Matt Corral and the Masters of the RPO
NFL Draft - How do you evaluate a quarterback prospect like Matt Corral of Ole Miss, who attempted over eight RPO passes per game in 2021?
That's not a rhetorical question. I'm really asking, because I don't know how to do it.
Corral is projected as a first-round pick in some circles, especially after earning positive reviews from his Pro Day last week. He's the QB3 on many 2021 NFL draft media boards and certainly some team boards. He definitely has some NFL attributes. But it's hard for me to wrap my head around Corral because he throws ... so. Many. RPOs.
As always, Football Outsiders' 2022 NFL draft coverage is presented by Underdog Fantasy!
Let's back up for a moment. As you probably know, RPO stands for "run pass option." It's an offensive concept that the 2017 Philadelphia Eagles popularized when Nick Foles led them to victory in Super Bowl LII over Tom Brady and the New England Patriots (he typed, somewhat vindictively). The basic design: the quarterback has the option to either hand off or fake the handoff and attempt a short play-action pass, often a quick slant or a toss into the flat, usually based on a pre-snap or immediate post-snap read.
The RPO requires rudimentary quarterback skills, such as the ability to quickly deliver accurate short passes and make simple pre-snap decisions based on factors like the number of defenders in the box. When an NFL team deploys an RPO-heavy offense, it is usually doing so to protect the quarterback, either from himself or a terrible offensive line. Teams like the Chiefs and Packers do use lots of RPOs to diversify their play-action game, but no NFL coach is looking for a quarterback prospect who is "great at the RPO," any more than they may seek a prospect who is "great at handing off from the I-formation."
So when a scouting report includes some remark like "he's really effective running the RPO," it's really saying "he did just fine in a paint-by-numbers offense designed for remedial quarterbacks." That's not to suggest that Corral is a "remedial" quarterback, but the fact that Ole Miss treated him that way is notable. Corral's film and his raw stats are so littered with RPOs that it's difficult to get a sense of what else he does.
Let's run some numbers. Here are the RPO passing statistics of each of the major 2022 draft prospects, courtesy the Sports Info Solutions college football datahub.
|RPO Passing Statistics|
|Kaleb Eleby||Western Michigan||131||99||75.6%||1,198||7||2||5||8.9|
|Matt Corral||Ole Miss||107||81||75.7%||874||3||0||3||6.3|
|Bailey Zappe||Western Kentucky||124||107||86.3%||667||7||1||0||7.9|
|Sam Howell||North Carolina||82||55||67.1%||696||4||2||3||8.3|
Eleby isn't quite a "major" prospect, but he's getting some late-round buzz, so he's worth mentioning here. He led the nation in RPO pass attempts in 2021, often flicking the ball to Skyy Moore on quick slants. When an NFL starting-caliber receiver prospect like Moore lines up in the slot against MAC competition and the play design says to throw him an automatic quick slant if there's no defender head up on him, guess what? He's going to be open, and the quarterback will get an easy completion.
Zappe is another late-round prospect and Senior Bowl attendee whose statistical profile is a real journey. Let's put him on ice for a moment.
Corral completed 75.7% of his RPO passes, which is non-noteworthy: the RPO is a high-percentage play. His adjusted net yards per attempt of 6.3, however, was depressingly low for a play he executed so frequently. Ole Miss RPOs were like pretzel bits: more filling than nourishing. On the stat sheet, they add bulk completions and completion rate while nerfing Corral's yards per attempt. On film, they're a blur of easy throws that will inevitably result in one misleading impression (fetishizing his quick release or accuracy) or the other (this kid can't do anything else).
Corral's RPO totals don't look all that different from Desmond Ridder's at first glance, even accounting for the fact that Ridder played one more game in 2021. But now let's look at the non-RPO passes of the top prospects. While we are filtering, let's also take out passes behind the line of scrimmage to get rid of quick screens, shovel passes, and other stat-sheet filler. As you can see, Ridder threw downfield far more often than Corral:
|Non-RPO, Past Line of Scrimmage|
|Bailey Zappe||Western Kentucky||449||286||63.7%||4,813||48||10||11.9|
|Sam Howell||North Carolina||213||128||60.1%||2,177||17||6||10.5|
|Matt Corral||Ole Miss||190||106||55.8%||1,940||12||5||10.3|
|Kaleb Eleby||Western Michigan||190||106||55.8%||1,945||16||4||11.0|
A whopping 36% of Corral's pass attempts in 2021 were RPOs when passes behind the line of scrimmage are filtered out. Eleby threw a higher percentage of RPOs then Corral (those similar attempt and completion numbers are not a typo; I triple-checked), but again: Eleby is a late-round prospect from a midmajor program, not someone being talked about as a potential New Orleans Saints or Pittsburgh Steelers quarterback of the future. Corral's completion rate with the RPOs and screen/swing passes filtered out is also disturbingly low. Conventional downfield passing is something Corral did not do very much and appears to have not been all that great at.
The non-RPO table helps explain the appeal of Kenny Pickett and, to a lesser extent, Carson Strong. Their film is full of conventional forward passes with minimal fluff. Pickett may check down a lot, and Strong heaves a ton of yolo balls, but at least they aren't just tossing 4-yard slants to uncovered slot receivers eight or nine times per game. Pickett's checkdowns are also more feature than bug for many scouts and insiders: they require more NFL readiness than prescribed passes to receivers with a 15-yard cushion.
And then there's Zappe, who set the all-time single-season yardage record with 5,967 yards for the Hilltoppers in 2021. Zappe threw 124 RPO passes at an 86.3% completion rate, but those RPOs represented just 18% of his total attempts. Zappe's 2021 completion rate of 69.2% is juiced by RPOs, but not by much.
Zappe appeared at the Senior Bowl and didn't embarrass himself, but he has failed to generate much pre-draft buzz. I checked my scouting notes on Zappe and found this line that sums up the issue with evaluating him: "I am basically watching a guy stand motionless in a whistle-clean pocket and throw to wide-open receivers." Who needs RPOs when you have overmatched Old Dominion and Tennessee-Martin defenses?
Malik Willis, Sam Howell, and Running by Design
Corral was also a gutsy, productive rusher: 614 yards and 11 touchdowns for Ole Miss in 2021, though with just 4.0 yards per attempt. Perhaps we're selling him short by harping on the RPOs: Corral could be a dual-threat capable of doing some of the things Kyler Murray does for the Arizona Cardinals.
Raw college rushing stats are a catch-all. Sacks are included in the data, which makes analytics types want to punch a hole in the sun. Designed runs, meanwhile, are so common that they really need to be separated out just to get a sense of how often a prospect did something he will rarely do in the NFL (run the quarterback draw or midline iso) versus something he will do a lot if he sees the field as a rookie (run for his life).
The Sports Info Solutions datahub allows us to sort scrambles from designed runs for college quarterbacks, just as it does for NFL quarterbacks. Here's each top prospect's scrambling data for 2021. B+MT stands for "broken and missed tackles":
|Rushing Production on Designed Runs|
|Sam Howell||North Carolina||74||667||9.0||31||7|
|Matt Corral||Ole Miss||71||454||6.4||18||11|
The 2022 draft prospects at quarterback break down as two players (Willis and Howell) who can be regular zone-read threats in the NFL, three (Corral, Ridder, Pickett) who could run now and then as a wrinkle, and one (Strong) whose feet will be dead-bolted to a spot in the pocket. While Corral ran frequently, the numbers above show a rather substantial difference between him and Willis/Howell in terms of results. Also, he's a rather lean lad with a habit of putting himself in harm's way.
Howell ran a nation-leading 39 quarterback draw plays, averaging 8.5 yards per rush. If you think the RPO makes a quarterback tricky to evaluate, try filtering out several runs right up the middle per game. Howell is a rugged runner, but he's not Cam Newton, and he does a lot of Carson Wentz stuff when asked to stand in the pocket and be a semi-conventional quarterback. Howell would be fun to watch in the USFL and will make some preseason games lively. In the NFL, he might end up in some goofy Kevin Hogan/Jacoby Brissett Wildcat/short-yardage specialist role.
Before we talk about Willis, let's run the scrambling table:
|Rushing Production on Scrambles|
|Matt Corral||Ole Miss||54||312||2.5||13||0|
|Sam Howell||North Carolina||50||438||5.1||26||3|
College quarterbacks should not scramble a lot, because their systems are generally designed to reduce pressure and they often play for superior programs. College quarterbacks who throw RPOs on 36% of their pass attempts definitely shouldn't scramble a lot, because the RPO gets the ball out of their hands immediately by design. Sorry to keep ripping Corral, but his stat breakdowns are just one yellow flag after another.
Willis' scramble rate is a little troubling. He scrambled against Ole Miss seven times, which is fine: the Flames were outgunned/outmanned/etc. He also scrambled seven times against Louisiana and Army and nine times against Louisiana-Monroe. Willis is an outstanding scrambler and was the best athlete on the field in many of those games, so he probably received lots of latitude to take off and run from his coaches. But that's exactly the sort of habit that must be relearned and replaced by something else in the NFL.
Pickett also scrambles quite a bit, but there's a difference between taking off five times per game against Virginia and Miami and seven times against Army. We're not going to sing any more hosannas about Pickett's decision-making than he's already getting from insiders who are trying to talk him up as QB1, but he's mobile enough to escape harm's way in the NFL. It's also worth noting that, despite his dainty little hands, Pickett only fumbled once when running by design or scrambling.
Ridder's low scramble rate is illuminating. Ridder runs well but rarely chose to do so. He played for a midmajor powerhouse that did a fine job protecting him, but he also faced Alabama, Notre Dame, Indiana, and some AAC programs that can take care of themselves (Houston, UCF) in 2021. Ridder's low scramble totals supplement the argument that he will enter the NFL as an experienced, relatively game-ready decision-maker. Pickett's scramble totals don't refute his similar claim, but they don't support it.
Strong is completely immobile, but he makes up for it by having an overrated arm. We'll let Strong have his revenge in a week or two when we break down each prospect's 20-plus-air yard passes, but the lack of Strong chatter coming off his Pro Day last week speaks volumes. Here's Strong's "money" throw from that workout, complete with a double-hop windup, if you are into that sort of thing:
Hello, Carson Strong.
— Nick Penticoff (@NickPenticoff) March 22, 2022
How to Evaluate a QB in a System Full of Bullsh*t
Let's circle back to our initial question: how are we supposed to get a true read on a quarterback prospect who showed us so little of substance?
Before the RPO, there were the bubble and tunnel screens. College programs were addicted to them about a decade ago. Half the quarterbacks in the nation were tossing six to 10 glorified sweeps to wide receivers in spread offenses per game. Paxton Lynch threw for 8,863 yards in three seasons at Memphis, but I swear he only threw for about 250 air yards. Lynch was the prime example of a prospect whose evaluation turned to gibberish because so much of his production was piffle.
Those receiver screens were in vogue at around the same time as the "spot on the field" throw, a Big 12 staple that made collegiate rock stars of Graham Harrell and Kliff Kinsgbury and could make Landry Jones look like an NFL prospect. We're talking about the Mike Leach Air Raid in its early incarnation, as well as various successors and imitators, plus other spread concepts and variations from 10 or 20 years ago, all of them popular in a conference that became notorious for defense-optional sandlot football.
Blaine Gabbert stood in a chapel-quiet pocket and floated short pass after short pass into the wide-open prairies. His receivers knew where the ball was going and their defender didn't, so the receivers won. Sam Bradford did much the same thing. Yes, these are oversimplifications, but Gabbert was a tricky evaluation because he threw so few passes under duress into tight coverage. Every Leach quarterback has been a system product, even as elements of his offense have been integrated into half the programs in the nation and the NFL. Bradford arrived in the NFL as injured vaporware and stayed that way.
It has always been this way, going back before Football Outsiders existed, when scouts were tasked with projecting triple-option quarterbacks and Big 10 BMOCs who handed off 40 times per game into the NFL. But don't despair! Here's how you can filter out all the noise to determine what a quarterback in a hinky system is really capable of:
- Become a team scout with NFL resources at your fingertips who can focus on prospect evaluations as a full-time career;
- Cross your fingers and proceed by dead reckoning;
- Recognize that team scouts with NFL resources at their fingertips who can focus on prospect evaluations as a full-time career are crossing their fingers and proceeding by dead reckoning.
NFL evaluators can see lots of things on film that even an experienced media scout cannot. They can talk to the prospect, his coach, and his weight bench sanitizer so they can learn things about his understanding of NFL offenses, work habits, and likelihood to end up on the police blotter that the rest of us can only guess at. Yet the NFL's success rate given all of this inside information speaks for itself. It has been five years since the "historic" 2018 quarterback class and the only first-round picks with any dignity left are the dude that #DraftTwitter and #AnalyticsTwitter treated like a smoldering poop bag on the porch and the dude The Establishment wanted to move to wide receiver.
Face it: we all suck it this. And yet we keep trying. So if we cannot succeed very often, let's at least fail for the smartest possible reasons.
Corral's film shows a pesky-but-scrawny quarterback who executes a pea-shooter passing game. He runs hard, but he got injured in his final college game. The stat breakdown shows that he relied heavily on a system which gave him easy reads and lots of high-percentage short passes.
So what are Corral's starting-caliber NFL quarterback attributes? A quick release? Great short accuracy? Mobility? Fine. Now: are those truly elite qualities for Corral, or are they simply all we have to talk about? Competitiveness? OK, boomer. If the best argument for someone being a starting-caliber prospect is that someone else put him on the list of starting-caliber prospects, then we have surrendered to groupthink. Which is another thing the professionals who run NFL teams do all the time.
The statistical breakdowns of all the 2021 draft quarterback prospects' rushing and RPO data bring us right back to where we started. Willis is gifted but unready. Pickett and Ridder are experienced and do some things well but lack va-va-voom factor. Howell is this class' designated option guy, Strong the flypaper for Mike Mayock types.
And Corral? He's a Lynch- or Gabbert-style system quarterback, just in a different system. His RPO breakdowns are warning us that he will max out as a feisty NFL backup.
We'll find out in a month whether teams like the Falcons, Steelers, and Saints heard the message.