Study: Trading up in the draft is a sucker's bet

[Skyy God]

I'm crediting blind luck rather than Pioli's savvy for the success of the Baldwin/Houston trade.

Today the Philadelphia Inquirer profiles Cade Massey, a professor at Penn's Wharton School of business. Already with a study under his belt arguing that the conventional wisdom of the Draft Value Chart is all wrong, Massey was contracted by an unnamed NFL team to study the history of the draft for market inequalities. He discovered something that won't come as a surprise to football fans: the draft is kind of a crap shoot.

There is skill in making individual picks, Massey says, but the fact that draft success isn't sustainable points to the conclusion that every team is fairly evenly matched. What seem to be indicators that drafting is a talent, like the Lions' drought or the Patriots' boom of a decade or so ago, are statistically expected aberrations.

"Some teams have great years, other teams have bad years - and it matters," Massey said. "But those differences aren't persistent year-to-year, which tells me that they are chance driven. Something between 95 and 100 percent - I'm not exaggerating - of team differences in the draft is driven by chance."

If you take issue with that, you'll have to math it out with the math guy; I'm just passing things along. But Massey's field of study seems perfectly designed to tackle the NFL draft— according to his site, his expertise in psychology and economics hones in on "judgment under uncertainty, with a focus on optimism, overconfidence, and learning."

Last year, he co-authored a study with the University of Chicago's Richard Thaler. Entitled "Loser's Curse: Overconfidence vs. Market Efficiency in the NFL Draft," it aimed to determine whether there are any patterns in how front offices value draft picks, and if those patterns expose an opportunity for greater value.

The study is embedded below, but here's a talking point: teams overvalue higher picks in part because they overvalue their own judgment in evaluating players, and that "overconfidence is exacerbated by information." The more front offices know about a prospect, the more they think they know, and they also assume other teams value that prospect as highly, creating a feedback loop that pushes players higher up the draft board than they may deserve.

In terms of practical takeaways, the study says to toss out that hoary chestnut, the Draft Value Chart. You know the one: teams have consulted it for decades to put a rough value on any trade including multiple picks. (Example: the second overall pick would be exactly equal to a sixth and a 16th.)

Massey says the chart is wrong, because it doesn't take into account salaries in a salary cap league. His study looks at "surplus value," or what a player actually gives you compared to what would be expected for his contract worth. Because rookie contracts keep salaries artificially low, the surplus values of draft picks are nearly always positive. But some are more valuable than others.

That treasured first pick in the draft is, according to this analysis, actually the least valuable pick in the first round! To be clear, the player taken with the first pick does have the highest expected performance, but he also has the highest salary, and in terms of performance per dollar, is less valuable than most players taken in the second round.
Here's the chart showing the surplus value of picks from drafts between 1994-2008 (when, it must be noted, the lack of a draft slotting system severely inflated the top picks' contracts). According to the research, teams gained more value from drafting late in the first round than they did early in the first.

Massey and Thaler have advice for GMs: only suckers trade up. Over those 14 years of drafts, they calculated the outcomes of every possible 2-for-1 trade for a first rounder using the Draft Value Chart, and found "overwhelming evidence that a team would do better in the draft by trading down." The team that would have traded down would have gained an average of 5.4 man-starts per season, with roughly the same amount of Pro Bowl appearances, at a cheaper cost.

The study was naturally controversial, in part due to misreadings. It deals only in probabilities, not in individual picks. If you trade up and land a player who turns out to be a superstar, it was a good trade. The study merely says that in most cases, that possibility doesn't justify the risk.

It also has little chance of making an impact in actual front office behavior, because of the very psychological barriers it cites. In terms of "impact"—tickets sold, media coverage garnered, general excitement—a top pick is always going to be desirable. Even more important is the corollary: the fear of missing out on a superstar. That's the kind of thing that costs GMs their jobs, and leads to moves that look sexy in the short term but don't work out over time.

http://deadspin.com/study-nfl-teams-...the-1378701238