It's always interesting watching mock drafts evolve over a long period of time. A year out from the NFL draft, we already know most of what we need to know about most of the players who will hear their names called. In his May 2022 way-too-early rankings for 2023, Mel Kiper Jr. listed Alabama's Will Anderson Jr. and Bryce Young, Ohio State's C.J. Stroud and Georgia's Jalen Carter as his top four prospects; in his most recent rankings, all four were still in his top five. Six of last May's top 10 were still in Kiper's top 10.
There are always shifts, however. Some players have disappointing seasons and fall in the mocks. (Sorry, Kayshon Boutte.) Some veterans finally put all the pieces together and rise. (Hello, Joey Porter Jr.) And almost inevitably, more and more quarterbacks coalesce toward the top as bad teams with top picks begin talking themselves into trying to win the QB lottery. Last May, Kiper had two quarterbacks in his top 10. Now? Four in his top six.
We know why this happens, of course. Quarterback is the most important position in team sports, and as Bill Barnwell recently wrote, there is a pretty big financial incentive when it comes to building around a quarterback on a rookie contract. (Financially, it almost doesn't make sense to draft anything but a quarterback, left tackle, edge rusher or receiver in the first round.)
Still, just because a QB gets picked in or near the top five, that obviously doesn't mean he'll be a top-five performer. In the past five drafts, 18 quarterbacks have been selected in the first round -- 12 have gone among the top 10 picks, seven among the top three. Of this batch of 18 first-rounders, only three have ranked among the top 12 quarterbacks, per Total QBR, over the past five NFL seasons: Josh Allen (No. 7 pick in 2018), Lamar Jackson (No. 32 in 2018) and Justin Herbert (No. 6 in 2020). All had statistical flaws on their résumé, all landed with teams that proved equipped to maximize their talent and none went in the top five.
Certain recent top picks are certainly trending in the right direction -- Joe Burrow and Trevor Lawrence, to name two -- but some either have flamed out (Josh Rosen) or are well on their way to doing so (Zach Wilson, Baker Mayfield, Sam Darnold). The jury remains out on others, due to either injury (Trey Lance), injury and stagnation (Kyler Murray), or a small and confusing sample size (Justin Fields).
We're never as good at predicting player success as we think we are, and as I wrote a couple of years ago, in the game of nature vs. nurture, we always underestimate the importance of nurture -- i.e. the competence of the team selecting the player. (It probably isn't a total coincidence that two of the "on their way toward flaming out" quarterbacks mentioned above were drafted by the New York Jets, for instance.)
This piece therefore isn't about which top QB prospect from the 2023 draft class will succeed or which I would personally take with the No. 1 pick. (I'm sure my preferences will become semi-obvious, however.) Instead, it's about simply setting the table. What can we learn from each quarterback's statistical profile? What are each player's noteworthy red flags and fatal flaws that teams will need to understand and accommodate?
First, the numbers.
From a purely statistical standpoint, three quarterbacks from Kiper's top 11 stand out: Young (the 2021 Heisman winner), Stroud (a two-time top-four Heisman finisher) and Georgia's Stetson Bennett (a two-time national champion). They were the most successful college quarterbacks on the planet, so their superiority makes sense.
Tennessee's Hendon Hooker, No. 5 on Kiper's list, stands out pretty similarly, and with quite a few more dropbacks than anyone in the top four. He has a couple of red flags going against him -- he's already 25, and he's coming off a November ACL tear -- and while he is pretty fast in open space, he was the worst QB in this sample when it comes to escaping pressure. His downside is almost as clear as his upside.
Meanwhile, if you're looking solely at stats, the fact that Kentucky's Will Levis and Florida's Anthony Richardson grade out as top-10 picks doesn't make a lot of sense. Levis lacked explosiveness, Richardson lacked efficiency, both threw interceptions more frequently than the other prospects, and both showed up near the bottom in terms of Total QBR. Both bring a level of mobility to the table, but Total QBR does take that into account a bit.
The general rule is that pressure and out-of-pocket improvisation numbers are a bit more volatile and that, to the extent that college stats are reliable, those generated from a clean pocket with no pressure might be a bit more telling. So let's look at those numbers.
Note: Looking at the opponent-adjusted Total QBR isn't an option once you drill into situational data.
Stroud's numbers leap off the page in virtually every category -- second-highest completion rate, most yards per completion, highest touchdown rate, third-lowest INT rate, highest raw QBR. He also had the best receiving corps, of course, throwing to first-rounders Chris Olave, Garrett Wilson and Jaxon Smith-Njigba in 2021 and future high draft picks Marvin Harrison Jr. and Emeka Egbuka in 2022. Hooker, Young and Bennett all thrived in clean-pocket situations, too, and among the lower-ranked prospects on the list, UCLA's Dorian Thompson-Robinson absolutely thrived in these situations; if opponents couldn't hurry him, he picked them apart.
Again, Levis and Richardson struggled here. Their status as soon-to-be top picks stems totally from what they could deliver in the future, not what they delivered in college. And if there's encouragement to be derived, it comes from Josh Allen's college stats. Before becoming a star for the Buffalo Bills, Allen posted wholly mediocre numbers as a Wyoming Cowboy. His pocket-no-pressure numbers in Laramie: 62% completion rate, 12.0 yards per completion, 80.5 QBR. It took two full years of struggles at the NFL level, plus nearly perfect support from the Bills in terms of improving his supporting cast and giving him time to grow, but he did eventually become a star after posting numbers not too dissimilar from Levis' or Richardson's.
What about the supposed big arm the NFL requires? Without getting too far in the weeds (and without diluting the sample sizes too much), by looking at how each player threw on specific big-armed routes, let's look at who had the most success throwing the ball more than 20 yards downfield.
Again, Stroud's numbers are incredible, as are what Hooker produced from Josh Heupel's cheat code of an offense. (Meanwhile, if you were thinking more highly of Thompson-Robinson because of the clean-pocket stats, these stats are a bit of a wet blanket.)
Once again, Bennett's numbers pop too, despite the fact that his best weapons at UGA were tight ends and a slot receiver (Ladd McConkey) catching a lot of his passes near the line of scrimmage. A former walk-on and a player thrust into the starting role at Georgia only after a number of injuries, Bennett's physical measurables (5-foot-11, 192 pounds) are holding his draft stock back considerably just as they (initially) held him back on the Bulldogs' depth chart. But his stock is rising a bit, and I'm fascinated to see what kind of opportunities he finds.
Again, Levis comes up short here, though we start to see a split between him and Richardson. No matter how strong Levis' arm might be, he was the most heave-and-hope deep-ball passer of this bunch, and he was picked off far more on these passes than any other prospect.
Josh Allen on deep balls at Wyoming: 38% completion rate, 18% TD rate, 5.5% INT rate. Levis doesn't clear even that low bar, though Richardson does.
The stats above certainly tell us who's getting drafted because of known production and who's getting drafted based on potential. And while there's a loose correlation between good college and pro numbers (and players almost never exceed their college rate stats in the pros), Allen's success, combined with the struggles of players like Mayfield, reinforces the idea that college production isn't the end-all, be-all for a prospect facing a degree-of-difficulty jump.
The top four names -- Young, Levis, Stroud and Richardson -- are virtually guaranteed to go in the top 10. Let's take a deeper dive into these four. The teams drafting these players will determine a healthy percentage of their success based on supporting casts and general developmental competence. And while we don't know who among the top prospects will succeed or fail, for those who fail, we probably know what traits, if unaccounted for, will sink them. Let's talk fatal flaws!
Jump to a quarterback:
Levis | Richardson
Stroud | Young
Bryce Young's fatal flaw: He's small
Bryce Young's radar chart. The discolored blob in the middle shows the national averages for each category. A quick key for less familiar stats: Pass Disp = average pass displacement, or how far from the receiver's center of mass passes were on average; Chunk % = percentage of plays gaining at least 15 yards; Explosive % = percentage of plays gaining at least 30 yards. (Source: StatsBomb.)
Of the four players we're looking at, Young was the only prospect who wasn't below average in at least one of the categories in the radar chart above. He was a revelation at the collegiate level, a five-star prospect (No. 5 overall in the 2020 recruiting class) who won the Heisman in his first year as a starter, then produced nearly identical numbers with a more flawed supporting cast in 2022.
2021: 67% completion rate, 13.3 yards per completion, 47-to-7 TD-to-INT ratio, 6.3% sack rate, 87.2 Total QBR
2022: 65% completion rate, 13.6 yards per completion, 32-to-5 TD-to-INT ratio, 4.2% sack rate, 86.2 Total QBR
Granted, saying Young's supporting cast was more flawed doesn't mean it was bad. He still played for Alabama, after all. But he took the field in 2022 without left tackle (and top-10 pick) Evan Neal and five of 2021's top six receiving targets, and he faced more awkward downs and distances in part because of penalties. Still, he almost perfectly replicated his Total QBR numbers and actually lowered his sack rate.
If Young were 6-foot-2, 215 pounds, he would be a no-brainer No. 1 pick. But he's not. He measured at 5-foot-10, 204 pounds, nearly the same as Kyler Murray (5-foot-10, 207) but without Murray's elite speed and mobility. He's fast but not Murray fast. I've seen a Drew Brees comp pop up more frequently of late -- and a half-Brees, half-Murray split isn't entirely inaccurate -- but even Brees measured in at 6-foot-0, 213. Brees was one of the best pocket passers in NFL history; will 2 fewer inches of height make a big difference?
Young has just about the quickest release you'll ever see and seemed to make decisions at hyperspeed in college. His deep-ball passing was merely fine, and while his out-of-pocket numbers were easily the best of anyone in this draft class, (a) again, those stats are more volatile, and (b) out-of-pocket passes accounted only for about one-sixth of his overall sample. He is quick enough that he should avoid taking too many huge hits (which balances out some of the size factor), and the extra slivers of pocket adversity that he faced in 2022 might end up coming in handy. But his size is nearly without precedent for a pocket passer, and if that ends up scaring off Carolina at No. 1 -- or if it turns out his ceiling is lower than that of other prospects once they actually get reps at the pro level -- those measurable limitations will likely explain why.
(By the way, it doesn't appear Carolina is being scared off. As of Monday night, Caesars lists Young's odds of going No. 1 at -1000, equivalent to more than a 90% chance. He's probably going No. 1.)
Will Levis' fatal flaw: He's going to take so many unnecessary hits (and throw unnecessary picks)
Levis' radar chart. (Blue = Levis' stats, black = national averages.) His completion rate was well above average, and he did make occasional huge plays, but his negative play rates were far worse than normal, and for someone with supposed plus-athleticism, his per-carry rushing numbers were dreadful. (Source: StatsBomb.)
We've learned a lot about ceilings and floors in recent years. In Josh Allen, Buffalo drafted a quarterback whose production suggested a terribly low floor and whose athletic ability suggested a ridiculously high ceiling. It took two full years of struggling and learning for Allen to click at the pro level. Meanwhile, in the same draft, Baker Mayfield -- 6-foot-1, 215 pounds, a 4.8 40 and with less perceived athleticism -- went to Cleveland at No. 1 and improved his Total QBR from 51.2 (23rd) to 54.4 (19th) to 65.5 (10th) over his first three seasons. It's been a free fall ever since. Mayfield was slowed by injury early in 2021, but his stats have never recovered. He ranked 27th in Total QBR in 2021 (40.9) and 31st (24.5), dead last among eligible QBs, in 2022.
Opponents figure you out and force you to continue improving over the course of your pro career, and even with early success, a lot of pros hit their heads on their respective ceilings pretty quickly. Whoever drafts the 6-foot-4, 229-pound Levis will be desperately hoping that he has an Allen-esque ceiling because his college stats suggest his floor is lower than that of any top-level prospect.
Levis' stats do come with caveats. For one, his supporting cast wasn't nearly as strong as Young's, Stroud's or even Hooker's, especially last fall. In 2021, his first year in Lexington after transferring from Penn State, he benefited from an ultra-efficient run game and the presence of second-round receiver Wan'Dale Robinson and ranked 18th in Total QBR. But in 2022 the Wildcats' run game regressed, and the receiving corps was led by a pair of freshmen; Levis battled a foot injury as well and plummeted to 61st in Total QBR. After averaging 6.0 yards per non-sack carry with a 5.3% sack rate in 2021, he averaged 3.4 and 11.3% last fall.
Still, he was interception-prone in both seasons, and while his mobility was severely hindered in 2022, his escapability numbers were never particularly good, at least by the standards set by either Richardson (the other Allen-type QB in this draft) or Allen himself.
Scrambles per dropback (2021-22 unless otherwise noted): Hendon Hooker 11.4%, Josh Allen (2016-17) 9.4%, Will Levis (2021 only) 9.0%, Anthony Richardson 8.7%, Bryce Young 6.2%, Will Levis (2022 only) 4.8%, CJ Stroud 3.5%
Yards per scramble: Richardson 12.5, Young 9.2, Allen 7.9, Hooker 7.5, Levis (2021) 6.3, Stroud 6.2, Levis (2022) 4.2
Sacks per pressure: Richardson 11.1%, Stroud 12.2%, Young 18.3%, Allen 18.6%, Levis (2021) 20.4%, Levis (2022) 32.1%, Hooker 33.8%
If athleticism shows us a player's ceiling, Levis had neither the deep-ball success that suggests a big, accurate arm, nor the scrambling and sack avoidance that suggests excellent mobility and instincts. His 2021 mobility numbers were certainly stronger than 2022's, but they were never Allen's -- he scrambled less frequently and less effectively, and even in 2021 his ability to escape pressure wasn't quite at Allen's level.
Meanwhile, he made more passing mistakes than any of the other top prospects. Maybe he will eventually thrive, in the mold of Allen, but he will almost certainly make a ton of mistakes in the meantime. Whoever drafts him should prepare for major struggles in 2023.
C.J. Stroud's fatal flaw: We have no idea how he will handle steady pressure
Stroud's radar chart. Nearly perfect in the explosiveness categories and as low as anyone in the mistake categories. The only category in which he didn't thrive was pass displacement, which would shock anyone who watched him play. (Source: StatsBomb.)
Stroud's displacement numbers shocked me anyway, as my eyes told me his accuracy was by far the best part of his game. He completed 69% of his passes in 2021 and '22, and while he had plenty of easy pitch-and-catch opportunities in there, he also threw the most successful deep balls.
So what gives? Why were his displacement levels seemingly poor? Let's look at another chart to search for the answer.
Stroud's pass placement relative to his receiver. Completions are filled in, and incompletions are empty rings. Shorter passes are more purple, longer passes more yellow. Stroud's receivers caught a lot of shorter passes that were either high or away from their body to the right or left. (Source: StatsBomb.)
Was Stroud misfiring a bit and getting bailed out by amazing receivers? Or was he leading his receivers into open-field catch-and-run opportunities? Watching Ohio State live, I would guess the latter. Jaxon Smith-Njigba was dynamite on hooks and shorter routes in 2021, for instance -- he averaged an incredible 16.8 yards per catch on passes caught between six and 10 yards downfield that season -- and for as good as Smith-Njigba is, I'm guessing he isn't capable of those numbers if he's not catching passes in stride. And in 2021 and '22, Ohio State receivers ranked seventh nationally in drop rate; again, they're spectacular, but Stroud seemed to put them in a perfect position to succeed.
What happens if Stroud doesn't have a superior receiving corps? He escaped pressure well in Columbus, but does that continue if he's facing far more pressure? And what happens when he's taken out of the most beautifully spaced offense in college football?
The duo of head coach Ryan Day and offensive coordinator Kevin Wilson crafted a college offense that combined both extreme matchup advantages and perfect sideline-to-sideline spacing. It has meant six straight seasons of finishing in the offensive SP+ top five, but it has also meant that recent Ohio State quarterback prospects -- Justin Fields and the late Dwayne Haskins -- entered the pros having faced minimal pressure and adversity. Both were used to having far more processing time than they encountered at the pro level.
Acknowledging that pressure numbers can be unstable, it is probably worth noting that Stroud's pressure numbers were far better than those of either Haskins or Fields.
Ohio State quarterbacks under pressure
Stroud (2021-22): 5.8 yards per dropback, 52% completion rate, 2.2% INT rate, 12.2% sack rate
Fields (2019-20): 3.8 yards per dropback, 49% completion rate, 2.5% INT rate, 32.5% sack rate
Haskins (2018): 4.8 yards per dropback, 58% completion rate, 2.4% INT rate, 18.1% sack rate
Stroud escaped pressure better, and while his completion rate wasn't as good as Haskins', he averaged 14.7 yards per completion under pressure to Haskins' 11.4. He was far better at keeping his eyes downfield under duress. Stroud had randomly awful games under pressure -- in Ohio State's four losses in his two seasons, he completed 45% of his passes under pressure with two interceptions, 11 sacks and a paltry 2.0 yards per dropback. That two of those losses were to rival Michigan has forever dampened his legacy in Columbus, but in 21 other games, including six against ranked foes, he was nearly perfect under duress. If his future pro team can keep him at least reasonably comfortable, he could shine quickly ... as long as his receivers can catch balls in stride.
Anthony Richardson's fatal flaw: He might never be accurate enough
Richardson's radar chart. He was a master of mobility, and he created solid big-play numbers without a wild number of interceptions. But his efficiency and accuracy numbers lacked significantly. (Source: StatsBomb.)
It appears the top six or so players picked in the 2023 draft will include two statistically masterful quarterbacks and two extreme works in progress. Two prospects with high floors and potential for most immediate success versus two players with high ceilings based on their athletic ability and, to put it politely, room for growth in terms of actual passing.
If you're looking for a long-term project, however, I'm not sure why you wouldn't lean toward the guy who is three years younger, 15 pounds larger, faster (Richardson recorded a 4.43 40 time at the combine), more elusive, more explosive and has taken far fewer hits to date.
If you're looking for the closest thing to a Josh Allen model in this draft, it seems Richardson checks a lot more boxes than Levis. He is the most dangerous runner of any of these prospects -- he averaged 12.5 yards per scramble (an average more than 3 yards higher than that of any of the other top prospects or Allen) -- and perhaps most impressively, he escaped pressure better than anyone, turning only 11.1% of pressures into sacks. A 6-foot-4, 244-pound SEC product, he has naturally drawn comparisons to Cam Newton, and while I understand the comp, it actually sells Richardson short athletically: Newton ran a 4.6 40 at the combine and recorded a vertical jump of 35 inches to Richardson's 40.5.
Of course, that comparison also sells Newton far short as a passer.
Cam Newton in 2010: 90.0 Total QBR, 66.1% completion rate, 15.4 yards per completion, 10.7% TD rate, 2.5% INT rate, 7.6% sack rate
Richardson in 2022: 71.2 Total QBR, 53.8% completion rate, 14.5 yards per completion, 5.2% TD rate, 2.8% INT rate, 3.5% sack rate
Newton was a year older in 2010 than Richardson was last fall, and for all we know Richardson might have improved in a similar manner had he returned to Gainesville in 2023. We'll never know, but what we do know is that he was not effective or accurate enough as a passer in 2022.
Richardson's passing bins chart. Green = better than average, and the size of each bin represents the proportion of passes thrown there. His big arm did not create above-average completion rates along the sideline, and while his numbers on shorter throws weren't horrible, they weren't particularly good either. (Source: TruMedia.)
The only area where Richardson managed even standard-level completion rates was on quick hitters behind the line of scrimmage. Among the top prospects his footwork was easily the least consistent, and while the team picking him inevitably will employ a quarterbacks coach who has said "Oh, I can fix him for sure" to the general manager, QB coaches are only so effective in that regard. Instincts are difficult to mold.
Worse, Richardson wasn't necessarily able to rely on his legs when the pass wasn't working -- a Newton specialty -- at least not against good defenses. He faced four opponents that finished in the top 20 of defensive SP+ (Georgia, Kentucky, Texas A&M and Missouri) and played well against only A&M. His numbers in these four games: 57-for-114 (50%) for 681 yards (11.9 per completion) with four TDs, three INTs and only 146 total rushing yards at 5.0 per carry.
I definitely understand the draw here, and if you're aiming to land the lump of clay with the highest possible ceiling, your gaze should probably shift toward Richardson, not Levis. But the accuracy numbers are scary, and there's no guarantee they'll ever improve.