The defenses with the lowest turnover value almost all do better the following year. And on the other side, the vast majority of defenses that generate the most turnover value (the far right) drop the following season, dropping on average 34 EPA generated from turnovers the following season.
Good players and good teams obviously impact turnovers, but turnovers are just really hard to consistently cause and there is some chance to it. We saw this last week when Darius Slay dropped an easier pick that would have sealed the game but then the improbable when Kyzir White tips a ball into James Bradberry‘s arms (while Bradberry was facing downfield) which gets returned for a touchdown.
Which teams may be headed for regression in 2022?
Dallas, Tampa, Kansas City, Cincinnati, Indy, and Arizona were the top quintile teams in 2021 with extremely high defensive value generated from turnovers. You can see how extreme Dallas was, by far the top team in 2021 and 4th best in 5 years. The Cowboys excess turnover value alone was worth over 2 wins last season.
Each of these teams have bounced around the league average over the past 5 seasons, with maybe Tampa Bay being the exception that has consistently generated turnovers recently. And each should be expected to regress defensively this season. Even before Dak was injured, this was the primary reason I had the Cowboys modeled under 8 wins.
On the other end, Jacksonville, Las Vegas, Baltimore, Carolina, the Jets, Detroit, Chicago, and Denver all were at the bottom of the league. A lot of bad teams here that lack defensive playmakers with Baltimore being the team that doesn’t belong in the list. The Ravens were ravaged with injuries, especially on the defensive side and in the secondary. They already started 2022 with 3 turnovers and will surely be one of the teams that raise their turnover value in 2022.
What about the Eagles?
It seems like we are always disappointed with the Eagles and lack of turnovers, but they have been pretty close to league average the past 5 seasons, with the exception of the Super Bowl year.
Sirianni and Gannon’s focus on stripping the ball has not panned out yet, but again, there is so much chance to it. I’ve written before on the pass rush and how opposing QBs have attacked the Eagles with historically quick (and short) passing games but with the investment in the defense with Bradberry, Kyzir, Jordan Davis, Haason Reddick, and C.J. Gardner-Johnson, the Eagles should be expected to show improvement here. At least that’s where I’m willing to stand.
Fly Eagles Fly.
Thanks to our friends at WagerWire