The Patriots’ screen plays are still going as badly as you think

Back in the dark, dark days of 2022, when every other offensive play the New England Patriots ran seemed like one you probably recognized from college intramural flag football, it seemed like we probably would’ve been ecstatic about just about anything we recognized from the Josh McDaniels offense.

Matt Patricia carrying on the proud tradition of dialing up a too-clever-by-half screen play on gotta-have-it third-and-longs was probably not one of them.

You know the one.

And now, we take you live to all of our reactions last year when the Patriots, unable to move the ball under just about any meaningful way last season, snapped the ball and Mac faked the screen one way just to throw it the other direction:

Yeah. Definitely don’t miss that.

For all his flaws and infuriating quirks, the Josh McDaniels screen game was something the Patriots both executed well and had little problem carrying over as different skill-position players and offensive linemen came and went. And to his credit, yeah, Josh tended to spam a lot of the same plays over and over, but at least he made an effort to fake out the defense with pre-snap motion, play-action, and weird formations and route combinations to avoid telegraphing “yup, it’s a screen again”.

Fast forward to the present day, and Bill O’Brien taking over offensive coordinator duties was supposed to mean, or at least we could hope for, a return to at least running screen plays like a professional football team, and not a Turkey Bowl squad. Maybe BoB didn’t have the track record of gratuitous play-action or pre-snap motion quite as much as Josh McDaniels did — and who loves both of those things more this side of Kyle Shanahan than Josh, really — but in theory, at least, O’Brien would know enough to tell the guys, “OK, you go here. You go here. Then when the ball is snapped, you pull this direction, you block this guy” and so on, and so on. You know, be able to teach the concept well enough to execute it in a competent fashion.

Well, now we have some data on how this specific aspect of the Patriots’ offense is going in 2023, and while the Buffalo Bills upset this past weekend was a heck of a pallet cleanser, the numbers are stark.

To paraphrase Brian Fantana, “60 percent of the time, it doesn’t work every time”.

Observe!

If you’re reading this on your phone on the train or something and the chart is too small, here’s the TL;DR — the Patriots run screens on 12 percent of their offensive plays, and when they do, those screen plays generate a positive EPA (expected points added) roughly 30 percent of the time.

Let’s focus on the latter half of that stat for a second. Here’s how Expected Points Added is defined, and what that 30% means:

EPA per Play is simply the mean of the offense’s or defense’s total EPA. An advantage of this stat is that it isn’t too hard to interpret: since the mean of EPA per Play is 0, a positive value indicates a good play whereas a negative value indicates a bad play.

Put another way, roughly one out of every three screens the Patriots are running are actually putting the team in a better position to win. That doesn’t even mean getting a first down or getting eight on 1st-and-10 or anything like that; it just means “this play increased our chances of scoring more points, and therefore winning the game, even if just by a little bit”.

Move up to the far top-right corner of the graph — the teams that both run a lot of screen plays, and usually rack up some serious yardage when they do — and you’ll notice one thing they all have in common.

It’s the 49ers, the Dolphins, the Chiefs, and the Lions. They’re all really freaking good.

Not only are these teams, by and large, full of awesome players, but several of those coaches are well-renowned for being masters of the screen game. Specifically, Andy Reid of the Chiefs, who’s been putting defenses in a blender with it since his Philly days:

(watch this whole thing. There’s so much detail that goes into this, it’s hard for a blog boy like me to imagine executing it in anything close to real time)

Back to the Patriots, though. If Arjun’s chart on how well (or not-well) the Pats are executing screen plays didn’t quite hit home for you, how’s this for a stat:

Arjun did the same chart for the 2022 season — the Matt Patricia season, mind you — and the Patriots’ success rate was actually slightly higher, despite running them even more often. Not a typo!

Yeah. As infuriating as seeing another throw behind the line of scrimmage was last year, the Patriots were actually getting better results with it in the Matt Patricia offense than they are right now!

Throw it on the pile of “we better figure this out sooner rather than later, because getting a professional offensive coordinator was supposed to fix these things”, I guess. Sure, an upset win against a title-favorite division rival will salve a lot of wounds, but this team is still 2-5 and quite possibly staring down a top-10 pick in the 2024 draft.

It’s not like they can afford to be bad at anything this basic.