By popular demand (and this is clear when we look at the popularity of his posts) guest blogger Steve Walters is back. For those of you with extremely short memories, Steve teaches economics at Loyola College in Maryland but remains a fan of his hometown Red Sox, Celtics, Patriots, and Bruins.
On the second day of last weekend’s NFL draft, the Baltimore Ravens did something fabulous. In the fifth round, they snagged this year’s Heisman Trophy winner, Ohio State QB Troy Smith.
I argue that this was a genius move by Ravens GM Ozzie Newsome not because I attach any significance whatever to the hardware that sits on Smith’s mantle. I haven’t run the numbers, but my guess is that Heisman winners have, in the past, disappointed more often than they’ve impressed as pros.
What’s remarkable about Smith is that despite having the coolest trophy in sports, and despite having played superbly in one of the most visible programs in one of the best conferences in college footballdom, Smith has somehow managed to be under-rated and, therefore, under-valued. All because he is (gasp) SHORT. A mere six feet (recoils in horror).
The ongoing bias against short field generals is progress of a sort. Not too long ago, you may remember, when quarterback jobs were open, Blacks were told they need not apply. Heaven knows what college coaches told themselves as they converted talented Black QBs into defensive backs and wide receivers, or how NFL GMs rationalized ignoring the relatively few who’d managed to slip past the position sentries in college. But it’s possible that their “reasoning” sounded a lot like former Dodger exec Al Campanis’s ill-fated remark, in explaining why there were so few African-American managers and execs in baseball, that they “lack some essentials” to get the job done.
The prejudice against short QBs can also be rationalized. They get a lot of passes batted down at the line. They can’t see downfield as very well. You have to design a different, rollout-style offense especially for them.
Now, these things might actually be true, at least to some extent. But if they were, in technical jargon, truly “disqualifying,” then you wouldn’t be able to find any successful short QBs in the pro game.
But, of course, you can. There are two pretty good six-footers active right now named Drew Brees and Michael Vick. If you go back just a couple of years there was Doug Flutie (another Heisman winner), generously listed as five-ten. You can probably come up with several more examples. If six feet was the magical height threshold below which all passes got stuffed or all downfield receivers were invisible, we’d never have heard of any of these guys.
So if short stature is not disqualifying, is it a non-trivial disadvantage? I don’t know. And what’s more, I’m not sure anyone in the NFL has any good, hard evidence on this score, or has ever commissioned a study to quantify any possible trade-off between height and efficiency (e.g., if shorter QBs are less effective passers, do they make up for that by being better runners or sack-avoiders?).
Maybe some of you dear readers will fill the empirical void. There are lots of possible ways to do this, but it won’t be easy. One problem is that there’s just not much variation in the heights of QBs these days: those now active range in height from six feet even to six-five. There aren’t any Eddie LeBarons (variously listed as 5’6” to 5’8”) anymore, and we don’t know whether that’s due to irrational prejudice or real inability to compete. Statistically speaking, though, the lack of variation in height in the remaining sample will make it difficult to test whether stature actually affects performance (holding other influences constant), or to accurately measure any trade-off.
Until and unless the numbers have been carefully and thoroughly crunched on this, all we’ll have is anecdotal evidence that something odd and possibly inefficient is going on in the league. As one Ravens official remarked after the draft, “can you believe this guy lasted until the bottom of the fifth round? If he was two inches taller, he would’ve been a first-round pick.”
In a league with a hard salary cap, exploiting possible inefficiencies is even more important than it is in baseball, where teams can overcome prior mistakes by simply inflating the payroll. And Ozzie Newsome has struck gold before. Eleven years ago, he stole a “stumpy” linebacker in the second round; Ray Lewis (6’1”) has been the core of the team’s defense ever since, and is a sure Hall of Famer.
There are no guarantees that Troy Smith will even make the Ravens’ practice squad in ’07, of course, much less start a journey to Canton. But at the cost of a low-round pick, the Ravens might have found a successor to Steve McNair (6’2”). Fans of underdogs everywhere will want to watch this situation closely.
–Steve Walters
Wages of Wins Journal 
JChan
May 1, 2007
Yeah, it’s amazing to me how they make decisions based on an inch or two of height and a tenth of a second in the 40. Shouldn’t they be choosing based on performance? I was amazed when I looked at the scouting report on BYU’s John Beck (not a BYU fan, but I live in Utah, so I was interested). He’s listed at 6’2 1/8″, and the “weaknesses” section said:
“Lacks prototypical NFL quarterback size. Possesses below average height and decent bulk.”
I just don’t get it.
Guy
May 2, 2007
That’s a tough one. I’m a bit more inclined than you to believe the trend toward selecting and playing only tall QBs in an indication that height confers a real advantage. But it could of course be bias. Assuming it’s hard to detect a difference between current 6-1 and 6-4 QBs, the only way I can see to look at this is to analyze 1) college QBs and/or 2) NFL QBs from an earlier period (say, 1970s-80s). That would give you more variance to work with. (However, even if you find no height-performance link there, it wouldn’t necessarily prove that height isn’t important, given the increased height of today’s NFL pass rushers.)
TDDG
May 2, 2007
I think Ray Lewis was Baltimore’s second pick in the first round.
You’d think you could do a fairly good study of height among QB’s by looking at Division 1-A stats. Granted, the level of competition isn’t the same, but I’d think the height of the O-Line and D-Line is similar. If someone could show that there was absolutely no difference among QBs with differing heights in college, I’d find it pretty compelling, even when thinking about the NFL. It might not be the sort of evidence that academia would accept as proof, but weaker statistical studies are used to “prove” things all the time on Wall Street.
I think most teams should probably pick a QB in the 5th round most years. If you strike gold and its another Tom Brady, great. If not, it was only a 5th round pick. Compare that risk/reward with the history of 1st round QBs!
Steve Walters
May 2, 2007
Guy & Tom: The Div-I idea has promise. Using a logistic regression model to test “survival” of all QBs drafted over the years might also be fruitful.
I did a quick-and-very-dirty regression with a sample of 45 QBs who threw 100+ passes in a recent season, and regressed various measures of QB efficiency on height and height-squared. The coefficients on height flirted with significance at the 90% confidence level (which is to say, we really failed to reject the null of no effect of height on QB efficiency). One thing that was interesting was the estimated relationship was U-shaped: QB efficiency started high for the short guys, went down to a minimum at 6’2″, then up again. The problem is that with a small sample, the extreme observations (Brees, Vick at one end; P.Manning, Palmer at the other) have disproportionate influence. So this is worth studying.
I also like JChan’s focus on “prototypes.” I suspect personnel guys have “prototype bias” because if you make the safe, classic choice and he bombs, you can shrug and say he fit the mold; if you take a chance on a shrimp and he bombs, you’re an idiot.
Oddly, while I’m leaning in the direction of there being some bias in the NFL draft, I find myself suspicious of the econometric results in the much-publicized “NBA refs are racists” story discussed on the later posts. And in this I’m in agreement with Sir Charles Barkley, who pronounced the study “asinine” on ESPN tonight. We need more research on several fronts! Oh, and Tom is right about Ray being drafted in the first round, though late (26th).
Guy
May 3, 2007
Steve: I think “survival” is indeed an important measure to include in these studies. We’re assuming height is valued only because it impacts ability to see the field. But teams may also feel that you have to be big and strong to survive — literally — the hits that QBs take from 300+ lb. rushers. I’m old enough to remember when “Too Tall Jones” was a big man in the NFL, and today he’d be nothing special. The game has really changed, and it may be that most 5-11 QBs just couldn’t take the punishment. (Of course, this assumes there’s a link between height/weight and avoiding injury, which seems likely but may not be true. )