Skip to content

Market design and the NFL draft

2010 April 22
by Skip Sauer

Today’s WSJ has a good story on issues with the design of the NFL draft — good, with the exception of the silly title, “The NFL Draft Drives Economists Crazy.”   The title is off base because oddball problems like the NFL Draft are the puzzles upon which we economists thrive!

At his Market Design blog,  Al Roth provides an expert’s perspective on the issue.  Of particular interest is this reform proposed by three of his Harvard students, as quoted in the WSJ article:

Three researchers at Harvard Business School—who studied under Alvin Roth, a Harvard professor and a pioneer in market-design theory—have proposed an alternative to the NFL draft.

Under their plan, all 32 teams would be given seven picks. They would have to abide by a spending cap that would go higher to lower—with the worst team (based on its record the previous season) having the most money to spend. When the bidding opened, the most sought-after players would draw multiple bids. Teams could then raise their bid as high as they’d like for a player they coveted.

Theoretically, a team could get any player it wanted—so long as it was prepared to pinch pennies on everyone else. Meanwhile, a team that didn’t want to break the bank on any particular player could pick up lots of useful parts by spreading its money around evenly. Teams could also thrive by focusing on the bidding and looking for bargains.

“I think that it would significantly help teams get the right guys,” said Lucas Coffman, one of the study’s authors. If nothing else, Mr. Coffman said, the auction format might be more exciting than the draft, which allows for long gaps between picks.

This is an intriguing idea, and surely worth some thought.  The proposal’s inverse relation between team success and draft spending probably codifies  what happens with the current draft structure (and is an oddity of N. American sports).  I like the idea of uncoupling draft order and selection, in principle,  and wonder about the impact on pre- and post-draft trading of picks and players.

7 Responses
  1. April 22, 2010

    Good one Skip, thanks.

    Didn’t Steve Ross have this kind of idea for a modification of salary caps 15 or 20 years ago?

    BTW Massey and Thaler’s now famous working paper on irrational trading up for earlier NFL Draft picks was updated earlier this month too.

  2. Donald A. Coffin permalink
    April 22, 2010

    For an interesting take on the NFL draft–in terms of outcomes, not design, this is worth a look:

    http://joeposnanski.com/JoeBlog/2010/04/21/nfl-draft-in-the-1990s/
    and
    http://joeposnanski.com/JoeBlog/2010/04/22/nfl-draft-in-2000s/

  3. Frank permalink
    April 22, 2010

    This is just like Fantasy sports leagues that use “Auction” style picking as opposed to drafts, isn’t it? (with adjustments in salary cap to help worse teams)

    It’s a much better system for fantasy.

  4. CGG permalink
    April 22, 2010

    I like Roth’s idea, but there seems to be the risk for allocations if a highly value player is brought up later in the bidding and goes for more than expected. In which case, the team which would value him the most could bump up against their cap and regret some of their prior bids.

    The way around this, of course, is the simultaneous multi-round auction, like that adopted by the FCC in auctioning spectrum in the US. All bidders bid for every player simultaneously, and you don’t know what other bidders are putting up for each player until after the end of the round. Bidders see what the got, and didn’t get, and then you go through to the next round where bidders adjust their strategies. Once every bidder has chosen their optimal strategy, there will be around in which all activity ceases, and the auction is declared over.

  5. Chris A. permalink
    April 23, 2010

    I personally like it, but the NFL would never allow it because it would drive up the salaries of the best rookie players, in turn driving up the salary demands of the best established players once their contracts are up for renewal.
    Some, maybe most, NFL teams would do their best to sign as many players as possible to reasonable salaries. But it only takes a handful of extravagant contracts to be used as yardsticks by the established players when their contracts are up for renewal.

  6. Louis permalink
    April 26, 2010

    One has to be careful with exactly what this style would accomplish. By nature, it is a general rule that the bad teams in the league have more (in quantity) needs than the good teams. Perhaps the Colts could use a defensive tackle, but the Lions need just about everything. Chances are you see the Colts pay up for someone like a Suh, as the Lions can’t justify paying that much for one player when they have so many needs…and they end up with what amounts to 7 guys between 2nd and 4th round talent.

    Not to say this isn’t the best way to build a franchise (many have argued that a #1 pick is a curse because of today’s huge guaranteed money), but the way the league has crafted recent rules makes it a league of superstars and skill positions, and the highest profile names going to the worst franchises being the end game.

    Then again, with GM’s and coaches just playing the CYA job security game, odds are this format would end up no different than the current format. Keep in mind, teams can trade down at will, but rarely do, and teams can also just not pick and wait to use their pick in a cheaper spot, but NEVER do. Execs are too afraid of doing something unconventional and having it not work out.

  7. Steve permalink
    April 26, 2010

    When I was a kid playing little league this is the system that was used to put the teams together. They would have a “draft” which was actually an auction, and each team had a certain number of points to bid on various players given their number of open roster spots (there was some carry over between years) and finish the previous year.

    The trouble with that system was the fact that little league can be dominated by one or two great athletes, so they naturally were bid up and up and up. Teams could spend almost all their points securing the best players. Any players left over after all the points had been spent were put in a hat, and any team which needed more players literally drew names out of the hat to fill their rosters.

    I agree with Louis, there are differences in needs in terms of both quantity and quality of players. Some teams don’t have roster spots for even 5 new players. Some are so bad that 15 draft picks may all make a roster and make a team better.

    We also forget that draft picks are already a currency of sorts, being traded for players or being traded around to move up in the draft and get a player you want.

    @ Chris A

    But we already have that in the NFL. Rookie salaries, at least for the top picks, grow more ludicrous by the year. At the point at which Jamarcus Russell finally signed his NFL contract, he was making more money per year than LaDanian Tomlinson, who the prior season had been named the MVP, was the rushing champ, and had broken the all time scoring record. And looking back, Russell’s contract is less insane than the one given to Matt Stafford or the one about to be bestowed on Sam Bradford.

Comments are closed.