Something to chew on for Josh Allen’s 2021
Buffalo Bills quarterback Josh Allen was spectacular in 2020. That’s not a hyperbolic statement. He ranked second in completion percentage over expectation (CPOE), third in expected points added (EPA)/play, fifth in adjusted net yards per attempt (ANY/A), third in QBR, fourth in passer rating, third in defense-adjusted value over average (DVOA), and fifth in PFF grade among NFL quarterbacks this past year. For those keep track at home, that means his STEW score (Bruce’s proprietary holistic metric composite) is 3.57, which was second in the league behind NFL MVP Aaron Rodgers. In addition to this, his volume statistics—4544 passing yards and 37 passing touchdowns, 421 rushing yards and 8 rushing touchdowns—made him the number-one fantasy quarterback. No matter how you slice it, Allen had a fantastic year.
The next question that logically comes to mind for many Bills fans after that initial statement passes the sniff test is: “Will Josh Allen regress in 2021?” If you’re hoping this article will give you the definitive answer through the crystal ball I’ve been hiding, you’ve come to the wrong place.
I would like to instead offer you three particular forms of regression Josh Allen could experience that, even if they occur, shouldn’t be construed as overly negative outcomes. All three seem like possibilities, but it’s important to set proper expectations BEFORE stimuli to better react AFTER stimuli. The way I see it, each of the following forms below the break represent a “type” of regression (raw statistical e.g. passing yards and touchdowns or efficiency e.g. DVOA, QBR, passer rating, etc.) followed by a “cause” of regression and the pairing of these two can help Bills Mafia decide how to respond.
1. Minor efficiency regression from a superlative peak
If Aaron Rodgers doesn’t go absolutely bonkers in 2020, the reigning MVP of the NFL plays in Orchard Park, looks good in shorts, and corrects media members when they fail to realize there’s only one New York team. Josh Allen’s STEW score was higher than Lamar Jackson’s during his MVP run in 2019.
Allen can regress simply because last year was so good.
Allen could regress and still have a top-five QB season. He could regress and this team could win a Super Bowl in plurality due to his play. Because of how good he was in 2020, we have to establish boundaries for how significant the regression even has to be before we even would call it such.
Right now, someone is out there saying: “He’s gotten better every year, so why wouldn’t he continue to get better?” He might.
But there are two counters to that: first, a raw statistical jump from 2020-2021 equivalent to the same raw statistical jump from 2019-2020 would mean that Allen would throw for almost 6,100 yards and 57 touchdowns. The 358 passing yards per game over 17 games would be 50 yards higher than the career average of the number-one QB in yards per game in NFL history (Patrick Mahomes).
Secondly, everyone who has spoken with even a small amount of fondness about Josh Allen since he was a prospect coming out of Wyoming has said he had a “high ceiling.” The operative word there is “ceiling.” Every quarterback has one. They don’t get better infinitely. We know now that Allen’s ceiling is one of the best quarterbacks in the NFL. He was that in 2020 unquestionably and normalizing that production would make the Bills Super Bowl contenders for a long time.
2. Raw statistical regression out of lack of necessity (better defensive play)
Josh Allen and the Bills’ passing attack carried the team in more than a few games in 2020. Early in the season, the Miami Dolphins and Los Angeles Rams games frankly required MVP-level play from Allen. Had the defense performed better, the Bills could have gone to a four-minute offense sooner and Allen’s passing yardage and passing touchdowns numbers wouldn’t be as impressive on paper.
If the Bills’ defense improves in 2021, Allen may not need to get that last 80-yard drive in under pressure, which is primarily passing the ball to deal with the clock. Taking away a few of those over the course of the season could lead to a raw statistical regression when no actual regression in ability has occurred.
3. Raw statistical regression out of lack of necessity (better running game)
Drives are a finite resource in the NFL, as are yards. In addition, there is a cap on potential scoring based on opponent drives and the clock that makes it a pseudo-zero-sum game. If you get more yards when you run the ball, there are less yards on that drive available for you to potentially get passing the ball. The Bills’ running game was middling in 2020, with both head coach Sean McDermott and general manager Brandon Beane remarking this offseason that they wanted it to change in 2021. If it does (much like the defensive point above), it may lead to a raw statistical regression when Allen didn’t actually regress in quarterback talent or execution at all.
These three types of “regression” may show up when you look at Allen’s 2021 on paper post-season and compare it to 2020. All of them would be classified by this writer as being “reasonable regressions” and wouldn’t be cause for even a little bit of alarm.
…and that’s the way the cookie crumbles. I’m Bruce Nolan with Buffalo Rumblings. You can find me on Twitter @BruceExclusive and look for new episodes of “The Bruce Exclusive” every week on the Buffalo Rumblings Podcast Network!