How Bad Are NFL Teams at Identifying the Best QB in a Given Draft?

It seems obvious that Trevor Lawrence is the best quarterback in the 2021 NFL draft. He’s perhaps the best prospect to enter the NFL draft pool in years. Since Lawrence was in high school, scouts have raved about him as god’s gift to football. He has every trait that people covet in modern QBs: arm strength, throwing accuracy, size, mobility, pocket presence, and a decent helping of intangibles like leadership and moxie. Even his hair is elite. He’s the closest thing to a can’t-miss quarterback prospect that could possibly exist, and the obvious choice for the Jaguars with the no. 1 pick.

But for as perfect as Lawrence seems to be, and for as foolish it would seem for the Jaguars to take anyone else, we have to remember a simple truth: More often than not, NFL teams draft the wrong person as the first quarterback off the board. In the last five years alone, teams have traded up to take Mitchell Trubisky and Jared Goff ahead of players like Patrick Mahomes, Deshaun Watson, and Dak Prescott. There have been 51 drafts since the 1970 AFL-NFL merger. Only 20 times (39.2 percent of the time) has the first quarterback selected gone on to lead his QB draft class in career approximate value, according to Pro-Football-Reference.

This, of course, is a simplistic approach. We’re boiling this examination down to one imperfect statistic. (In explaining what AV is, its inventor wrote that it “is not meant to be a be-all end-all metric”; the stat was designed to compare players across eras, and its value measurements tend to be more accurate over a player’s full career than for any individual season. By its nature, this metric benefits quarterbacks who had long careers.) But I’m trying to measure something relatively simple: How often was the first quarterback off the board the one who went on to have the best career in a given draft? Time and again, teams have proved more likely to pick the wrong quarterback than the right one.

On the one hand, this finding feels unremarkable; the draft is widely seen as a crapshoot, with busts as common as stars. And no metric effectively measures the impact that coaching and developmental infrastructure has on a player’s career. On the other hand, this is wild. How can it be that at the most important pick at the most important position, NFL teams are more likely to miss than hit? This goes beyond the failings of individual teams; consensus leaguewide opinion on the top QB prospects—including from media members like me—is often woefully wrong.

Let’s break down what historical draft results tell us about the science of drafting quarterbacks.

The Last Decade Has Been Brutal

It seems like the NFL has improved just about every facet of the passing game over the last 10 years. Quarterbacks are playing better than ever, teams are calling more pass plays than ever, and pass play designs are more inventive than ever. It would naturally follow, then, that the NFL has also gotten better at drafting quarterbacks, especially compared to past eras when the passing game was a novelty.

But that’s not what has happened. Let’s take a look at the quarterbacks taken in the draft’s last nine years:

Only twice in these nine drafts has the first quarterback off the board accumulated the most career weighted AV to date. (If you think Marcus Mariota is a better quarterback than Jameis Winston—a bold take, but not completely ridiculous—it’s down to one of nine.) While it’s still early in many of these players’ careers—maybe by 2030 Joe Burrow will prove himself superior to Justin Herbert—I think the AV measurement generally reflects how these quarterbacks stack up. If I could change one thing about the list, I’d say that Josh Allen, not Lamar Jackson, seems to be the best quarterback in the 2018 class, although Jackson has won an MVP award, so he’s not exactly a bad pick either. Regardless, it doesn’t change the way we look at the Browns drafting Baker Mayfield. Better QBs were available.

Beyond the regularity with which the first quarterback taken in a recent draft class has been the wrong guy, the other thing that jumps out from this table is how late some of the league’s premier quarterbacks have been found. Dak Prescott and Russell Wilson were available deep in the 2012 and 2016 drafts, respectively; both classes had two quarterbacks go off the board with the top two picks. In addition to the Colts taking Luck, the 2012 draft was notable for Washington trading a package of future picks to acquire the rights to select Robert Griffin III. And in 2016, the Rams and Eagles both traded up, to draft Goff and Carson Wentz. None of those four quarterbacks are now with the teams that drafted them. Dak and Russ (for now) are.

If anything, this list fails to capture the extent to which some of these picks were misguided. I’d argue that Bortles was the fourth-best quarterback in the 2014 draft, behind Carr, Teddy Bridgewater, and Jimmy Garoppolo. Trubisky is the third best from 2017, behind Mahomes and Watson, and Mayfield is also third best in 2018, behind Allen and Jackson. These teams didn’t just flip a coin and get the wrong guy—there were multiple other QBs on the board who have now established themselves as better options.

I was surprised to find that the best decade for drafting the top quarterback available was the 1980s. Look at this run.

NFL teams correctly identified the best quarterback on the board in five of the eight drafts between 1983 and 1990. And the misses are all excusable: Elway versus Marino is a debate worthy of its own blog post; Testaverde’s career wasn’t significantly worse than Gannon’s, and no quarterbacks were taken in the 1988 draft until the third round, so the Cardinals’ Tupa pick (no. 68 overall) was hardly a disaster. Not one of these teams made a mistake as costly as using a top-three pick on the third-best quarterback in the draft, which has happened multiple times in the last decade.

If I had to guess why teams have recently fared so poorly, I’d say that the glut of acceptable quarterback prospects has made picking the right one more difficult. Back in the 1980s, there may have been only one appealing quarterback in a class; now there are routinely three or more. But the point remains: In an era dominated by the passing game, NFL teams are worse than ever at drafting passers.

Teams Have a Better Track Record Picking QBs at No. 1—but Not Good Enough

Of the 24 quarterbacks who have been the no. 1 pick in the NFL draft since 1970, 11 of them (46 percent) went on to record the most career AV of anyone in their class at the position.

• Kyler Murray, 2019
• Jameis Winston, 2015
• Cam Newton, 2011
• Sam Bradford, 2010
• Matt Stafford, 2009
• Carson Palmer, 2003
• Peyton Manning, 1998
• Drew Bledsoe, 1993
• Jeff George, 1990
• Troy Aikman, 1989
• Terry Bradshaw, 1970

Meanwhile, in the drafts in which a quarterback was not the no. 1 pick, teams went 7-for-27 (25.9 percent) in taking the right guy first. This would seem to make intuitive sense: NFL teams have a higher hit rate with those rare QB prospects who are deemed sure things.

Even some of the top picks who were surpassed by other quarterbacks from their class in terms of career AV turned out to be superstars: Elway, Luck, and Michael Vick, for example, are not draft misses. And no matter what the numbers suggest, you can’t tell a Giants fan that Philip Rivers and Ben Roethlisberger had better careers than Eli Manning—the rings help them sleep well at night.

There are only a few instances of complete whiffs with quarterbacks who’ve gone no. 1 overall. The least productive one is Jamarcus Russell by a long shot, but he was part of a strange 2007 class from which no quarterbacks had lengthy careers. (The QB AV leader from that draft: Trent Edwards.) Same goes for David Carr—a bust, but in a draft whose top quarterback was David Garrard. These are outliers. Some quarterbacks who are now generally considered lackluster no. 1 picks (such as Jeff George and Sam Bradford) actually were the best quarterbacks in their respective classes, even if they didn’t live up to the superstar expectations that come with being the top pick.

This is cause for optimism for Jaguars fans. There is a difference between the drafts in which the top quarterback prospect is clear and the drafts in which no such player is available. The odds are high that Lawrence will have a great career. Relying on career AV data and nothing else, though, there’s a better than 50 percent chance that a quarterback other than the presumptive no. 1 pick will have the most successful career in this class. For all of the “Tank for Trevor” talk last fall, history indicates this could go down as the Justin Fields draft, the Trey Lance draft, or the Zach Wilson draft. (Or, uh, the Mac Jones draft.)

The NFL Is Good at Identifying Bad QB Classes

The flip side here is also broadly true. When NFL teams aren’t that hot on any quarterbacks in a given draft, it tends to be for good reason. The last draft in which no QBs were selected in the top 10 picks was in 2013—the aforementioned EJ Manuel year. That’s also the last draft that failed to produce any Pro Bowl quarterbacks (excluding the 2020 class).

This holds throughout history. There have been 13 drafts since 1970 in which no quarterbacks were taken in the top 10. These drafts produced only four combined quarterbacks who rank among the top 40 at the position in Approximate Value since 1970: Esiason, Cunningham, Brady, and Brett Favre. That means that nine of the 13 times no quarterbacks were taken in the top 10, the league’s consensus opinion proved to be more or less right.

The oddity here? Teams have fared surprisingly well when picking the first quarterback in these drafts with dismal classes. There have been five drafts since 1970 with no quarterbacks in the first round—and in four of those five drafts, the first quarterback taken was the right choice by AV (Danny White in 1974, Esiason in 1984, Cunningham in 1985, Tony Banks in 1996). This means the stats listed above are actually being juiced by these late-round picks. If we limit our examination to how effective teams are at picking the best available QB in the first round, the success rate drops markedly. Teams are 17-for-46 in that respect, which comes out to 37 percent.

This is meaningful info entering the 2021 draft, which is projected to have five quarterbacks picked in the first round. All signs point to Lawrence being a good NFL quarterback, probably very good. He’s the most sought-after QB prospect since Andrew Luck. But it’d be wrong to simply assume he’ll be the most productive quarterback in this class. When one team screws up a top pick, we can make fun of that team. When the NFL in its entirety shoots well under 50 percent at selecting the best quarterback in the draft, that’s a sign that the science of quarterback scouting is flawed.