One of the starting quarterbacks in Sunday's Super Bowl, Kurt Warner, famously wasn't drafted out of college, so he had to bag groceries, then became an Arena Football League quarterback, then an NFL quarterback, then a league MVP and Super Bowl winner, then he became a has-been, and now he's back in the Super Bowl at age 37.
This kind of thing is not hugely uncommon in the NFL (consider the career of Jeff Garcia, who didn't make it to the NFL until he was 29, has been twice given up on, and still was 9th in the NFL in passer rating this season at 38), even though, as the top job in American sports, a huge amount of expertise is devoted to evaluating potential quarterbacks.
In the New Yorker, Malcom Gladwell says:
This is the quarterback problem. There are certain jobs where almost nothing you can learn about candidates before they start predicts how they'll do once they're hired. So how do we know whom to choose in cases like that? ... The problem with picking quarterbacks is that [U. of Missouri quarterback] Chase Daniel's performance can't be predicted. The job he's being groomed for is so particular and specialized that there is no way to know who will succeed at it and who won't. In fact, Berri and Simmons found no connection between where a quarterback was taken in the draft—that is, how highly he was rated on the basis of his college performance—and how well he played in the pros.
From that, Malcolm extrapolates that we should completely change the way teachers are selected in America. Which may or may not be a good idea, but, let's first figure out if he's right about NFL quarterbacks.
When Malcolm makes a quantitative statement, it's usually time to fire up Excel and check for yourself. I went to Pro-Football-Reference.com and looked up all 277 quarterbacks chosen in the NFL draft in the 1980s and 1990s. (I wanted recent QBs but not so recent that we can't get a sense of how there careers will turn out.) Here are the average career achievements (keeping in mind that some, like Peyton Manning and Donovan McNabb, aren't done yet):
# of QBs |
| Pro Bowls | Seasons Starting | Games | Yards |
7 | Number 1 Picks | 4.1 | 11.0 | 171 | 37,089 |
17 | Top 5 Picks | 2.4 | 7.8 | 124 | 25,480 |
23 | Top 10 Picks | 1.8 | 6.1 | 103 | 20,296 |
31 | Top 20 Picks | 1.6 | 5.6 | 97 | 18,643 |
54 | Top 50 Picks | 1.5 | 5.1 | 91 | 17,338 |
43 | Picks 51-100 | 0.3 | 2.1 | 57 | 6,461 |
70 | Picks 101-200 | 0.2 | 1.1 | 32 | 4,307 |
110 | Picks >200 | 0.1 | 0.3 | 13 | 1,531 |
In other words, seven quarterbacks were chosen first overall in the draft, and, on average, they earned 4.1 Pro Bowl recognitions each, started for 11 years, played in 171 games, and threw for 37,000 yards. On average, the QBs picked number #1 ended up being worth it, although not necessarily to the teams that drafted them (e.g., Vinnie Testaverde's two Pro Bowl selections came at age 33 and 35, with his third team, and he started six games as a 44-year-old -- an odd career, but a pretty good one).
In contrast, 110 quarterbacks were chosen 201st or worse in their draft year, and, on average, they achieved 0.1 Pro Bowl selections, 0.3 years as a starter, a 13 game-long career, and threw for 1,531 yards.
For these two decades, draftees can be lumped into roughly four categories:
- the seven #1 overall picks (Tim Couch, Peyton Manning, Drew Bledsoe, Jeff George, Troy Aikman, Vinnie Testaverde, and John Elway), who had 29 Pro Bowl appearances among them. Hall-of-Famer Steve Young might have been another first-player-chosen in the NFL draft if he hadn't signed with the upstart USFL. (He was the #1 choice in the NFL's subsequent "supplemental draft" of USFL players, as was Bernie Kosar the next year.) On the other hand, the quarterbacks taken first overall in the entire draft in this decade (Michael Vick, David Carr, Carson Palmer, Eli Manning, Alex Smith, and JaMarcus Russell) probably won't match their predecessors. So far, they only have five Pro Bowl appearances.
- Picks in each draft from #2 to #50 overall. The two top yardage quarterbacks fall here: Dan Marino was the 27th player picked his year, and Brett Favre the 33rd. Gladwell's contention was closest to truth here, where there didn't seem too much of a trend between being a #2 and being a #50. The quarterbacks who were picked 51-100th went to 12 Pro Bowls, while the guys who were picked 1-50th went to 81, so, when Gladwell says, "there is no way to know who will succeed at it and who won't," don't believe him. There are ways. They are far from perfectly reliable, but 81 to 12 is a pretty good indication that the NFL guys aren't just throwing darts. There is some bias in most of the statistics toward the high draft picks in the sense that once a team makes a big investment in a quarterback, they often feel obligated to play him. But the Pro Bowl statistics are fairly objective.
- Picks from #51 upward -- Lots of good quarterbacks were taken down here, like Rich Gannon (#98), Mark Brunell (#118), Matt Hasselbeck (#187) (and Tom Brady went #199 in 2000) but the average achievement level is low because the pyramid is so broad. The lowest drafted quarterback during these two decades to make the Pro Bowl was Doug Flutie at #258. He was also no doubt the shortest Pro Bowl quarterback.
- And then there are the undrafted quarterbacks, such as Warner, Garcia, Tony Romo, Jake Delhomme, and Jon Kitna (and Warren Moon back in the late 1970s), who emerged out of the couple of thousand or so college quarterbacks who went undrafted during these two decades. No doubt there were other undrafted quarterbacks who, with the right breaks, could have been stars in the NFL, but the percentages have to have been very low -- the pyramid gets very, very wide down here.
In conclusion, contra Gladwell, the NFL teams can predict quarterback performance in the NFL a lot better than random chance would dictate. And yet, considering the huge amount of effort that goes into selecting the most promising college quarterbacks in the NFL draft, there is much that remains delightfully unpredictable, as Kurt Warner's career demonstrates.
One of Malcolm's biggest problems is that he has very little sense of where he is on a bell curve. He looks at people on the 99.999th percentile (top 50 draftees) and says that nobody can predict who will make it to the 99.9999th percentile, and, therefore, we should throw out prediction methods. Well, swell, but that doesn't mean that you can't predict ahead of time with some degree of accuracy who will wind up at roughly the 10th, 50th, and 90th percentiles out of the general population. But, Malcolm just doesn't get it.
My published articles are archived at iSteve.com -- Steve Sailer