From The Athletic:

J.J. McCarthy. How do I know EPA per dropback is a trustworthy QB stat? Peyton Manning leads all quarterbacks in that stat since 2000. Patrick Mahomes, Drew Brees, Tom Brady and Aaron Rodgers are among the top 10.
Well, after Minnesota's McCarthy completed 12 passes for 87 yards and two interceptions in a 23-6 loss to the Packers, he now has an EPA per dropback that ranks 275th among the 282 quarterbacks to throw at least 100 passes in that era.
That's three spots below Josh Rosen, also a former No. 10 pick, who was replaced by the Cardinals after a brutal rookie year. Could McCarthy follow a similar trajectory? Mac Jones, Joe Flacco or any of the following should tempt Kevin O'Connell ...
Hey, I like that you're using statistics not normally seen that actually are, in the end, considered relevant and meaningful. But it might help those that aren't familiar with the statistic if you explained it a bit—not necessarily the nuts and bolts of how it was derived, though that isn't impossible to speak upon broadly—but that we generally aren't supposed to look at EPA as an indvidual-based statistic until there's some sort of consistent value that can't be explained away from a team perspective. People, including the good statisticians whose work matters more to the them than the notoriety of the warpath, will tell you that because of the impossibility of isolating an individual in football performance a team's expected points should not be used to judge an individual.
People should use it with massive caution for RBs or WRs or TEs. But for QBs, it's becoming increasingly common to compare QBs by isolating the result per each dropback. I don't care for it. I think unless you have a massive sample size you're picking up coaching strategy and competence, offensive line quality, receiver talent, and the unbalanced schedule as confounding factors. But I'm also not an expert. And it has a lot to do with the crossover of gambling, gambling props, and statistics that is partly responsible for the bleeding of this stat into individual assessments of talent.
I think it's fine for Manning, but not for a kid who has played six games or so. Actually, the more I write, the angrier I am at this writer. This is a misuse of the stat.
An aside: RYOE (rushing yards over expectation) is beginning to be used often and used loosely because it does help teams determine usage between
their own backs—you can take certain, similar backs that you are endeavoring to run similar plays, and that's what a lot of fantasy writer's readers are concerned with so the writers give them the stat. But it's technically not, or shouldn't be, just RYOE the teams are using to do that with because RYOE is the difference of the initial figure the model thinks the back should get and what he actually got; so the teams are likely using both the back's RYOE and, more importantly (probably, this is guess by me) the original estimate of how far the back should get according to the model.
So it's not just RYOE, which uses the back's ****AND THIS IS KEY*****
own MPH and acceleration (measured by chips in the player's shoulder pads that give their location at every tenth of second) to calculate his expected distance. They calculate that speed, in addition to where his blockers are (they disappear at handoff in the model) and where his pursuers are and how fast they are going at the exact moment he gets the handoff, and therefore, the whole stat (if used to compare backs on different teams) is totally meaningless and should absolutely not be used for cross-team comparison.
eta* this is edited to try and convey it with brevity and clarity, but that is being lost the more I edit this post for total transparency and accuracy. You guys should get with this and look at it before all the quant kids are winning all the leagues and feeling all the pride, sowing their fantasy oats, and making a living cashing quant checks (do they still use the big ones for photo-ops?) while Troy Aikman still spits the word "analytics" out on national television like he ate the part without the "ytics" and just realized what exactly was in his mouth.
eta2* And I'm realizing that this might sound condemnatory but it's so far from that. Keep bringing good discussions and patience (you have more than I do) and you'll have people reading facts and then opinions in short work.