NFL Draft 2022: What QBOPS can tell us about the quarterbacks

The Green Bay Packers probably won’t draft a quarterback this year. For one thing, it’s not a great year for quarterbacks in the draft according to, well, pretty much everyone. And of course, the Packers still have Aaron Rodgers, and to a much lesser extent, Jordan Love. Still, we probably shouldn’t ignore the position completely, as the Packers haven’t exactly given Love a huge vote of confidence, and the end of the Aaron Rodgers road is still in sight.

The Packers have obvious needs at receiver and edge rusher and could use some depth on the offensive line, but should they see some value in a developmental quarterback, they have enough picks to spend one on the position. And because this class is so scattershot, some talent may find its way into the latter rounds. This team used to routinely take late-round picks on quarterbacks like Matt Hasselbeck, Mark Brunell, and Aaron Brooks, and given where they are as a franchise, I won’t be too surprised if they start kicking a few tires.

QBOPS Primer

I’ve written a few pieces on Wide Receiver OPS (WROPS), the primary statistic I use to assess receiver productivity in college. For quarterbacks, I use the extremely similar Quarterback OPS (QBOPS). Like WROPS, QBOPS boils down production to a single number scaled to baseball’s OPS statistic. Where WROPS is made up of Catch% (WROBP) and Y/R (WRSLG), QBOPS is made up of completion percentage (QBOBP) and yards per completion (QBSLG). QBOPS seeks to answer the questions of whether a given quarterback was accurate considering how many big plays down the field the quarterback is made. It will credit extremely accurate short throwers as well as somewhat less accurate deep ballers, and is fairly similar to, though less sophisticated than, Completion Percentage over Expected (CPOE).

QBOPS shares one other similarity with its baseball counterpart that is important to note. In the traditional OPS statistic (which is literally just “On base Plus Slugging”), slugging percentage is over-weighted compared to OPS. Here too, slugging is overrated, and the On-Base portion (that is, completion percentage) should be considered first. A quarterback with a .350 QBOBP will almost never be good, even if he brings an absurd QBSLG to the table to compensate.

QBOPS works quite well historically, and the best professional quarterbacks tend to come from a group I call the .400/.600 club. The very best tend to have a QBOBP north of .415. Going back to the 2011 class, it saw Russell Wilson (.430/.606/.1036) as essentially equal with Andrew Luck (.421/.670/.1.091) and RG3 (.427/.619/.1.046), but it did even better than that. 127 quarterbacks qualified for a QBOPS score that year, and in addition to the big 3, it also picked out the less heralded Case Keenum (.419/.691/.1.110) and Nick Foles (.408/.721/.1.129). It’s not perfect by any stretch, and I would be dishonest if I didn’t point out that it also liked Brandon Weeden (.427/.619/.1.046) and Matt Barkley (.407/.674/.1.081) that year, and scouting is still an enormous part of quarterback evaluation, but like WROPS, it forces evaluators to answer the important question, “why are you down on this guy given how productive and accurate he was in college?”

Moving to more modern times, in the 2020 NFL draft, the .400/.600 club identified the obviously great Joe Burrow (.450/.691/.1.141), but also the less heralded Tyler Huntley (.431/.689/1.120), who ascended to the backup position in Baltimore despite his status as an undrafted free agent and who played extremely well against the Packers last season. And while QBOPS was probably too high on Tua Tagavailoa (.421/.773/.1.194), it also saw a solid future for Jalen Hurts, who continues to play reasonably well in Philly (.411/.796/.1.207). Finally, in the 2019 draft QBOPS was a fan of Kyler Murray (.407/.674/.1.081) and hated Daniel Jones (.357/.542/.899). I mention Jones because while QBOPS often likes a few too many guys, quarterbacks with poor QBOPS (outside of Josh Allen) are almost never any good.

For example, QBOPS was also down on Drew Lock (.371/.589/.960), Mitch Trubisky (.401/.589/.990) and Jordan Love (.365/.569/.934). Even Love’s far superior sophomore season (.378/.773/1.151) still isn’t great when compared to his peers, simply because he didn’t complete as high of a percentage of passes. Finally, I always like to mention the absolutely atrocious Christian Hackenberg who, in 2015, had a .316/.644/.960 line, and was somehow drafted in the second round. Hackenberg’s raw completion percentage ranked 96^th out of 111 qualifiers that season.

Before we get into the up-and-coming class, it’s worth noting that there has been some completion percentage inflation in college as offenses grow even more sophisticated and create easier throws for their quarterbacks. In the 2020 season, Mac Jones, Zach Wilson, Justin Fields, Matt Corral, Trevor Lawrence, Kyle Trask, and Sam Howell all put together outstanding seasons that would have marked them as sure-fire stars back in Russell Wilson’s day, or even in Mitch Trubisky’s day. A .400/.600 isn’t what it used to be, and the amount by which a prospect clears the threshold is quite important. There’s old analytics saying, that when a metric becomes a target, it ceases to be a useful metric, and I think we’re seeing some of that. It’s important to take the extra step with any quarterback to determine how much of his success is due to the surrounding talent, the scheme, and the level of competition faced.

As an example, Matt Corral of Ole Miss (.399/.630/1.029) has been exceptional over the past two seasons according to QBOPS, but almost all of his production comes on RPOs, and almost every snap he takes is from the shotgun. There’s nothing inherently wrong with this, but when you do scout him on the few standard throws he does get to make, his efficiency evaporates. Because RPOs are simple read plays, Corral struggles in looking off defenders, and he throws far more interceptable balls than he should, especially when a play breaks down. He’s also on the small side, and while his arm is adequate for an NFL quarterback, it’s only that. Given that he’ll need to learn the under center game from scratch at the next level and has no success as a read-processor, he’s fallen down draft boards. My stats say he’s good. Scouts disagree. After watching him play, I agree with the scouts.

The 2022 Class

There are two standouts in this class, and only one of those is currently considered to be a high round talent. The chart below lists every prospect in this class, their QBOPS along with its component parts, QBOPS+ which adjusts for interception percentage (look for those with a difference of 0.1 or more between their QBOPS and QBOPS+, which is substantial), CPOE (per Pro Football Focus), PFF grade, PFF Big Board Rank, and Relative Athletic Score (RAS). Let’s start with the guys QBOPS likes before we start poking holes in everyone else’s game.

One final note: I ranked 132 quarterbacks via QBOPS at the conclusion of the 2021 season. The highest mark was set by the perennially great Grayson McCall of Coastal Carolina, at .431/.800/.1.231. McCall will be a very interesting case when his college career is complete, because while he’s obviously talented, no scheme does more for their offensive players than Coastal’s. On the other end of the spectrum, clocking in at .335/.456/.790, we have Ken Seals of Vanderbilt, who managed to rank 112^th in completion percentage while ranking 2^nd to last in yards per completion, and throwing 8 picks in just 224 attempts.

The Good

Kenny Pickett – Pittsburgh

QBOPS Splits: .396/.634/.1.030 (31^st overall)
CPOE: +7

While Pickett falls just barely short of the threshold for QBOBP, he’s close enough to warrant consideration, and upon closer inspection, he easily comes out on top among his peers. While QBOPS sees Pickett as almost identical to the previously discussed Matt Corral, the difference between the two is immediately apparent both on tape and via some additional statistical color. While the Pitt offense has its fair share of simple throws built in, Pickett is also adept at going through his progressions when things break down, making solid reads, and getting the ball downfield. While he isn’t the quickest processor I’ve ever watched, he is easily the quickest in this class.

QBOPS’ case is bolstered by its much more sophisticated cousin, CPOE, where Pickett’s 7% is easily the best in this class. CPOE isn’t just a measure of accuracy; it’s also a good proxy for decision-making. A good CPOE is driven as much by physical accuracy as it is in seeing receivers open downfield and targeting them on-time and with some authority. Pickett does all of this, and does it well.

Pickett isn’t without his flaws, as his hand size measurements at the Combine were extremely small (though they grew preposterously larger at his pro day). The biggest knock on Pickett though, and the reason I suspect he is consistently mocked after Willis, and occasionally after Howell, Ridder, and Corral, is his arm strength. There are a few cannons in this draft (especially Willis), and Pickett’s arm is merely above average. Arm strength can be a real issue at the next level, and as much as I love Baltimore’s Tyler Huntley as a prospect, arm strength is the one factor likely to hold him back. Pickett’s arm is well above Huntley’s, and within the acceptable range for a starting NFL quarterback. He should be fine.

In addition to WROPS and CPOE, Pickett is also an outstanding athlete, having posted a 9.55 RAS, and PFF’s graders had him as the best QB in college football with a 92.2 overall grade.

Pickett’s production was outstanding, his scouting was outstanding, and the knocks on him are fairly minimal. I think he is easily the top quarterback in this draft, and while I’ve seen many analysts project him as the second quarterback off the board behind Malik Willis, he just as frequently falls to 5^th, and currently sits 48^th on the PFF Big Board behind every other consensus top quarterback, and alarmingly close to Carson Strong at 68. Pickett is likely to be a bargain for a team in need of a quarterback prospect. In the top third of the draft, he is the only one worth taking.

Skylar Thompson – Kansas State

WROPS Splits: .410/.636/.1.046 (19^th overall)
CPOE: +5

And here is where things get a bit crazy. No one, and I mean no one, is talking about Skylar Thompson at all. He’s 255^th on the PFF Big Board, and by most accounts, is unlikely to be drafted. This would be a mistake, but let’s start with the knocks on Thompson before we get into the positives, because there are some real knocks.

First and foremost, he’s quite old as 2022 will be his age 25 season. The NFL does not care for older prospects as they lose valuable production from that player’s prime and because older players, due to their experience, can sometimes put up artificially high numbers in their final college seasons. It’s also worth noting exactly why Thompson is old. He lost almost all of 2020 to a serious upper body injury, but used the NCAA’s grant of an extra year of eligibility due to the Covid pandemic to come back for a 6^th season. He played brilliantly during that extra season. However injuries continued to plague him as he suffered knee and ankle issues, though nothing too serious.

Finally, Thompson doesn’t have the sample size of some of the other quarterbacks. Given injuries and Kansas State’s run-heavy offense focusing on star back Deuce Vaughn, there just wasn’t a ton of passing. Complicating matters, Vaughn played a large role in the passing game, leading Kansas State in targets in 2021.

So, what’s to like? First, Thompson was incredibly accurate given how often he challenged defenses down the field, completing almost 70% of his passes. It would be an easy thing to credit that high percentage to checkdowns to Vaughn, but if that were the case, QBOPS would see it, and punish Thompson for a lack of playmaking. Instead, it sees quarterback who made phenomenal use of his outside receivers down the field. Junior Phillip Brooks was impressive as Thompson’s primary possession receiver, averaging 12.6 yards per reception and repeatedly moving the chains, but I think the best case for Thompson is to look at his deep passing.

Thompson completed 162 passes on 233 attempts in 2021. 57 of those completions found receivers Malik Knowles (15.2 Y/R), Landry Weber (14.3 Y/R) and tight end Daniel Imatorbhebhe (26.2 Y/R). As a unit, those 57 balls gained 983 yards, averaging 17.3 yards per completion. On tape, Thompson was an underrated processor, primarily because Vaughn and Brooks were generally his first reads, but Thompson was willing and able to stay in the pocket, buy time, and wait for his downfield threats to come open. Thompson is easily the most patient passer in the class.

There’s also some evidence that Thompson’s physical tools are underrated.

And this bit of tracking data seems to comport with what I see on tape. Thompson’s arm strength is well above average (I think superior to Pickett’s), and he has no issues pushing the ball outside or downfield. He’s also more than adequate as an athlete, and even managed to be a semi-dangerous runner with some regularity.

Finally, Thompson was decidedly NOT helped by his scheme or the level of talent around him, as Kansas State runs a more traditional passing offense, and isn’t a Big 12 power. That said, being in the Big XII provided solid opposition for Kansas State, allowing us to see how Thompson would handle high level competition. In 2021 he played quite well against Oklahoma, LSU, and at Texas Tech, though he did struggle against Baylor.

Thompson isn’t some generational talent in hiding, but he is better than the vast majority of quarterbacks expected to go in the first two rounds. Given that he will cost you almost nothing, any team in need of a cheap backup that may have more to offer should be all over him in the later rounds.

The Questionable

We’ve covered Matt Corral (.399/.630/.1.029) above, so let’s dive into the rest of the highly ranked passers in this draft.

Carson Strong

WROPS: (.414/.559/.973)
CPOE: N/A

Carson Strong has been accurate, but hasn’t done enough damage down the field, especially given that his 8 interceptions is on the high side considering his number of attempts. The low QBSLG sees him as nothing more than a game manager, and it’s worth noting that .559 is really low. Occasionally, a very talented quarterback is held back by his college system, and winds up looking more like a game manager than he should. Such was the case with Justin Herbert who, in 2019, posted .394/.595/.989 splits at Oregon, just barely missing the .400/.600 threshold. Herbert’s physical tools should have allowed Oregon to do so much more, but even a held back Herbert was ripping off 12.1 yards per completion. Strong was at just 11.4 last season, ranking 94^th of 132 qualifiers in 2021.

I wouldn’t blame anyone for dreaming on Justin Herbert a bit because Strong is, well, strong. He has an absolute cannon, and like Herbert, Strong’s chief issue is his work under pressure, as he takes far too many sacks (and has suffered injuries because of it). The Chargers were able to turn this completely around for Herbert, but with Herbert, the big play seeds were far more apparent. At the moment, Strong’s decision-making under pressure is a disaster, and any team that selects him is gambling they can completely remake a slow-processing game manager into something much more. Not impossible, but a tall order.

Desmond Ridder

WROPS: (.383/.651/.1.034)
CPOE: +0.3

Desmond Ridder fell just short of my accuracy threshold and also had a higher than average pick total for his number of attempts. Successful quarterbacks do occasionally share Ridder’s statistical profile, (the .375/.650 club) with Patrick Mahomes being the most obvious example (.387/.689/.1.076), and Ridder’s RAS was excellent (9.61) meaning there is still some projectability there. That said, on film it’s clear that Ridder’s receivers (especially Alec Pierce) did him a lot of favors, as did Cincinnati’s scheme. Ridder has trouble getting through reads, and is much more of a see-it, throw-it processor with the ball. Watching Ridder made me want more Pierce and less Ridder.

Ridder’s arm strength will also be an issue at the next level. While he has no trouble ripping off throws when he’s set, when moved off of his spot or under duress, his throws lack zip and he gave opposing defensive backs far too many opportunities for turnovers. His issues with reads also manifest as an almost total inability to look off opposing defenders, which doesn’t help.

Ridder, to his credit, clearly responds well to coaching, and has improved his Y/A every year. You can see the difference from his junior season to his senior season, as he was once prone to take off running, as a senior he was much better at ascending the pocket and finding receivers downfield. I love a coachable prospect, and Ridder strikes me as a smart player, but his propensity to take sacks and that 0.3 CPOE testify to his remaining issues as a processor. It is the single most difficult problem to fix at the next level.

Sam Howell

WROPS: .369/.690/.1.059
CPOE: +1.8

Sam Howell had major accuracy problems and threw far too many interceptions for me to seriously consider him. While his 2020 was better, it was only marginally so, and I prefer prospects who improve in their final season. That dynamic sophomore season was driven primarily by receivers Dazz Newsome (6^th round, Chicago Bears) and Dyami Brown (3^rd round, Washington Commanders), and without his two top targets, he locked on to sophomore Josh Downs far too frequently. Downs’ 101 receptions were 70 more than the next closest Tar Heel receiver. Carolina runs one of the most quarterback friendly schemes in college football, and I expect far more out of any potential pro prospect from that system. Accuracy and sacks are the major red flag here to go along with 9 picks.

Malik Willis

WROPS: .360/.676/1.037
CPOE: +2.7

Finally, scouts love Malik Willis (.360/.676/.1.037), and I understand why. He has the best arm in this class, and he’s an exceptional athlete who routinely destroyed opponents with his legs as well as his arm. That said, of any prospect in the draft he was the least accurate passer, and that is a lot to overcome. His CPOE wasn’t atrocious, at 2.7, which means his accuracy may be slightly better than it appeared given the number of downfield throws he made; however, we also need to consider the level of competition he faced while at Liberty. Willis faced off against Matt Corral and Ole Miss on November 6^th, and Willis was unable to do much of anything in a 27-14 loss (16 of 25 for just 173 yards, and 3 picks). While he did manage to punch in a touchdown on the ground, he was also held to 71 yards rushing on an insane 27 attempts.

One game hardly makes a prospect, but when looking at small school guys, it’s nice to see at least some success against better competition. In 2020 Willis did have an impressive game against Virginia Tech, outdueling Hendon Hooker 38-35, but VT was just 5-6 that season, and did not have their typical class of defense.

If you take Willis, you’re hoping you can do for him what Buffalo did for Josh Allen, or perhaps dreaming on Lamar Jackson with a better arm. While Willis isn’t exactly old, this will be his age 23 season, while Allen entered the league at 22 and Jackson at 21. With a developmental prospect, extra years matter a lot. Raw quarterback prospects work out more in the modern game than they used to in the past, but Willis is still a long shot. Given the upside that quarterbacks provide, it’s understandable why teams continue to take that shot, but Willis is going to need to land in a good situation, and have a lot go right.

The Rest

It is a testament to the coaches at Iowa State that they managed to coax a 71.7% completion percentage out of Brock Purdy (.423/.535/.958). Purdy plays like quarterback like a man who is literally on fire, and throws more near-picks than any other pass I watched this season. He is my least favorite quarterback in this draft and there is no reason to waste a pick on him.

I briefly considered Bailey Zappe (.409/.614/1.023) as a potential diamond in the rough given his elite accuracy numbers. Unfortunately, he has the worst arm in the draft, falling below the minimum level of arm strength for success at the next level. Under any kind of pressure, he throws lollipops.

Old friend Jack Coan (.386/.610/.997) has the second worst arm in the draft, though I suspect he’ll be a valuable clipboard holder at some point.

Finally, I would urge everyone to watch some tape of Southeastern Louisiana’s Cole Kelley, who somehow tied Kenny Pickett for the highest PFF quarterback grade of the season at 92.2. That’s a ridiculous grade for a player who is a complete mechanical mess, and who won by overwhelming lesser competition, but Kelley is unquestionably fun as a 6-7 behemoth with giant, gangly arms, who was also used in the run game to a hilarious degree. He has a good arm, but like a lot of tall quarterbacks, no idea where the ball is going.