Writers of Pro Football Prospectus 2008

Most Recent FO Features

CoxMar15.jpg

» Varsity Numbers: Honing in

Bill Connelly again looks at which college football teams the F/+ ratings are sure about, and which teams remain a mystery (led by Appalachian State).

Loser's Curse Haunts the First Round

by Michael David Smith

Football fans take it on faith that the San Francisco 49ers, who hold the first pick in the NFL's annual draft on April 24, have the best chance to improve of any team. The team, after all, has its pick of the best and brightest, and should emerge with the most talented and promising amateur player in the world.

But two business professors who have studied the draft extensively say the big contract given to the first pick is a millstone that actually puts San Francisco in the worst position of all in the first round.

The professors, Cade Massey of Duke University's Fuqua School of Business and Richard H. Thaler of the University of Chicago's Graduate School of Business, analyzed every selection, every trade, and every player in every NFL draft since 1988. They published their findings last month in a paper titled "The Loser's Curse: Overconfidence vs. Market Efficiency in the National Football League Draft."

The professors' conclusions centered around the exorbitant cost of players chosen early in the draft. The NFL's salary cap dictates that every dollar spent on a high-priced rookie is a dollar that can't be spent on improving the rest of the team. The player taken with the first pick typically costs four times what the last first-round pick costs and, therefore, is actually of less benefit to his team.

So while the Giants were delighted to acquire Eli Manning last year, their young quarterback's six-year, $54 million contract is exorbitant compared to the production teams typically get out of the first player selected. Ben Watson, the tight end who went at the end of the first round to the Patriots, may actually end up a better pick because he comes with a contract of only $13.5 million for six years.

Because of the salary cap, a better way to judge players than "who is the best?" is "who is the best per dollar?" And by that measurement, every selection in the second round is better than the first pick in the first round. It turns out that the most valuable pick in the draft in terms of getting a good player without breaking the bank is the 43rd overall, which is the 11th pick in the second round.

Last year, that pick belonged to the Dallas Cowboys, who took running back Julius Jones out of Notre Dame. Jones ran for 819 yards and scored seven touchdowns in just eight games and cost Dallas only $4.37 million on a six year contract. The professors' graph shows peak value at pick no. 43 and then a steady decline through the rest of the draft, with one small blip toward the end caused by a single outlier, Patriots quarterback Tom Brady, who was selected with the 199th overall pick in 2000. Brady aside, top picks are better players than later picks - but not by a wide enough margin to justify their significantly higher salaries.

Teams often target a specific player and trade picks to move up in the draft order because they're sure other teams want the same player. This is another instance of misperception, Massey and Thaler argue; teams wrongly assume they won't be able to select the player they want unless they trade up.

Cleveland Browns coach Butch Davis illustrated that classic mistake last year. Davis coveted University of Miami tight end Kellen Winslow so much that he shipped his team's second round pick to the Detroit Lions solely so he could move from the seventh pick to the Lions' sixth and take Winslow.

Davis was afraid the Lions would take Winslow for themselves or trade the pick to another team. But reports since then have indicated that the Lions' intention all along was to take Texas receiver Roy Williams with the sixth pick. So the Lions got a free second-round pick for moving down and taking Williams, who they wanted anyway, while the Browns sacrificed their all-important second-round pick to move up and grab Winslow, who would have been available anyway. And the reversed order meant the Browns had to pay more for Winslow than the Lions had to pay Williams.

Every team knows the market for picks; a widely used chart gives a numeric value to every pick in the draft, and Massey and Thaler have found that the chart closely approximates the actual value that teams have spent in trades. But that chart overvalues the highest picks, and in some cases, like the Browns' draft-day trade last year, teams overpay even compared to what the chart suggests.

Though the professors' research is new, it corresponds with the innovative draft strategies employed by the Washington Redskins in the 1980s. General Manager Bobby Beathard felt strongly that trading out of the first round and getting lower-priced players in the second round was the best choice. Beathard traded away first-round picks seven straight years in the 1980s; the team he built won three Super Bowls.

The best illustrations of the professors' line of reasoning come from the actual results NFL teams have had. The Arizona Cardinals' recent draft history has been filled with players such as Andre Wadsworth, a defensive end selected third overall from Florida State whose career was marked by major injuries (knee problems limited him to only eight sacks in his Cardinals career) and major paychecks (Arizona paid him more than $12 million).

Even young players who look promising, like the Cardinals' first-round pick (no. 3 overall) in 2004, Larry Fitzgerald, exact a heavy financial toll. Fitzgerald's six-year contract gives him $20 million in guaranteed money. Meanwhile, Michael Clayton, the receiver Tampa Bay selected with the 15th pick, got only $6.2 million in guarantees on his six year deal. And Clayton had a better rookie year than Fitzgerald.

On the flip side are the Patriots, who have had only two top-10 picks in the last 10 drafts. One of those picks, Terry Glenn, has long since departed, while the other, Richard Seymour, remains an integral part of the team. But the absence of high picks has given the Patriots the leeway to spend money on other players, and that wise spending has made them the best team in the league.

One reason teams overvalue high picks is that they're overconfident of their ability to pick the best players. Every team makes lists of all the available players and sorts them by position, ranking, for instance, the top 50 offensive tackles. Although teams are accurate in the general assessments, such as which player deserves a first-round grade and which deserves a second round grade, they do badly when picking within those general assessments.

Scouts and general managers (who use hours of film study and reams of data to determine which player is, say, the 12th-best at his position and which is 13th) fare about as well as they would if they flipped a coin. Take any player chosen in the draft, and there's a 49% chance that the next player chosen at his position will make as many or more Pro Bowls and a 47% chance that he'll start as many or more games.

The ideal strategy for the 49ers would be to call every team and offer to trade down. If they could find enough trading partners, the 49ers could get six mid-second-round picks -- even though all six of those picks are actually more valuable than the first overall.

That strategy illustrates just how bad a move it was for the Giants to give up the rights to Philip Rivers, a third round pick in 2004, and first and fifth round choices in 2005 in exchange for Manning.

"We thought it crazy for the Giants to give up so many picks for the opportunity to move up from the fourth pick to the first one," the professors write. "Rather than a treasure, the right to pick first appears to be a curse."

This article originally appeared in the Friday edition of the New York Sun.

Comments

1
by Mike (not verified) :: Thu, 08/25/2005 - 7:50pm

I know it's an old topic, but....

"As has been stated, a team would get absolutely crucified by the Media if it traded the #1 overall for like a 2nd, 3rd, 4th,5th, 6th, and 7th round draft choice."

I don't agree. If Ditka got destroyed for giving up eight picks for Ricky, then a GM shouldn't be criticized for being on the other end of a similar deal. In the two highest-profile cases of a team giving up many picks for a high-profile player - the Herschel Walker deal being the other - the team stockpiling picks got much the better of the deal. Given that, I don't think a team would be crucified for giving up a high #1, even the #1 overall, for a ton of picks.

2
by Ken (not verified) :: Fri, 08/26/2005 - 9:37am

Another reason why players new to the league should have a fixed play level. If you get drafted in round 1 you get X, round 2 you get Y... etc.

3
by bobby mozitis (not verified) :: Fri, 08/26/2005 - 11:36am

#71

Your idea of taking a player who isn't projected as highly and then paying him less than a number 1 pick would normally get is similar to the draft strategy the Oakland A's use. The Niners could have picked their favorite player who was otherwise projected to be a second rounder and given him a contract worth more than he would have gotten in the second round, but significantly less than the #1 overall pick would usually get. Of course, you still have to deal with the PR hit and sports-talk radio complaints, but a good franchise worries about putting the best product on the field, knowing that Wins and Losses are what ultimately bring in the money, not "star" players.

4
by Bad Doctor (not verified) :: Fri, 08/26/2005 - 12:55pm

#77 ... I like the thought, but one fundamental difference is that in baseball you can't trade draft picks. Oakland had four first rounders as a result of not being able to pay their own free agents ... but they also weren't able to pay four normal first-rounders either, and they couldn't trade the picks for value. Their method was a necessary resolution to that problem. The Niners could trade down ... I know there wasn't much interest in the first pick this year, but even if they only got what is the usual return for the 5th pick or 10th pick, that would probably prove more valuable than using the A's Moneyball method.

5
by bobby mozitis (not verified) :: Fri, 08/26/2005 - 3:19pm

#78

yeah, you are right. The easiest solution is just to trade the pick, even if they don't get much in return. they better really hope Alex Smith works out for them. Like aaron points out in the 49ers section of the Prospectus, all the effort they have made to scrap their old team and re-build in the mold of the Pats or Eagles could be sidetracked by Alex Smith's contract.

6
by Chuangtzu (not verified) :: Fri, 08/26/2005 - 5:13pm

You forgot another prominent example of a team giving up tons of picks to move up in the draft to number one. That would be the debacle that crippled San Diego as they traded their future to draft Ryan Leaf.

7
by Goldman (not verified) :: Sat, 08/27/2005 - 2:26pm

I was wondering why this study has been getting so much attention this week. I knew about it months ago, thanks to FO. It seems like every other sports site on the Internet just heard about it. Weird.

8
by Tim (not verified) :: Mon, 08/29/2005 - 1:46am

After finally actually reading through the study, I have one question about the methodology. They make the tacit assumption that teams are attempting to maximize their "surplus value," which is the difference between the amount the team pays the player and (roughly) the amount the team would pay a veteran free agent for the same performance level.

(I'm also not certain about their development of the equivalent veteran costs. The performance level for the veteran derives from his performance during his first 5 years in the league. However, he is paid in his 6th year. That pay amount is compared with the pay amount of drafted rookies at a draft position whose average first 5 years' production level is equal to that veteran player's performance to date. So while it doesn't immediately seem like it, they're comparing apples to oranges. The rookies are getting paid for the 5 years they're about to have, and the veterans are getting paid for whatever years they're about to have, but they're valued in the study based on the 5 years they've just had. If you asked a team, "Whom would you pay more for, a rookie who in the next 5 years will go to one Pro Bowl and start every game for 4 of those years, or a 6-year veteran at his position who has gone to one Pro Bowl and started every game for 4 of his first 5 years," most teams would take the rookie. That's over and above the observed surplus effect, which essentially shows that rookies are underpaid relative to their production because of the draft and rookie allocation pool structure.)

Even assuming all their numbers are correct, I'm still uncertain that teams are better off getting the most value for the dollar, as opposed to the best player(s). However, I don't have any data to prove either way, and I don't even know what sort of form that would take anyway.

9
by wyote (not verified) :: Mon, 08/29/2005 - 8:19am

Great analysis.

I would definitely trade down consistently if I were an NFL GM.

If I were the 49ers, I'd have traded down; then traded down again, then traded down again and again. Four or five 2nd round picks are worth much more than the 1st pick of the draft.

10
by Jim A (not verified) :: Mon, 08/29/2005 - 11:37am

Tim, I thought I was the only one crazy enough to actually read the entire paper!

Technically, you're right on the first point, but I think the authors were reasonable to use veteran market-value salaries as a baseline to measure rookie performance/salary/surplus value. It's not perfect for the reasons you state, but in terms of using to compare the different draft positions, I wouldn't think there would be any unpredictable biases. They are simply evaluating draft slot values, nothing else. Unfortunately, the media has somehow picked it up as saying all rookies are overpaid, which couldn't be further from the truth. And after reading Mike Florio's attempt to twist this study to push his own agenda on ProFootballTalk, it's pretty clear that not only didn't he bother to read the study, he didn't even read the AP article about it that he links because the research findings don't support his view at all.

I think your second point is a more important one. Because of the scarcity of available players (due to highly specialized skills and the fact that most players are already under contract with other teams) you simply can't just go out and acquire a team full of good players who are all underpaid. Although it's important to keep your player salaries in check, you still need a quorum of top players to be competitive on the field, and those players are more likely to come earlier in the draft.

BTW, there's lot of other interesting related sub-topics covered in this paper. The discussion of the psychological overconfidence of player evaluation was fascinating, as was the authors' study of draft-day trades.

11
by fyo (not verified) :: Tue, 08/30/2005 - 9:16am

HALF of all probowlers were first round picks!

Check out Fig6 of the pdf (thanks JMM). Second round picks are only 40% as likely as first round picks to get pro bowl appearances. Sum it all up and about half of all pro bowl visits were by first round picks.

Even more surprising (astounding, even) is that IT DOESN'T MATTER WHERE IN THE FIRST ROUND THEY WERE PICKED! Check out the virtually flat line in Fig7.

That's just plain weird in my opinion. With a huge difference between the first and second round, how can there not be (much of) a difference within the first round itself? What's the reason behind this discontinuity?

Is there some sort of bias in the selection to pro bowl? Or something completely different?

12
by PL (not verified) :: Wed, 08/31/2005 - 1:39pm

MDS-

Are you just going one better than 42 for the answer to Life, the Universe and Everything, or is that a typo?

13
by Josh (not verified) :: Sun, 10/23/2005 - 3:51am

responding to 85 two months afterwards - I do think there is a bias in the selection to the Pro Bowl for certain positions, mainly defense and O-Line where conventional stats are not that useful as measures of performance. These are based mainly on reputation, and a 1st round pick who makes some plays here and there can easily develop a reputation that is better than he actually is.

14
by loopslike :: Tue, 05/10/2011 - 1:39pm

dental lab
I was damn sure they will win the running session. all their players were fit for the game. Their game-play was very attractive and nice. I support this team for their team-play. Their dribbling, passing is awesome in one word.

15
by Traitement de la cellulite (not verified) :: Wed, 07/20/2011 - 2:35pm

This game was mostly dominated by Loser's Curse. They really played very well.