With training camp around a month away for most NFL teams, fans across the league are looking for reasons to be optimistic about their favorite squads in 2022. And we'll attempt to provide that here, projecting which quarterbacks are most likely to improve this upcoming season.
To do this, we have to pick our metric of choice to evaluate quarterbacks. Keep in mind that no single metric perfectly describes every aspect of quarterback play, and there are pitfalls with all of them that we will address.
We'll use expected points added (EPA) per pure dropback as our measuring stick, partially because it is slightly more stable year-over-year than raw EPA per dropback, and partially because success on run-pass options, screens and play action is less reliant on a quarterback's overall ability.
This does not consider a signal-caller's ability on designed runs or how well their passing offense is constructed, and it doesn't fully isolate them from their surrounding talent. But we can use it for our purposes as an effective one number to approximate a quarterback's on-field output.
Here are the 2021 NFL QBs ranked by EPA/Dropback on "Pure" Dropbacks (I removed Play Action passes, Screens, and RPOs bc that's fake production) pic.twitter.com/LwGkxQoWDe
— Conor McQuiston (@ConorMcQ5) June 23, 2022
To determine how likely a quarterback is to improve in EPA per pure dropback in 2022, we first fit a linear regression based on their 2021 EPA per pure dropback figure, an indicator for if a quarterback is entering their second season and the yards per route run (YPRR) of their projected top three wide receivers in the previous season. This is built based on the 2014-21 NFL seasons, and the top three receivers are determined by those who ran the most routes in the previous season and remain on the roster. If a team has a first-round rookie wideout, they are projected to be the No. 3 receiver and their YPRR is projected based on their draft slot, not their most recent college results.
The results from this linear regression were then used as a basis for a Monte Carlo simulation, where each fit was considered the mean of a normal distribution of outcomes for a quarterback. For each run of the simulation, the quarterbacks were ranked by their EPA per pure dropback as determined by the pull from their normal distribution. They were considered to have improved if their rank in a run was better than their rank in 2021. Passers who retired following 2021 or are unlikely to start in 2022 were removed. We ran 100,000 simulations.
This method does skew toward quarterbacks who ranked poorly, as it is only possible for them to improve. However, there are interesting differences in the degrees of improvement. It’s important to note that this does not consider tight ends or pass-catching running backs, since there are only a handful of quarterbacks in any season where that's significant. Additionally, it does not consider offensive line play, given the difficulty in untangling pressure rates from quarterback play and individual offensive linemen. Coaching changes are also not considered because of the immense difficulties in quantifying the value add of individual coaches. This also does not account for how a quarterback's injuries may have affected their play.
With that out of the way, here are the 10 quarterbacks who are most likely to improve in 2022:
Wilson is the worst returning quarterback in the NFL by EPA per pure dropback, and so he can really only improve. The model projects a 15.7% chance of him remaining at the bottom of the league, but it's more likely the second-year BYU product will improve with rookie Garrett Wilson and a healthier Elijah Moore in the wide receiver room. The model is also pessimistic about the degree of his improvement, giving him a less than 20% chance of being a top-16 quarterback in the NFL in 2022. There is a chance it is not fully accounting for the typical size of quarterbacks' second-year leaps.