Adjusted Quarterback Efficiency 3.0

Refining adjustments, including adding benefits from play-action usage, temperature and humidity to weather and PFF pass blocking grades

You can find the details behind the methodologies for adjusted quarterback efficiency (AQE) 1.0 and 2.0 here.

New adjustments

The only brand new adjustment this week is adding a scheme-based metric for play-action usage. Spamming play-action passes doesn’t produce the same level of outperformance as we’ve seen in the past, but I still found around 0.1 EPA benefit for using play-action on early downs when you control for quarterback and other factors.

The play-action adjustment in AQE 3.0 looks specifically at how often a team uses play-action versus league-wide expectations. Simply put, if a team is constantly in third and long situations where running the ball isn’t a credible threat, they aren’t going to have the same relative opportunities to use play-action. Other times when play-action usage is fairly nonexistent are trailing by more than two touchdowns in the fourth quarter, and in two-minute offenses outside of the red zone.

The play-action adjustment for each quarterback looks at the specific situations they were (down, distance, etc) on a play-level basis, calculates the expected play-action rate and compares it to their actual play-action usage. Every time they use play-action above expectations, they are debited by roughly 0.1 EPA.

It’s not all offenses should be calling play-action all the time, there are clearly some quarterbacks who can generated better efficiency that expected under play-action in different circumstances, so their play-callers wisely allow them to use their extraordinary abilities to enhance overall efficiency without relying as heavily on play-action. But the assumption baked into the adjustment is that it requires more quarterback skill to outperform with lower play-action usage, and those quarterbacks should be credited for that ability in relation to their unadjusted efficiency.

Refining adjustments

The two adjustments I altered, and hopefully enhanced, in AQE 3.0 are for weather and blocking. I incorporated more granular weather information from FNTData to add temperature and humidity to previously only using wind and stadium (indoor or not). This gives a little more credit to quarterbacks playing in colder conditions, or in rain. Wind is still the primary weather factor affecting passing, but those others have a material effect and needed to be added.

For the blocking adjustment, I’m incorporating PFF player-level pass blocking grades in addition to my pressure survival curve calculations based on play-level data from FTNData. While I’m confident in my survival curve methodology, it often enhances modeling to have multiple strong data sources than just one, and the PFF grading likely can catch some contextual elements of scheme and play-type missing in the pressure data.

For the team blocking grade, I weigh each position group (tackle, interior offensive line, other) by an estimate of its importance and then aggregate up value by the blocking snaps and grades for individual players.

2022 AQE results through Week 16

I’ll now display and walk through the results from the adjusted detailed above, and what they say about more isolated quarterback efficiency this season. I’ve added a few quarterbacks excluded last week due to not having enough dropbacks. Here’s is AQE 3.0:

*You can find all the detailed data for AQE, in a downloadable format, in the Unexpected Points Google stat sheet for paid subscribers.