Which Quarterback in the 2023 NFL Draft Class Is Statistically Superior?

After a one-year hiatus, high-end quarterback prospects are back, baby.

In last year's crop of NFL draft hopefuls, none of the top five quarterback prospects ranked in even the 80th percentile of my pre-draft model. And, not surprisingly, only one of those guys wound up being drafted in the first two rounds.

If you wanted elite quarterback prospects, it was a dud of a season.

But for 2023, not only are more guys getting first-round buzz, but more guys deserve that hype. And with a handful of quarterback-hungry teams near the top of the draft, things could get really fun.

But, who, exactly, should those teams covet most?

The answer, as always, is complicated. There's no one-size-fits-all approach to scouting quarterbacks; otherwise, we wouldn't see them bust at the rate they do.

Today, we're going to dig into just one aspect of the evaluation: the numbers. Based on simply each guy's statistical resume, how should they stack up? And which guys fit the mold of previous successful first-round picks?

Obviously, this is just one piece of the puzzle. Film matters, and scouts are broadly very good at what they do, despite the misses. But in looking back at previous drafts, you can see that statistical resumes -- even if they're not perfect -- do point you in the right direction.

Let's start with parsing that out and then dig into this year's class.

Why the Stats Matter

In order to flesh this out, we need a measure of what success looks like in the NFL. There are a bajillion ways you could do this. Here, we're going to look at numberFire's Total Net Expected Points (NEP).

NEP is numberFire's expected points model, and Total NEP includes the expected points a quarterback adds as a rusher on top of his output as a passer. Basically, we want to measure who moved the needle most in the key aspects correlated to scoring points and, thus, winning games.

Since 2010, 40 quarterbacks have been first-round picks. Of those 40, 35 have either been in the league for three-plus years or have already recorded a top-10 season in Total NEP. It's still a small sample on guys like Justin Fields and Kenny Pickett, so we'll omit them for now and focus on just the other 35.

Among that group of 35 quarterbacks, 12 have finished in the top 10 in Total NEP in at least one-third of their qualified seasons (minimum 200 drop backs). These are your elite passers: Patrick Mahomes, Joe Burrow, Josh Allen, Lamar Jackson, etc. If we want to focus on identifying guys with the ability to win you games single-handedly, this is the group we want.

Here's the average statistical resume of those 12 quarterbacks coming out of college versus the other 23 first-rounders in our study. Here, "age" refers to the player's age the day they were drafted. "Pick" is where they were taken in that year's first round. "Games" is the number of games in college in which they attempted at least 10 passes. "QBR" is their Total QBR from ESPN in their final collegiate season. Finally, AY/A is their adjusted yards per attempt in their final collegiate season.

As you can see, the successful first-rounders were younger, more experienced, and more efficient in college, and they typically went higher in the draft. This furthers the point that scouts are good at what they do. But it also shows that a player's collegiate resume can tell you plenty.

Based on this information, prior to last year's draft, I decided to create a prospect model that weighs these factors, puts them in a blender, and spits out a rating for each quarterback. Through last year's class, my model includes 219 quarterbacks invited to the combine since 2010.

There is one version -- the one we'll lean on today -- that doesn't factor in draft capital. We don't have a firm answer on draft capital yet while everything in the pre-draft model is already set in stone.

So, how do our 12 high-end quarterbacks compare to the others when we plug them into the model? It's not surprising given the data outlined above, but the gap here is pretty massive.

Of those 12 hits, only one quarterback ranked lower than the 80th percentile in the pre-draft model. That was obviously Allen, who was in the 29th percentile. The model whiffed on him big time. But on the whole, successful first-round picks often graded out better in the pre-draft model than their less successful peers.

That 80th-percentile range did seem to be where things fell off. Of our 35 qualified first-rounders, 20 of them were in the 80th percentile or higher of the pre-draft model. Sixty percent of them have at least one top-10 season in Total NEP, and almost all of that group is in that range regularly.

Allen, of course, is the one outlier from the bottom group. The others with a single appearance in the top 10 in Total NEP are Daniel Jones, Carson Wentz, and Ryan Tannehill. Those are the biggest success stories in that bucket other than Allen.

That's not to say there are no misses in the group with a great model rating. Zach Wilson's up there. So are Blake Bortles, Johnny Manziel, and even Paxton Lynch. Grading out well in the model doesn't guarantee success, nor does having a poor grade guarantee failure; the odds of hitting just increase as the model rating goes up.

So, that's the lens through which we'll view this draft. We'll dig into whose resume is best based on this criteria, discuss why that's the case, and outline what that means for their outlook in the pros.

For this exercise, we'll be focusing specifically on the quarterbacks given a grade of 75 or higher by Scouts Inc. As mentioned, scouts are good at what they do, and if they decide a player isn't worth a first-round pick, the odds the player pans out are low. This means the five quarterbacks we'll look at today are the top five in those rankings: Bryce Young, CJ Stroud, Will Levis, Anthony Richardson, and Hendon Hooker.

Of that group, whose collegiate resume is most impressive? It's the lone non-SEC passer on the list.

1. CJ Stroud, Ohio State

Model Rank: 93rd percentile | Age: 21.6 | Games Played: 25 | AY/A: 10.9 | Total QBR: 87.7

As a reminder, the data listed here is for Stroud's age on the day of the draft, his games with 10-plus pass attempts, and his final-season efficiency stats.

He doesn't check all the boxes with less experience than you'd like; everything else lines up gorgeously for Stroud, though.

Stroud finished the year ranked third in the nation in Total QBR. That was actually a step back for him after he led the nation with a mark of 91.6 last year when he had Garrett Wilson, Chris Olave, and a healthy Jaxon Smith-Njigba. Obviously, Marvin Harrison Jr. is a nice consolation prize, but it was encouraging to see Stroud duplicate his success even after losing multiple first-round pass-catchers.

The value of a stat like QBR is that it accounts for the schedule a player faced. Stroud's was especially tough. He played three games against top-10 defenses by Bill Connelly's SP+ and five against the top 15. He's the only guy in our sampling with more than two games against top-10 defenses. For comparison, the highest-ranked opposing defense Young faced was 15th-ranked Texas in Week 2.

Although Stroud struggled in a couple of his tougher games, his broad sample was very good. Stroud posted a 9.7 AY/A against top-50 defenses by SP+. In the Georgia game, specifically, Stroud flashed his ability to create and use his legs, something he didn't do often before then.

That game likely made Stroud a lot of money. While you'd rather not over-weigh just one game, when it's doing stuff like that against the now two-time national champs, we can make an exception.

As mentioned above, the one blemish for Stroud is the lack of experience. His 25 games played is fewer games than 29 of 40 first-round picks since 2010, and a lot of the guys behind him wound up as busts. Manziel, for example, also had 25 games of experience and solid efficiency marks coming off his age-21 season, and we know what happened there.

A positive for Stroud is that we know why he didn't have more experience. When he got to campus, Ohio State already had a future first-rounder in Fields at quarterback. It's not as if Stroud couldn't bust through a thin quarterback room or transferred for more playing time. And when he did get his chance, he shined.

The broad consensus entering the draft seems to be that Stroud is behind Young as the top quarterback on the board. The marks against Stroud -- such as the offense he played and the talent around him -- all make sense. But at least from a pure statistical perspective, Stroud's resume is the cleanest of the group, and he belongs in that discussion as the first guy selected.

2. Bryce Young, Alabama

Model Rank: 87th percentile | Age: 21.8 | Games Played: 27 | AY/A: 9.9 | Total QBR: 83.7

All those glowing words of Stroud should not be taken as a dismissal of Bryce Young; he's also well worth the hype.

Let's go back to our group of 12 quarterbacks at the top who have been in the top 10 in Total NEP in one-third of their seasons. Here's their average resume compared to that of Young. For additional context, the average marks of all first-rounders since 2010 are included, as well.

Young's resume almost exactly mirrors that of past superstars. As with Stroud, you'd like a bit more experience, but with Mac Jones being the roadblock in Young's first season, it's once again easy to forgive.

The most impressive thing about Young is how good he looks when you dig in deeper. In obvious passing situations, he absolutely lit it up.

Specifically zeroing in on 3rd and 6 or longer, Young had 69 drop backs in those scenarios this year when the defense could pin its ears back and focus exclusively on the pass. Young's AY/A in those scenarios actually increased to 12.5.

The other situation when the defense can allocate more resources to stopping the pass is when they're ahead. Young led our group in AY/A in that split, too.

We're dealing with small samples in these splits, so it's important not to fret over them too much. But it's encouraging to see that Young excelled even in adverse situations.

The big question is how Young stacks up compared to Stroud. Stroud has the size advantage, and he grades out better in the model. But Young seems to be held in higher regard by scouts.

Personally, I don't think the split between the two in the model matters much. Sure, you'd rather be higher in a vacuum, but the list of first-rounders between the 80th and 90th percentile in the model contains some of the league's best.

So, sure, we could use this data to say that the model prefers Stroud, and that is accurate. But I think the bigger takeaway is that it views both guys as potential superstars who are worthy of a top-five pick. If you miss out on one, the numbers say the other is a delicious consolation prize.

3. Hendon Hooker, Tennessee

Model Rank: 74th percentile | Age: 25.3 | Games Played: 38 | AY/A: 10.9 | Total QBR: 89.4

Frankly, I was surprised to see Hendon Hooker this high when I ran the numbers. Age is a big factor in the model, and Hooker is one of the oldest prospects in this database. But his efficiency stats were phenomenal, even before his final year at Tennessee. We'll get to his age and his outlook post-ACL repair in a second, but the guy was a legitimate force in college.

Hooker had four collegiate seasons in which he attempted at least 150 passes. He had an AY/A of at least 10.0 in three of them with the lone exception being a mark of 8.6 in 2020. Levis and Richardson combined for zero seasons with an AY/A of at least 8.5.

This was true even against stiff competition. I mentioned earlier that Stroud had a 9.7 AY/A against top-50 defenses by SP+. Hooker actually topped that.

And that's Hooker's AY/A, which doesn't include his contributions as a rusher. He did plenty there, too.

Obviously, these numbers won't account for the quality scheme Hooker played in or the talent around him. Those are fair criticisms. It's worth noting, though, that Hooker didn't juice up his efficiency stats by dinking and dunking. He actually had the highest aDOT of this group, according to PFF.

So, those are the pros with Hooker. Let's dive into the negatives.

To me, at least, the age is a bigger red flag than the ACL. Not counting this year, 22 quarterbacks invited to the combine since 2010 have been 24.0 or older on the day of the draft. Half went undrafted, and only Brandon Weeden went in the top 99 picks. The most successful pro of that group is Case Keenum, but even he had a ceiling.

A big part of the reason that age matters is that more talented quarterbacks declare for the draft earlier while others remain in college. The fact that it took this long for Hooker to enter the draft means the NFL must not have been clamoring for him earlier, and that matters.

There are two rebuttals to this. First, Hooker is better than most of the other older prospects. He and Keenum (78th percentile) are the only guys in that group to rank higher than the 65th percentile by the model. Hooker's final-year QBR of 89.4 is the best in that group and factors in what he added as a rusher and the competition he faced.

Second, Hooker wasn't sitting for all those other years. He was playing, and he was putting out good efficiency marks in the process. He had a 10.6 AY/A in his age-21 season. It came on just 162 attempts, and Virginia Tech leaned heavily on his rushing abilities. But it's better than nothing.

As for the ACL, it likely guarantees that Hooker won't carry huge draft capital, which lowers his odds of hitting (both due to the value of what scouts think and the decreased incentive to force him onto the field). It's a bummer given how good he was in college; you'd love to see a guy like this get a shot.

It wouldn't be a huge shock to see Hooker be relevant at some point in the NFL, though. Other non-first rounders in the 70th percentile or higher in the pre-draft model include Jalen Hurts, Russell Wilson, Dak Prescott, Geno Smith, and Jimmy Garoppolo. If nothing else, Hooker's an interesting prospect to think about despite the warts around his age and his knee.

4. Anthony Richardson, Florida

Model Rank: 36th percentile | Age: 21.9 | Games Played: 14 | AY/A: 7.6 | Total QBR: 70.6

This is where things get interesting. Neither Anthony Richardson nor Will Levis fits the mold of previous successful first-round picks. What does that mean for their outlook?

Between the two, Richardson grades out better, sitting in the 36th percentile. If he goes in the first round, it'll be the ninth-lowest mark of any first-rounder since 2010. Unless Levis also goes in the first, of course.

Two things ding Richardson: his low QBR and his lack of experience. Let's dig into both separately.

Starting with the QBR, it's important to remember it does include what Richardson did as a rusher. And, whew buddy, did he do a lot of sicko stuff there.

There's a lot of value in that. And when you think about a player like Josh Allen, he was able to exceed pre-draft expectations in large part due to his athleticism. It's a big mark in Richardson's favor.

But again, QBR includes that, and he still finished just 36th in the nation in that stat.

Richardson's 14 games with double-digit pass attempts would also be the fewest of any first-round pick since at least 2000. And it's not as if he's insanely young, either; he's actually a hair older than both Stroud and Young but with one less year as a full starter.

Richardson struggled most as a passer when backed up in 3rd and long. As referenced above in the section on Young, Richardson's AY/A on 3rd and 6 or longer was just 4.0, more than 2 yards lower than any of our other top-five passers.

A lot of those incompletions were throwaways when receivers couldn't get open, so it could very well be on the pass-catchers. It's just noteworthy how far Richardson was behind the pack.

There is a positive in that same table, though. Richardson took a sack on just 6.3% of his drop backs in those obvious passing situations. His full-season sack rate of 3.8% ranks second in this group behind just Stroud.

Part of this was thanks to Richardson's athleticism. But it also speaks well of his ability to create when things break down.

Again, there's a lot of value in that, and it's encouraging to see Richardson doing stuff like that with such little experience.

When you consider the athleticism and the ability to avoid sacks, you can understand why an NFL team could talk itself into Richardson early. He's still young enough, and he has high-upside traits. His statistical resume just deviates heavily from that of previous successful first-round picks, meaning you're banking on an outlier trajectory.

5. Will Levis, Kentucky

Model Rank: 9th percentile | Age: 23.8 | Games Played: 29 | AY/A: 8.3 | Total QBR: 60.4

If Levis winds up going in the first round in April, he'll be the lowest-ranked first-rounder in my model since 2010. Ryan Tannehill (10th percentile) is the only other guy ranked lower than the 25th percentile. Here's why that's the case.

The big thing is the QBR. Levis ranked 61st in the nation there last year. If he goes in the first round, it'll be the fifth-lowest final-year QBR since 2010, ahead of just Jordan Love, Jake Locker, Blaine Gabbert, and Josh Allen. More on Allen in a second.

Sacks played a big role in Levis' low QBR. No first-round quarterback since 2010 has had a sack rate higher than 8.5%; Levis was at 11.3%. This is a similar criticism to what we saw with Sam Howell and Malik Willis last year, and they both went in the middle rounds.

You could blame this on Levis' injuries this year and a loss of talent in the offseason. After all, Levis' sack rate in 2021 was just 5.9%, roughly half what it was in 2022.

So, let's look at Levis' 2021 QBR. His mark that year was 76.8, much better than this year's output. But even that would still rank just 26th among first-rounders since 2010, right between Paxton Lynch and Tim Tebow.

This career trajectory is very similar to that of Allen. Allen also had a better QBR in his second-to-last season before a step back, and he wasn't ultra-experienced coming out.

Here's the key difference, though: Allen was two years younger coming out than Levis is. Allen was wrapping up his second season in the NFL when he was the age that Levis is entering the draft. It's a lot easier to talk yourself into improvement with a younger player, and Levis would be the third oldest first-round pick since 2010 if he goes that high.

The more realistic comps for Levis (on the optimistic side) are Tannehill and Carson Wentz. Both guys were older and less experienced coming out, and neither had elite efficiency numbers in college. Although neither has been a home-run pick, they at least flashed upside.

In drafting Levis, you have to be banking on that. He has shown moments of high-end play thanks to his big arm; you just need him to do that on a more consistent basis while trimming back on the negative plays. Maybe a clean bill of health can do that for Levis. It's just not something we've seen previously from a guy with his resume coming out of college.