Stat Analysis
Advanced analytics on player and team performance

QBASE 2020

LSU Tigers QB Joe Burrow
Photo: USA Today Sports Images

There's no risk in the NFL quite like the risk of drafting a quarterback. No defense revolves around a single player the way every offense revolves around the quarterback. Trust your offense to the wrong young quarterback, and your team isn't going to climb back into playoff contention. Quarterbacks get drafted earlier than players who rank similarly at their respective positions. And while quarterbacks may have more statistics measuring them than other players do, teams haven't been more accurate in drafting them. Just ask the Chicago Bears, who selected Mitchell Trubisky instead of Deshaun Watson and Patrick Mahomes three years ago.

The lessons of history can at least help us figure out how much of a risk each quarterback prospect will be. That's the point of Football Outsiders' Quarterback-Adjusted-Stats-and-Experience (QBASE) projection system. It looks at college performance, experience, and expected draft position (to incorporate scouting information that college stats will miss). To allow some time for development, QBASE projects a quarterback's efficiency (passing only) in Years 3 to 5 of his career according to Football Outsiders' Defense-adjusted Yards Above Replacement (DYAR) metric. 50,000 simulations produce a range of potential outcomes for each prospect, with players drafted later generally having a larger range of possibilities.

You'll notice that every listed quarterback prospect has a chance to be elite and every quarterback has a chance to be a bust. That reflects just how much we don't know about drafting quarterbacks, and the wide range of possible outcomes for each player.

QBASE favors quarterbacks expected to go high in the draft who also have a relatively long resume of college success according to the stats. Those stats include completion percentage, adjusted yards per attempt (adjusted for touchdowns and interceptions), and team passing efficiency (measured with Bill Connelly's passing SP+ stats). These numbers are adjusted both for the quality of the defenses that a prospect had to face as well as the quality of his offensive teammates. QBASE is meant to only be used on players chosen in the top 100 picks; after that, the judgment of scouts becomes even more important, and statistics become even less predictive.

Overall, QBASE thinks this is a good year for moderately promising quarterback prospects. None of this year's quarterbacks come close to the top projections in QBASE history, in part because no top quarterback prospect this year has four full seasons as a college starter. However, this is the first year with three prospects with mean projections over 600 DYAR since 2012 (Andrew Luck, Robert Griffin, and Russell Wilson), and the first year with five prospects with mean projections over 400 DYAR since 2006. Then again, those five quarterbacks in 2006 were Matt Leinart, Jay Cutler, Kellen Clemens, Vince Young, and Tarvaris Jackson -- once again demonstrating that projecting quarterbacks is very difficult.

Here are projections for eight quarterbacks who may go in the top 100 picks of the 2020 NFL draft.

Joe Burrow, LSU (Scouts Inc. ranking: 2)

Mean Projection in Years 3-5 759 DYAR
Bust (< 500 DYAR) 42%
Adequate Starter (500-1499 DYAR) 27%
Upper Tier (1500-2500 DYAR) 18%
Elite (>2500 DYAR) 13%

Joe Burrow in 2019 had perhaps the greatest season ever in college football history. He set a new all-time record with 60 touchdown passes. His completion rate of 76.3% was the second highest ever. His 12.5 adjusted passing yards per attempt ranks fourth. He did this against an above-average schedule in the SEC.

Burrow's forecast takes a small hit from playing alongside a very high amount of draftable talent, highlighted by likely first-round picks Justin Jefferson (this year) and Ja'Marr Chase (next year). But the experience variable is the main reason Burrow's QBASE projection is not higher. The system still identifies the risk that comes with drafting a quarterback with only two years of starting experience. Nonetheless, Burrow has the second-highest projection ever for a quarterback with less than three years of starting experience.

Top QBASE projections for 1-Year and 2-Year Starters, 1997-2020
Player School Year QBASE
Alex Smith Utah 2005 798
Joe Burrow LSU 2020 759
Cam Newton Auburn 2011 698
Tua Tagovailoa Alabama 2020 653
Kyler Murray Oklahoma 2019 595
JaMarcus Russell LSU 2007 570
Sam Bradford Oklahoma 2010 545
Vince Young Texas 2006 536
Dwayne Haskins Ohio State 2019 527
Tim Couch Kentucky 1999 474

Tua Tagovailoa, Alabama (7)

Mean Projection in Years 3-5 653 DYAR
Bust (< 500 DYAR) 46%
Adequate Starter (500-1499 DYAR) 27%
Upper Tier (1500-2500 DYAR) 19%
Elite (>2500 DYAR) 8%

As a sophomore, Tagovailoa had 12.8 adjusted passing yards per attempt, the third-highest number ever. As a junior, Tagovailoa had 13.4 adjusted passing yards per attempt before he injured his hip. That would be the highest number ever if Tagovailoa had not missed time and failed to qualify for College Football Reference's historical rankings. So the elements of QBASE based on past performance like Tagovailoa an awful lot.

On the other hand, QBASE has a variable to discount quarterbacks whose stats are helped by playing with a lot of draftable talent at the receiver and offensive line positions. And in the history of QBASE, which goes back to 1997, no quarterback played with more talent in his final college season than Tua Tagovailoa in 2019.

Three different Alabama teammates are forecast to be top-20 picks in the upcoming draft: tackle Jedrick Wills and wide receivers Jerry Jeudy and Henry Ruggs. Two more of Tagovailoa's offensive teammates are likely to be top-10 picks in next year's draft: tackle Alex Leatherwood and wide receiver Jaylen Waddle. Receiver Devonta Smith and guard Deonte Brown are also likely to be drafted in 2021. As a result, the projected draft value for Tagovailoa's teammates is over 60% higher than the previous quarterback who played with the most draftable talent, Danny Wuerffel in 1997.

(Joe Burrow also is now in the top five for teammate value, along with Johnny Manziel and Matt Leinart.)

Tagovailoa's schedule also wasn't as hard as you might expect for a quarterback in the SEC; due to his injury, he missed Auburn (fifth in SP+ defense) and the bowl game with Michigan (11th).

These elements only drop Tagovailoa a little bit, however. He still has one of the all-time best QBASE projections for a quarterback with only two years of starting experience.

Jordan Love, Utah State (22)

Mean Projection in Years 3-5 57 DYAR
Bust (< 500 DYAR) 64%
Adequate Starter (500-1499 DYAR) 23%
Upper Tier (1500-2500 DYAR) 10%
Elite (>2500 DYAR) 2%

Scouts are high on Love's arm talent and smooth release, but his final year of college was quite unimpressive. Thanks in part to 17 interceptions with only 20 touchdowns, his already mediocre 7.2 yards per attempt number becomes 6.4 adjusted yards per attempt. In the last decade, the only top-100 pick with a lower adjusted yards per attempt in his final college season was C.J. Beathard. Love had this poor performance against the easiest schedule of pass defenses faced by any of this year's top quarterback prospects.

But there's an asterisk here, which is that Love's performance declined dramatically between his sophomore and junior seasons. In 2018, Love had a far more impressive 9.4 adjusted yards per attempt, with 32 touchdowns and only six interceptions. If Love had come out a year ago, he would have had a higher projection (420 DYAR). Teams considering Love in this year's draft need to watch two years' worth of film and figure out what changed for Love in 2019.

Justin Herbert, Oregon (23)

Mean Projection in Years 3-5 689 DYAR
Bust (< 500 DYAR) 43%
Adequate Starter (500-1499 DYAR) 28%
Upper Tier (1500-2500 DYAR) 19%
Elite (>2500 DYAR) 11%

Herbert has started for two full seasons and two partial seasons at Oregon, so we counted him with three seasons for the purposes of QBASE. That gives him the second-best projection this year, ahead of Tagovailoa but still behind Joe Burrow. Where Herbert takes a bit of a hit is that Oregon was only 36th in FBS in passing SP+ this season.

Jake Fromm, Georgia (57)

Mean Projection in Years 3-5 439 DYAR
Bust (< 500 DYAR) 51%
Adequate Starter (500-1499 DYAR) 28%
Upper Tier (1500-2500 DYAR) 15%
Elite (>2500 DYAR) 7%

Fromm gets a boost in QBASE from three years of starting experience. Both his 8.1 adjusted yards per attempt and 61% completion rate last season are below the mean for top-100 picks over the past 20 years, but he played the hardest schedule among this year's top quarterback prospects. And like Jordan Love, Fromm declined in his final season, as he had a 67% completion rate and 10.1 adjusted yards per attempt in 2018. Unlike Love, Fromm's QBASE projection would have been lower had he come out a year ago with less experience (208 DYAR).

Jacob Eason, Washington (60)

Mean Projection in Years 3-5 -98 DYAR
Bust (< 500 DYAR) 68%
Adequate Starter (500-1499 DYAR) 21%
Upper Tier (1500-2500 DYAR) 9%
Elite (>2500 DYAR) 2%

Eason's 2019 season stats were close to the mean for top-100 quarterback prospects, but he put up those numbers against an easier than average schedule. Combine that with a lower expected draft position (end of th second round) and only two years of starting experience, and Eason does poorly in our QBASE projections.

Jalen Hurts, Oklahoma (98)

Mean Projection in Years 3-5 410 DYAR
Bust (< 500 DYAR) 51%
Adequate Starter (500-1499 DYAR) 25%
Upper Tier (1500-2500 DYAR) 16%
Elite (>2500 DYAR) 7%

The QBASE system likes Hurts even though he's projected to be a late-third- or early-fourth-round pick. He's got three years of starting experience and had awesome statistics in his final year at Oklahoma: 70% completion rate and 12.2 adjusted yards per attempt. That latter number is the sixth-highest qualifying AYPA number in FBS history. But when you look at the five players ahead of him, you sense the biggest question about Hurts. Those players include Baker Mayfield twice and Kyler Murray. Four of the top six APYA seasons ever come from Lincoln Riley's Oklahoma offense. How much of Hurts' performance is Hurts, and how much is the offense, especially considering the big jump in Hurts' stats from his two starting years at Alabama to last year at Oklahoma?

Anthony Gordon, Washington State (116)

Mean Projection in Years 3-5 -234 DYAR
Bust (< 500 DYAR) 72%
Adequate Starter (500-1499 DYAR) 20%
Upper Tier (1500-2500 DYAR) 7%
Elite (>2500 DYAR) 1%

It's hard to put too much faith in Gordon considering that his only year as a starter came as a redshirt senior. Even Gordon's good-looking stats aren't as impressive as they might seem at first glance. He was second in FBS with 5,579 passing yards and fourth with 48 touchdowns, but those stats were partially the product of ranking first in pass attempts in Mike Leach's pass-friendly Washington State offense. Gordon was also helped by an easier schedule in the offense-friendly Pac-12. His 8.4 adjusted passing yards per attempt ranked only sixth in the conference.

This article originally appeared on ESPN+.

Comments

19 comments, Last at 06 Apr 2020, 9:08pm

1 Tua

I take it that QBASE doesn't attempt to factor injuries/health in?

2 Only if

In reply to by serutan

Only if a player misses most of a season, for the purposes of counting starting seasons.

3 Burrow's rating doesn't…

Burrow's rating doesn't surprise me; I assumed he'd take a hit for his earlier season(s). I figured Herbert would do better because he'd started basically three seasons, and played with very little draftable talent. His receivers this year were better than Love's, but not by as much as you would think.

With Burrow you have to wonder how much credit for his 2019 season should go to coordinator Joe Brady and the talent around Burrow. None of these guys are close to safe bets; They all have bust potentials over 40%.

5 Burrow 2019 credit

Happy LSU fan here.

I think that it is a combination of: Coordinator Joe Brady; his receivers (including Randy Moss' son); RB Clyde Edwards Helaire (day 2 draftee), a good O-line, and Burrow being great.

If you watch the highlights of different games, you will see plays where Burrow has his first read schemed open, makes an accurate pass, and the receiver does the rest. You will see plays where Burrow goes through his reads and makes the best choice. You see some where Burrow throws a 50/50-type ball, trusting his receivers to make the catch--and they usually did. But you will also see some plays where the protection breaks down, or a blitz gets home, and Burrow makes a great pass under pressure, or eludes the pressure and runs, or eludes the pressure and makes a play. What is unbelievable to me is the almost 77% completion percentage. It would be hard to do that regularly in a padded scrimmage--and he did it for 15 games, against elite competition!

Now, I am not predicting that he is going to be the 2nd coming of Peyton Manning/Tom Brady/Drew Brees/et al. But I think he can have a career like Bledsoe or Stafford. I would be shocked if he didn't become an above-average starter. 

9 Double counting?

Do you think it’s possible that the model is docking guys from big schools twice? Take Tua- he’s docked for playing with NFL talent, and then he’s docked again for being behind an NFL talent in Hurts as a freshman.

14 I think this is definitely…

In reply to by ThunderstruckX44

I think this is definitely something to consider, especially as the gap between the haves and have-nots in college football continues to grow. The idea of one school having two NFL starters at quarterback (let alone two first overall draft picks) on the same roster used to be a rarity. It certainly seems more common now.

15 Exactly!

And I think one of the best ways that we're adjusting analytics is by controlling for teammates, but it seems like we're only doing that in one direction. I think a lot of what we've learned about prospect analytics with regard to how strong of a predictor usage seems to be at multiple positions is being called into question now that there are only 6 or so CFB programs that matter. 

I suspect the same thing is going to happen when Playmaker Score is down on Jerry Jeudy and Henry Ruggs. 

In any case, you guys do awesome work, thanks for the response!

16 <em>"The idea of one school…

"The idea of one school having two NFL starters at quarterback (let alone two first overall draft picks) on the same roster used to be a rarity."

This seems to be begging for data, as I don't have the impression that it used to be rare. Especially if you go back to less scholarship limited times. The big/major schools seem to have always had that kind of depth to me (2 future NFL starters on the roster... not the first overall draft pick thing. That's always going to be rare.)

18 OK, I went to PFR and…

OK, I went to PFR and checked every QB drafted, looking for every case where one school had two or more quarterbacks drafted in back-to-back years, or the same year. I threw out the AFL drafts when they were competing with the NFL, to avoid cases where the same player was drafted twice, and limited it to the first 250 picks, which is about how long the draft is now. Here's how often it happened:

 

Decade Instances
1940-1949 7
1950-1959 7
1960-1969 1
1970-1979 8
1980-1989 5
1990-1999 5
2000-2009 10
2010-2019 3

 

Only three times this decade, the fewest since the 1960s, which surprises me. It occurs to me that this may be the effect of transfers, which (I think) are more common now than they used to be. Each QB is only listed with his last school, so this ignores that Russell Wilson and Mike Glennon were teammates at N.C. State; or that Davis Webb was teammates with both Patrick Mahomes and Baker Mayfield at Texas Tech, even though none of them finished there; or the Michigan State trio that somebody else mentioned in this thread.

10 Does QBASE account for the…

Does QBASE account for the player's age? That's maybe the main thing that would concern me about Burrow - he's older than Lamar Jackson.

11 Nope

Nope, but I agree that it's a consideration. I think of it as a consideration more the other direction: players who come out at age 21 do have development time in front of them.

12 I agree there's a point…

In reply to by Aaron Schatz

I agree there's a point where it’s a consideration, but I'm not sure Burrow is at that point. This was a huge reason I couldn’t understand the love in draftnik circles for Brandon Weeden as a 1st round prospect.

13 How similar is Jordan Love…

How similar is Jordan Love to Josh Allen? Not in the numbers, but in the scouting. I just wonder how much a prospects perception is affected by how the last similar guy did.

19 Hmmm

I'd say Jordan Love is way more accurate than Josh Allen ever has been. Allen has a stronger arm although both are excellent and I think Allen is the better runner. The biggest issue with Love is decision making. That could be because the team around him was garbage and he needed to make plays, because if he didn't they'd certainly lose, so the INTs didn't actually cost his team much in win %. Maybe he's just way too overconfident in his ability to squeeze passes in there. This is one of those cases where it would be extremely helpful to interview him and pick his brain for why he made so many risky throws. It's one of the reasons why many people were much lower on Mahomes than hindsight says. People discussed Mahomes "upside" as being based on his arm talent, but the real source of his success is from his ability to mentally process stuff along with his freaky arm. If Love doesn't have that, it seems unlikely to work out for him in the NFL.