Accuracy remains one of the most puzzling aspects of quarterback scouting. In theory, it should be easy. Did the ball go where it was supposed to go? Yes? Great. But it’s not that simple, especially as we try to translate what college accuracy means for quarterback prospects heading into the NFL.
We’ve also reached a sliding scale on how much accuracy actually matters when compared to the value of throws that can be completed. With some of the big-armed alien quarterbacks that have entered the league over the past few seasons, the search is on for the next one.
Figuring out how to balance that scale is going to be big for the 2023 draft class, especially with projected first-round prospects Anthony Richardson and Will Levis. According to data from Grinding The Mocks, Richardson has an average draft position of 4.2 and Levis is at 5.9. These are two top-10 projected quarterback prospects who will have to improve upon accuracy issues in the NFL.
That, of course, brings us to the Josh Allen comps. Allen became one of the league’s best quarterbacks in his third season after completing just 56% of his passes in college. Since that time, Allen is second in EPA per dropback behind Patrick Mahomes in EPA per play, according to TruMedia.
Allen was viewed as a raw and inaccurate, but talented prospect coming out of Wyoming. His success since has warped the discussion around accuracy and quarterback development. But we might also have a slight misunderstanding of where Allen was as a prospect.
There is no ignoring Allen’s completion percentage to every area of the field. Among 100 quarterback prospects back to the 2016 college season, Allen had a 10th percentile completion percentage to the short area of the field (1-10 air yards), 31st percentile to the intermediate level (11-19 air yards), and 10th percentile on deep passes (20+ air yards) in his final college season, per data from Sports Info Solutions.
But SIS also charts on-target percentage, which better tracks the quarterback’s control of accuracy. SIS defines on-target as “a pass that hits a receiver in stride, regardless of whether the pass is completed.” On-target rate is going to be a one-for-one proxy of “accuracy” but it does give us a clearer look at ball placement.
If we look at how Allen fared in that metric, it’s still not great, but it’s much better than the bottom-third percentiles of his completion percentage. Allen ranked in the 71st percentile for on-target rate to the short area, 61st percentile on intermediate throws, and 47th percentile on deep passes.
Those metrics kind of fit who Allen is as a pro, even as he’s developed into one of the league’s most valuable players. His accuracy has improved but he’s still going to be variance from down to down. Last season, he was 19th among quarterbacks in overall on-target rate per SIS.
This is where that sliding scale for accuracy kicks in. Quarterbacks like Allen and Mahomes don’t have to be pinpoint on every throw because they have the ability to make up for it with bigger throws often enough to keep up the value.
We don’t have to use this to retcon who Allen was as a prospect, but understanding Allen’s baseline can give us a better guide to how we view the “raw” prospects who come in the future. Allen wasn’t starting from zero as some of the discourse around him would have suggested.
There have been other quarterbacks who struggled with accuracy in college but became high draft picks with early success.
We could look at Justin Herbert, who ranked in the 44th percentile on short throws, 64th percentile on intermediate throws, and 22nd percentile on deep attempts. Herbert was in a college system that didn’t completely take advantage of the quarterback’s strengths, so he stands as the projection hope that a new system can help out a talented quarterback.
The biggest leap in accuracy for a first-round quarterback might come from Daniel Jones. Jones was 22nd percentile on short throws, 11th percentile to the intermediate level, and 36th percentile on deep attempts. It took until Year 4 but the Giants were able to get the most out of Jones by completely eliminating what he did poorly and setting him up to succeed with low aDOT passes. Jones was just 25th in EPA per play over the past three seasons, per TruMedia, but finished 11th in his first year with Brian Daboll and a new offense.
On the other side, plus accuracy isn’t an automatic indicator of quarterback value. Mitchell Trubisky had a 99/92/65 percentile slash line across the three passing depths in his final college season. But accuracy isn’t quite as valuable if there aren’t enough complementary-or-better traits to prop it up — whether it be something like arm strength or processing speed.
2023 Quarterback Class
That brings us back around to the 2023 quarterback class. Here’s where they stand in the rates of where they threw, their on-target rates, and completion percentages plus how they stack up in this class.
If we focus on the likely first-round picks and zoom out on their standing among that full 100-prospect sample, we get a bigger picture. The top two projected picks, C.J. Stroud and Bryce Young, fall into acceptable ranges while there is one slight area of concern for each. Those don’t necessarily limit what each quarterback has been able to do with relatively high completion percentiles.
Stroud’s top concern is the short accuracy, which has been a thing for Ohio State quarterbacks. Ohio State quarterbacks routinely have the highest rates of passing attempts past the line of scrimmage, so they’re not often helped out with screens. That can make those shorter throws more difficult. Stroud was 54th in on-target rate among college quarterbacks on short throws outside the numbers, per SIS.
Young works the short and intermediate levels of the field — much better than the typical shorter quarterbacks — but his deep accuracy was not as strong. However, Young lived in the intermediate area (25% of throws) and did not throw deep all that often (11.6%). When Young’s deep passes were on target, they were often completed.
Richardson and Levis would be two of the least accurate passers selected in the first round over the past few seasons.
2023 Projected First-Round Quarterbacks, On-Target Rate & Completion Percentage By Depth Percentiles
data per Sports Info Solutions, based on quarterback prospects since 2017
Betting on the traits of Richardson and Levis would be a leap bigger than the likes of Allen, Herbert, and even Jones.
Richardson’s biggest red flag comes on short passes. That’s evidenced in his tape where balls either skip before or sail over the head of his intended receiver. This was also an area of the field Richardson wasn’t asked to throw to that often. In 2022, just 28.2% of Richardson’s pass attempts traveled between 1-10 air yards. That’s easily the lowest not just of the group of first-round quarterbacks, but of the 100 quarterback prospect sample since 2017.
The NFL is a shorter league than college when it comes to pass attempts. Last season, 70.5% of throws in the NFL had an aDOT of 10 yards or lower according to TruMedia. Nearly half of all passes in the league were between 1-10 air yards. Justin Fields had the lowest rate in the league with 39% of his passes to that depth range.
Fields is an interesting comparison for Richardson. In Fields’s final season at Ohio State, 49.5% of his pass attempts traveled between 1-10 air yards but it was the area he struggled the most with in accuracy (third percentile on-target rate). Fields has since pivoted from the short passes — especially after a failed rookie season focused on quick game — and has used his legs as a replacement for those shorter throws.
For Richardson, it’s a question of whether he can’t or if he just wasn’t asked to rely on throws to that area. While it was a place he struggled, there are also examples like the 2022 game against Utah where he mixed in his ability to bounce off defenders with more controlled in-rhythm passing and an aDOT of just 6.8 yards.
There are also questions about Richardson’s footwork, which appear to be the cause of a lot of the errant throws. The intention typically looks clear for Richardson passes, even if the full body doesn’t align. He’s also not a quarterback who works himself into more trouble by panicking in the pocket. Richardson was only sacked on 10.1% of plays when he faced pressure, which was the 11th-lowest rate among college quarterbacks in 2022. His pocket management is a plus while some of the connective tissue to other parts of his game need fixing.
Inexperience also stands out. Richardson has just one year as a starter and under 400 career pass attempts as a college quarterback. It’s not as if he’s been a long-tenured starter maxed out in experience and scheme. That leads us to Levis.
Levis impressed during the 2021 season at Kentucky in a Sean McVay-inspired wide-zone system under offensive coordinator Liam Coen. In 2022, Coen went back to the Rams to serve as offensive coordinator and a few skill position players left for the NFL. In 2022, Levis struggled under Rich Scangarello, who had spent the 2021 season under Kyle Shanahan as the quarterbacks coach in San Francisco.
It’s not just that Levis’s production dipped, it’s how much it fell off. The 2021 tape showed promised but that was also aided by an offensive system that has done a fantastic job of opening things up for quarterbacks. That system also didn’t ask him to push the ball down the field often with almost 30% of his passing attempts coming at or behind the line of scrimmage.
In 2021, Levis was an exceptional intermediate thrower but still struggled to throw deep. Then in 2022, his intermediate accuracy completely tanked while throws to other parts of the field didn’t make up for the loss.
Will Levis Accuracy Percentiles, 2021-2022
data per Sports Info Solutions
|1-10 OT Percentile
|11-19 OT Percentile
|20+ OT Percentile
While Levis has some of those throws on tape, it’s concerning how far his play fell as the circumstances around him changed. He was also a quarterback who made some problems worse. Levis was sacked on 30.4% of his pressured dropbacks in 2022, which was the fifth-worst rate among 121 qualified quarterbacks in college football season. However, he also averaged 7.8 yards per attempt under pressure, which was the seventh-highest rate.
Projecting these quarterbacks to the next level takes a leap of faith in taking what works and eliminating what doesn’t. At the combine, Indianapolis Colts head coach Shane Steichen talked about the need for accuracy as one of the most important traits in a quarterback, but also one that can be improved. Even some of the prospects that have paved the path for the traits-based bets have provided a higher baseline than what we’ve seen from Richardson and Levis.
“I think you can [fix accuracy],” Steichen said. “You can help accuracy with mechanics and part of it is the scheme you put him in. Don’t make him think too much and simplify the offense to build it around the quarterback.”
Stehcien and the Colts are a team that could be looking for a quarterback and might be in the Richardson-Levis zone with the fourth overall pick. It’s something they’ll have to consider and what’s fixable will be a gamble that comes along with the selection. It’s a swing teams desperate for quarterback play could want to take but it’s one only a few might be able to handle well.