This week's DVOA commentary is all about worsts. Come find out where Washington stands among the worst special teams in DVOA history, whether San Diego has the biggest gap between offense and defense, and whether Baltimore or Jacksonville has the worst running game we've ever tracked.
28 Feb 2013
by Aaron Schatz
Even before they ever took the field, most NFL observers knew the quarterback class of 2012 was going to be one of the better ones in recent years. Scouts generally listed Andrew Luck and Robert Griffin III as the two best quarterback prospects since at least 2004, and maybe even since Peyton Manning back in 1998. And when Football Outsiders applied our statistical formula, the Lewin Career Forecast, it came out in total agreement with the scouts. Luck came out as one of the top ten quarterback prospects since 1998, and Griffin topped Philip Rivers to set a new record for the best-ever projection using the LCF system.
Or at least, he would have, except that there was one other player who had an even higher projection than Griffin: Wisconsin quarterback Russell Wilson. The idea of Wilson as one of the strongest prospects ever seemed ridiculous at the time. We dubbed Wilson "The Asterisk" and tried to explain why his projection might be incorrect. The system is only meant to be used on players chosen in the first two days of the draft, and most people didn't think Wilson was going until the third day. We thought the system overrated his senior improvement because he had transferred from North Carolina State to Wisconsin. And LCF couldn't account for his height because no quarterback under six feet had ever been drafted in the first three rounds.
Seattle then drafted "The Asterisk" in the third round, and he led them to the playoffs with the highest DVOA rating of any of the rookie quarterbacks. The moral of the story: You always need to use common sense when analyzing a statistical projection system that doesn't incorporate everything that scouts can learn from watching film. But unexpected numbers may be telling you something.
That's an important lesson when looking at the 2013 draft class, because the LCF likes this class a lot more than the scouts. This year's top prospects rank among the highest we've ever tracked, even though none are seen right now as first-round talents. Will players like Geno Smith and Matt Barkley rank with Colt McCoy and Brady Quinn as the LCF's biggest failures? Or will one of these players surprise like Russell Wilson?
There are seven variables involved in the Lewin Career Forecast:
These last two variables work together to penalize both quarterbacks who scramble too often and quarterbacks who take a lot of sacks (since sacks are counted as negative runs in college), while pocket quarterbacks who are successful when they do run (Andrew Luck, for example) get a bonus.
The projection number represents an estimate for passing DYAR in years 3-5 of a player's career. (DYAR, or Defense-Adjusted Yards Above Replacement, is explained here.) The top prospects will be above 1,200 DYAR, and you should avoid quarterbacks below zero.
Smith is the kind of prospect that the LCF loves: He's a three-year starter with a high (67.4 percent) completion rate, and he's mobile but a passer first. The LCF also likes that his stats improved in his senior year, which suggests he's still learning and growing in the position. Unfortunately, the LCF isn't looking at Smith's game-by-game record in 2012, which was astoundingly inconsistent. West Virginia ranked 122nd out of 124 teams when it came to variance in Football Outsiders' Offensive FEI metric. He struggled against teams with strong pass defenses, like Kansas State and TCU, and those are the games teams need to look at before they decide if Geno Smith is worth a first-round pick.
Barkley is a four-year starter, and despite the success of players like Cam Newton and Aaron Rodgers, games started in college still correlates with NFL success more than any other factor. However, Barkley did decline slightly as a senior. What about the "USC curse?" Well, Matt Leinart fell far short of his LCF projection, but Mark Sanchez has only fallen slightly short of his because LCF didn't rate him highly to begin with. And while Carson Palmer may no longer be an above-average NFL starter, he far surpassed his projection. (People tend to forget how good Palmer was in his first few seasons.)
Nassib is a three-year starter who improved significantly in his final season, although his career completion rate isn't as high as you would like (only 60.4 percent). He seems to be climbing up the draft boards of the various draft experts across the Web, although that doesn't tell us what teams think about him. After the Bills hired his college coach Doug Marrone, Nassib to the Bills in the second round is too obvious not to happen.
The MVP of the Senior Bowl was a better college quarterback than you might realize, with a 66.9 percent completion rate and three years of starting experience. I've seen articles comparing him to Blaine Gabbert, and other articles comparing him to Terrelle Pryor, which is somewhat like saying that something tastes like both steak and jelly beans.
Wilson doesn't come out well because he has only been a starter for two seasons, and he declined a little in his senior year. There may be extenuating circumstances: according to FO college analyst Brian Fremeau, Arkansas played the sixth-hardest schedule of opposing defenses in the country last year. Scouts like Wilson's feel in the pocket, but he takes more sacks than you might expect. Remember, sacks count as runs in college football, and for his career Wilson had 110 runs for -44 yards.
Phrases you don't want to be associated with when you come out with a year of eligibility remaining: immature, inconsistent, "lack of decision-making ability." Completion rate you don't want to be associated with when you come out with a year of eligibility remaining: 58.6 percent.
Glennon has only two seasons as a starter, because he was sitting behind Russell Wilson. Glennon's stats as a junior (62.5 percent comp. rate, 6.7 yards per attempt, 136.4 passer rating) are actually better than Wilson's stats the year before as a redshirt junior with the same program (58.4 percent, 6.8, and 127.5). But this is where we see the importance of steady improvement. Wilson's performance blossomed in his senior season, when he transferred to Wisconsin, while Glennon's performance as a senior declined. (He led the nation with 17 interceptions.) The other problem with Glennon may be his body type, because he's extremely skinny. We'll have to see how he comes out at the combine, but last year he was listed at 6-foot-7 and 220 pounds, a 24.8 BMI. No quarterback since 1998 has been drafted in the first three rounds with a BMI below 26.
The LCF is built to apply only to quarterbacks chosen in the first three rounds of the draft. After that, quarterback success and failure becomes too difficult to predict. Part of the concept of the system is that scouts will do a good enough job identifying "system quarterbacks" so that those quarterbacks whose college stats are much better than their pro potential will naturally fall to the third day of the draft.
So here's Landry Jones, who will rank with the fourth-highest projection ever -- unless he doesn't get taken in the first three rounds. This time, we can't use excuses like we did with Russell Wilson. Jones has been starting at Oklahoma for four years and has prototype quarterback size. But scouts have pinned a number of red flags on him. He seems like the Colt McCoy of this year's class, although his issue isn't arm strength like McCoy's was. Jones has fine arm strength, but he's immobile and many scouts believe he wilts under pressure. Issues with mechanics and footwork were often hidden by an Oklahoma spread offense that involved lots of quick-developing screen passes.
Perhaps some team will figure out a way to solve Jones' problems, grab him in the third round, and turn him into a quality NFL starter. Or he may drop to the fifth and be forgotten as just another college system quarterback-turned-NFL flop. It's just another example how, when it comes to picking future talent, even the most stat-oriented of us know that numbers complement scouting. They don't replace scouting.
(This article originally appeared at ESPN Insider. There was one quarterback we left out of the ESPN article, because he probably won't go in the top three rounds. If he does, Zac Dysert of Miami of Ohio has a projection of 1,131 DYAR, including the penalty for non-BCS conferences.)
57 comments, Last at 28 Apr 2013, 6:54pm by Theo