Why Free Agent Signings Turn Out So Disappointing

Adam Archuleta became one of the most sought-after NFL free agents in 2006. Several teams were interested in the playmaking strong safety, but the Redskins won the bidding, making him the highest paid safety in history at the time. Owner Dan Snyder signed-off on giving Archuleta a 6-year $30 million contract, with $10 million guaranteed.

To call the Archuleta signing a bust would be an understatement. He started only 7 games the next season and was traded to Chicago for a 6th-round draft pick the following year. Archuleta never returned to his early-career form, and washed out of the league after the 2007 season.

Although Snyder has a well-known, and well-deserved, reputation for overpaying for disappointing free agents, he's not alone. There's a phenomenon of auctions that makes overpaying for top free agent players all too common.

Consider fictional star running back Freddy Adams, a top free-agent in whom several teams are interested. Each of the teams has a hole to fill at his position. The scouts and executives of each team all have their estimations of the player's value. On a scale of 1 to 10, Team A feels he's a 7.1 in terms of future expected performance, so they offer $7.1 million per year. Team B estimates Freddy is a 6.2, and offers $6.2 million per year. Team C and Team D think he's a 4.7 and a 3.8 respectively, and make offers accordingly.


Mr. Adams is no dummy, and decides to go with the high bidder, Team A.

At this point we can't know Freddy's true value, which will be revealed only after he plays out his contract. But as long as the teams' scouts, coaches, and executives have any degree of competence, we know where his actual value would tend to be. Each team has its own biases and errors, and some will overestimate while some will underestimate his value. It's very likely Freddy's true value will lie somewhere between the high and low estimates.



Let's say that over the course of his contract, Freddy's true value turns out to be 6.2. Since Team A offered $7.1 million per year, and he was 'truly' worth only $6.2 million per year, Freddy was a disappointment on net, worth -$0.9 million per year.

This result was bound to happen. The team that most grossly overestimates a free agent's value will very likely be the team that offers the most and win the auction. The upshot is that free agents tend to be signed by the teams that erred the most in predicting their true worth. That's why free agent signings turn out so disappointing so often.

Known as The Winner's Curse, this phenomenon is a well-documented characteristic of auction-style transactions. Whether the object of the auction is a part of the wireless frequency spectrum, licenses to drill for oil, or securing the rights to a Pro Bowl-caliber safety, the top bidder will likely be the one who most overestimates the value of the prize.

In truth, the Winner's Curse applies strictly to "common value" auctions, where the prize would be of equal value to all bidders. This is rarely the case in reality, so to account for differing values to bidders, we can add a "synergy" factor to the mix. A player might be of particularly high worth to a team with a single 'missing piece', or a speedy pass-rushing linebacker might be of special value to a defense that plays a 3-4.

In the end, however, the synergy value of the prize is just as susceptible to overestimation as its common value. A free agents would be just as disappointing to his team, which would likely be the team that overestimated both the general value and the particular synergy he could bring. (Check out this applet that demonstrates the Winner's Curse in action. You'll notice that the synergy factor needs to be extremely high to escape the Curse.)

Once the top free agent is signed in any given year, the market is now 'set' for other lesser players. The Winner's Curse tends to inflate the price of similar goods across the market place, creating a bubble. The real estate or stock markets may not be so different from the free agent market in the NFL. Which home buyer is going to be the one that ends up with that McMansion? The one who overestimates its value the most, that's who. And with each inflated sale, all home prices inch up one more tiny notch, at least until a market correction comes along.

Bidders who are aware of the Curse can mitigate its effect by suppressing their bids below what they believe is the true value of the prize, a technique known as bid-shaving. This makes it less likely a bidder will end up overpaying, but it also makes it less likely the wise bidder will win the auction. The unsavvy bidder (such as Dan Snyder, perhaps) won't shave his bids, and becomes that much more likely to win (and overpay for) the prize.

Perhaps the Winner's Curse explains why top teams tend to build their starting lineups through the top rounds of the draft and not through free agency. It may also explain why draft picks might be better bargains than veteran free-agents with similar expected performance levels.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

24 Responses to “Why Free Agent Signings Turn Out So Disappointing”

  1. Anonymous says:

    Brian,
    One thing to consider is a may be competing with a division rival. So the Jets slighly over paying for a player they need (lets say a DE) that is also needed by the Patriots may well be worth it.

    I think its the Synergy that helps make up for the difference. Free Agents have good chance of being over paid, so you had better be sue teh Synergy makes the contract worth it.

    Also - what is the alternative? Playing a really bad player? What is the value of the Free Agent vs the replacemnet level player? That varies from team to team.

    The Jets may over pay the DE because the replacement player (lets say Vernon Gholston) is awful).

    Does overpaying give you a better chance of making the playoffs and/or selling more tickets?

    I dont think this is a simple as the Java Applet makes it seem (i.e. dont over pay). Just don't over pay unless you have a good reson to.

    Free Agnets tend to be older so thats an additional negative.

  2. Charles Ranhofer says:

    Brian,

    While the winner's curse obviously plays a role in free-agent disappointments, I don't feel it fully explains why free agency is generally an unsuccessful strategy. It applies just as much to draft picks- a team will tend to draft the players it value more than the market such that the players on its team will tend to be worse than the players which were the market consensus of what was available at their draft positions.

    Rather, I think free agency is a matter of informational asymmetry. In the draft, all teams are at the same level with respect to information access to the same player; they go to the same pro days, talk to the same college coaches, see the same game-footage. (Obviously, some teams may do a better or worse job of developing this information, but the point is that everyone is on the same level in terms of what information they COULD get if they acted optimally).

    In free agency the situation is completely opposite. One team- the incumbent- has MASSIVELY more information than other bidders: they've seen the player in practice, in the weightroom, how he's interacted with other teammates, etc. The other team only has game footage and reputation- if the incumbent team has any interest in the player they're not going to be sharing info with a competing bidder to facilitate their bid.

    Therefore, in every free agent signing, the team with the most info, who should know the value of the player best, has valued the player at less than the competing bidders. Free agents should be systematically overvalued by their successful courtiers to a greater extent than that which would simply be explained by a winners curse in an informationally symmetic setting.

    Note that this argument does not apply in a sport like baseball where teams have vastly different economic power and an incumbent team can simply be unable to afford a player even if they view the new team as making a bid which is under "fair value". This also doesn't apply to the best FAs in basketball, since their value so-exceeds the level of the max contract that even a team which greatly underrates a LeBron James will pay him the exact same contract as the team which overrates him.

  3. ch says:

    "top teams tend to build their starting lineups through the top rounds of the draft"

    A simpler explanation then the winner's curse: good teams presumably are composed of good players. Teams renegotiate the contracts of the players they want to keep before they hit the FA market, and get to franchise one guy who won't negotiate (until the franchise number hits Julius Peppers numbers). Good players generally have a career that exceeds the NFL average length, and thus the percentage of your own guys tends to increase, probably levelling out somewhere around the average length of a good player's career, where you invest a high draft pick (or picks) on hopes of finding his replacement. You don't need a pricey quality FA for a spot you've got filled already.

    NFL players likely have a higher incentive to renegotiate an existing contract vs. other sports because of the risk of catastrophic injury is so much higher and NFL contracts aren't guaranteed.

    --

    I suspect that most teams would be super happy with getting $6 million dollars of value for a $7 million Freddie Adams contract. That's outstanding, Pro Bowl caliber performance. Archuleta's remembered as a FA bust because he was paid $10 million for a half season. That's a lot more then a -0.9 million deficit value.

  4. Chase says:

    Good article, and I think the "early rounds" of the draft is an interesting point. In a lot of ways, who gets a draft pick? The team that overvalues the player the most. This effect is significantly mitigated in the first round as there isn't much granularity among prospects. Last year, the Redskins saw Brian Orakpo basically fall into their lap, and almost everyone would have agreed that he was a steal there.

    But in the later rounds of the draft, nearly every pick is subject to the winner's curse. If you take a WR out of Minnesota in the 5th round, it's a safe bet that most teams didn't feel he was worth that pick. They would have picked him in the 4th round if they thought he was really good, or traded in front of you to get him. Usually, players in the later rounds (I think -- just speculating here) go to the teams that value them the most. I'd say most 5th, 6th and 7th round picks of other teams are viewed as "bad" picks by most GMs, but their own 5th, 6th and 7th round picks are viewed as shrewd ones.

    In the first and second (and maybe third) rounds, though, the lack of elite players combined with the lack of a liquid trade market (it's not particularly easy to just acquire a 2nd round pick) means sometimes you just get a really good pick.

  5. Jeff Clarke says:

    Charles is right. Information asynmmetry has a lot to do with this. Think of the old used car dilemma.

    90% of cars are good cars. 10% of cars are lemons. The owner knows whether a car is good or not. Its not immediately obvious to a prospective buyer. Owners of good cars decide to keep using them. Owners of lemons decide to sell them. Hence, while 90% of used cars are good cars, 90% of cars on the used car market are bad ones. This deflates the value for all used cars. Deflating the value makes it even less likely that good car owners will sell. Obviously, there are some people that sell good cars anyway (they can't drive anymore...they are moving to Manhattan....they absolutely need the status of being seen in the latest model...etc) and there are some tests you can do to weed out lemons.

    The basic point remains though. The owner knows more about how much the car is worth. If he is selling it for $3000, he obviously thinks its worth less than $3000 and he knows more about it than you do.

    The same thing really applies to the F.A. market. Dan Snyder routinely outbids incumbent teams and he routinely finds out about problems they already knew about.

  6. Anonymous says:

    I don't think you understand auctions very well. You seem to think that people always overpay at auctions, when in fact the converse is true. 90% of the items bought at auction are sold at below market value. This should be fairly obvious if you think about. Why would I pay more for a used car at an auction, if I can walk out the door and buy for less at a retailer down the street? I wouldn't and most of the time neither would anyone else.

    The winner's curse occurs only when the estimatation of value of a commodity varies greatly. This may be somewhat true about football players, but not to the degree that you state. I don't think that one team would estimate the market value of a player at 163% of another team. Probably much closer to 110%.

    You mention private valuation, but this not related to winner's curse. I would be willing to pay over market price for a piece of land next to one I own, but this land *is* more valuable to me than it is others. The fact that I paid more does not mean that I overpaid. That would depend on the accuracy of the estimation of my private value.

    The main reason that people pay too much at auctions is that they get into bidding wars and let their competitiveness override their judgement. This is infrequent however.

  7. Brian Burke says:

    "...but not to the degree that you state."

    "Probably closer to 110%."

    "This is infrequent however."

    How do you know these things? Are you Vinny Cerrato?

    Top free agents are not "commodities" as you state. They are the opposite. A commodity is something easily replaceable by another instance of the same thing--a gallon of milk, an ounce of gold, etc. Top FAs aren't "a used car" at all. They are an exotic Ferrari of which only a handful were built.

    There is no private value in football. Players have a value in terms of how much or little they help their teams win, or how much revenue they generate. There is a bottom line, absolute value at the end of the day. GMs don't get an inner satisfaction of having Adam Archuleta on their team like they would having a Picasso hanging on their wall.

    But hey, I guess I don't understand auctions very well, so forget I said anything!

  8. 陳哲毓念阿彌陀佛往生西方極樂世界 says:

    阿彌陀佛 無相佈施


    不要吃五辛(葷菜,在古代宗教指的是一些食用後會影響性情、慾望的植
    物,主要有五種葷菜,合稱五葷,佛家與道家所指有異。

    近代則訛稱含有動物性成分的餐飲食物為「葷菜」,事實上這在古代是稱
    之為腥。所謂「葷腥」即這兩類的合稱。 葷菜
    維基百科,自由的百科全書
    (重定向自五辛) 佛家五葷

    在佛家另稱為五辛,五種辛味之菜。根據《楞嚴經》記載,佛家五葷為大
    蒜、小蒜、興渠、慈蔥、茖蔥;五葷生啖增恚,使人易怒;熟食發淫,令
    人多慾。[1]

    《本草備要》註解云:「慈蔥,冬蔥也;茖蔥,山蔥也;興渠,西域菜,云
    即中國之荽。」

    興渠另說為洋蔥。) 肉 蛋 奶?!











    念楞嚴經 *∞窮盡相關 消去無關 證據 時效 念阿彌陀佛往生西方極樂世界











    我想製造自己的行為反作用力
    不婚 不生子女 生生世世不當老師








    log 二0.3010 三0.47710.48 五0.6990 七0.8451 .85
    root 二1.414 1.41 三1.732 1.73五 2.236 2.24七 2.646
    =>十3.16 π∈Q' 一點八1.34

  9. bytebodger says:

    Apparently, you're big in Japan

  10. johnnyjohnnywu says:

    Lame guy who wrote in chinese is trolling. This is the english translation:

    Do not eat Wu Xin (meat dish, in the ancient religion refers to a number of food will affect the temperament, the desire of the plant
    Materials, there are five main meat dish, collectively known as the five dirty, within the meaning of Buddhism and Taoism are different.

    Modern Times is claiming that the food contains ingredients of animal food is "meat dish" In fact, it is said in ancient times
    It as fishy. The so-called "Hun Xing" means collectively, these two categories. Meat dish
    Wikipedia, the free encyclopedia
    (Redirected from Wu Xin) Buddhist five dirty

    In Buddhism also known as Wu Xin, acrid of five dishes. According to "HeDongShiHou" records, for the large Buddhist five dirty
    Garlic, little garlic, Xing channel, Tsz onions, Ge onion; five dirty Health and taste by rage, people irritability; cooked food made lewd, so that
    Many people want. [1]

    "Herbal preparation to" comment goes: "Chi onions, shallots are; Ge onion, green onion mountain too; Hing Drainage, Western cuisine, cloud
    That China coriander. "

    Xing also said the onion drainage. ) Meat and eggs milk?!

  11. Brian Burke says:

    Yeah, me and Spinal Tap. I'll be there in June, so we'll see if I'm swamped by fans at the airport.

    You'd be amazed at some of the international interest in the NFL and this site. I get hits from all over the world, and not accidental ones either. They're coming from google searches for 'advanced nfl stats' or even 'brian burke nfl', stuff like that. Romania, Italy, Germany, UK, Japan, Pakistan. I think many are US embassy employees or military, but many are not. Lots of links are from international NFL message boards. I get hits on google.translate.com for Korean, Turkish, and lots more languages.

  12. Anonymous says:

    Brian,
    Just speculating, but maybe some of the visitors to your site are interested not in football, but in how you analyszed football. Maybe they want to use your techniques on analyzing Soccer.



    For teh NY Jets, how much of Thomas Jones' production was due to him, vs the offensive line?
    How will Ladanian Tomlinson do?

  13. James says:

    Honestly, not sure how much value you could get from Brian's posts on how to evaluate soccer from the way he looks at football. Soccer is MUCH more like basketball or hockey than football. How in the world would you calculate EP or WPA for dribbling past a defender?

  14. Brian Burke says:

    I think he means that searches are for "football analysis," meaning what we call soccer. There's is some of that, but there is a small but hardcore NFL fan base in Europe, mostly in eastern Europe. I can usually tell if someone is looking for soccer, they only visit with one click. The other guys hang out for hours clicking around dozens of pages.

    It's just something that surprises me and seems interesting.

  15. Anonymous says:

    "top teams tend to build their starting lineups through the top rounds of the draft"

    Do you have a working link? I'm looking for research on this topic.

  16. Brian Burke says:

    Try it now.

  17. James says:

    How much of FA disapointment is due to regression to the mean e.g. player has excellent season partly due to luck gets a big contract and then plays at his normal level the following season thus being a disappointment.

    There is a hardcore fan base in the uk of british fans thats why the international series games sell out in minutes. Sky sports (pay TV satellite channel) showed 2 games live every sunday last season and the previous season they showed 5 games on sunday and another channel showed MNF. Admittedly it is a cult in that the majority of UK sport fans have no interest neither does the media but those of us who are into it are probably as knowledgeable as many US fans.

  18. takeitdown says:

    There's a compounding problem, which goes to the synergy comment, within this whole FA issue. It's not just that teams misestimate the value of a player. That is naturally going to occur, and with something like a FA...a limited number of elements to choose from, a zero sum game, etc....one would expect the price of "absolute value" to be over and underestimated by something like 35% (yes a guess, but it means most teams would value a guy between 5 and 9 who's really worth 7). The further problem is simply there is no "absolute value."

    A player's value to different teams IS very different, because each system is so different. It's not so much that team A is just one piece away and team B isn't. It's more that DE A is very valuable to team A, because his weaknesses are an inability to read run quickly, and easily gets washed out of the play. His strengths are he's amazing at pass rush. For Team A, he's very valuable, because they have a heady thumber MLB, who's always at the right spot, and a pair of DTs who penetrate and clog up the G/T lane. Now, Team B has none of this, yet they're bidding on a "good DE." Team B has an extreme risk of overpaying. Though he was good on his former team, and could be great on Team A, he will likely be downright awful on team B.

    We see this year in and year out. Players seldom droff off a cliff at 27. Yet, FAs often bust out immediately. It's that teams don't evaluate very well how well that player will fit in their scheme. Teams are very cognizant of this in the draft, but a productive NFL player is already considered "vetted" and the same work doesn't apply. It's a bit like being accepted to Harvard because you were already accepted to Yale. "He played great for the Ravens, must be a really good player." I think this aspect contributes more to the bust nature. The Auction nature makes there be overpayment, but the lack of scheme specific knowledge makes it be far more likely the player truly busts.

  19. Anonymous says:

    an interesting side note, maybe:
    take this over into baseball and all bets are off. wealthy baseball teams basically just buy their way out of this scenario. i.e. it doesn't matter what the actual value is if you win it all. so what if you overpaid for everyone, you won. another argument for a salary cap in baseball.

  20. Tom says:

    Here's another reason. 16 games isn't a statistically valid sample, so in any given year, a player will generate statistics above or below his long-term average performance. If the last year before free agency is randomly above average performance, he is likely to regress to the mean in the first year of the new contract.

  21. Anonymous says:

    but if it were simply mean regression in a context where 16 games isn't enough to accurately value a player due to luck factors, I feel this wouldn't explain consistent UNDER performance of signed FAs. some players would actually be better than their previous performance indicated no, since luck by definition would be randomly distributed? so wouldn't such a phenomenon manifest itself in high variance rather than a skew? seems like the winner's curse and potentially the scheme fit explanations would be more reasonable.

  22. Anonymous says:

    The fatal assumption here is the Free Agency is a common value auction where the intrinsic value of a player is the same for each team. It could very well be that the value to each team is different because of scheme or current talent.

    FAs busting has more to do with information asymmetry and incorrect valuations than the 'winners curse.'

  23. NFL Stadiums says:

    Some players just don't succeed in some systems. We may have seen Archuleta fulfill expectations with a franchise like Tampa Bay or Baltimore. Who knows? I agree that the winner's curse doesn't exist. Just look at Jeremy Lin with the Knicks. He is thriving in D'Antoni's system.

  24. Anonymous says:

    Remember the Alamo!!!

Leave a Reply

Note: Only a member of this blog may post a comment.