Tuesday, July 03, 2007
Trading Lemons and CherriesPosted by Patrick DiCaprio at 7:54am
Have you ever thought "why is it every time I trade for a guy he ends up as the second coming of Neifi Perez (or maybe Ray Oyler for you older guys)?" There may be a good reason for this, and it isn't that those long hours searching MrSkin.com for scenes of Scarlett Johannsen are now coming home to roost; it is more likely that there are informational factors involved.
The 2001 Nobel Prize in Economics was, surprisingly, not awarded to Barbra Streisand for her trenchant observations on American tax policy, but to George Akerlof, A. Michael Spence and Jospeh Stiglitz, who had the following idea: suppose in a population of cars, some will be defective lemons and some will be reliable cherries. The owner of the car knows, from repair bills, which is which, however the buyer does not. This disparity in information results in the market having many more lemons than cherries. Why? Because anyone with a lemon will try to sell it but the owner of a reliable cherry will have little incentive to sell. This principle has been examined in the major league baseball trading context, by Phil Birnbaum of the SABR Statistical Analysis Committee in SABR's Baseball Research Journal #34, with the quite illuminating result that indeed there appear to be many more baseball lemons than cherries on the trading market.
Does this principle apply to the fantasy baseball trading market? In "real" baseball the assumption would be that the team has inside information, namely the players "repair bills," that others do not. In fantasy most players don't have this information, though the expert player has it or knows how to obtain it; there are a few such sources, including the columns here at The Hardball Times. Fantasy players all have access to some powerful information of their own to use, on the injury front and on a more involved analysis of a players statistics.
With the proliferation of information on baseball sites and with a proper understanding of statistics, this inside information is fully within any owner's grasp, and the disparity of information can be fully exploited by either the trader or the tradee. For fantasy purposes, a "lemon" is a player whose perceived value is much higher than his actual value. This can be because of injury, but it can also be because of the owner's knowledge of a player compared to other owners.
On the injury front, in one fantasy league I owned Bartolo Colon and traded him for a second round draft pick in the offseason, in a trade that I touched upon last week. The league is AL only, and each year there is a draft of all available free agents and minor leaguers that have played in AA. The key to the league is that minor leaguers cannot be picked up during the year, and must play at least one game at AA to be drafted in the following year's draft.
The parameters of the league, much like the terrain in a battlefield, guide the owner to a winning strategy: top prospects must be identified and much preferred in the drafts, even over average everyday major leaguers. Most top prospects moving up to AA during the year are available for draft. The league allows players to be kept for as long as you are so inclined, so these picks are tremendously valuable. In fact, what we have seen is that teams who draft minor leaguers well and eschew opportunities to pick pedestrian major league players have finished in the money for a decade in a row. As some examples, we have drafted Jason Giambi, Derek Jeter and Albert Pujols while in the minors.
A review of Colon's injury reports indicated that his condition was chronic, and since he opted for rehab over surgery it was not likely to result in a good outcome. So, we traded Colon virtually immediately; if this information was correct then once the season started he might start out OK but could rapidly deteriorate. Who needs that risk?? Had the opposing owner delved more deeply or had more information he probably wouldn't have made the trade. Note that Colon's performance is virtually irrelevant; even at Colon's full value a second round pick in this league is probably more valuable than what can be expected of Colon this year.
What are the key points for the Fantasy GM?? Here are two salient ideas:
1. You must be able to figure out when you have a lemon. This requires not just the information that everyone knows; you must dig deeper and seek out true expert opinions (just read some of the stuff Derek is writing here on the blog and you will know what I mean). It isn't just about injuries; an examination of a player's hit rate, for example may inevitably lead to the conclusion that he is due for a big fall, or a pitcher may have been very lucky with stranded runners. These players are lemons. As indicated in a prior column, I considered Stephen Drew to be a lemon.
2. You must know your opponent. This is a key principle of all games, and fantasy baseball is no different. In any trade you need to determine what your opponent's motivation is, and if they are an expert or at least very knowledgable, you need to be more discerning. An owner who does not, for example, know about the various expected ERA or Batting Average metrics such as BABIP, or hit rates and strand rates for pitchers, is probably not trading a lemon; the owner who does is much more dangerous. The owner getting Colon in the above example most likely didn't have enough information, or if he did he chose to ignore it. This cuts both ways, not just when you are the trader. Or perhaps he outwitted me, and knew Colon would be pitching and pitching fairly well!!
Another owner in my high stakes league was recently shopping Chipper Jones and Eric Byrnes to me (and others). Since the opposing owner is sharp and knowledgeable, I assumed that he was trying to sell high and expected a drop off, and closely looked at Eric Byrnes. Byrnes has a current 35% hit rate, and clearly should not maintain that rate, given his career batting averages. This owner likely knew that information. Likewise, Chipper also had a 35% hit rate. In a trade that I discussed on my blog, he posted a trade for Byrnes and Chipper that netted him Ian Snell and Carlos Delgado with Andy LaRoche throw in for good measure. This is a prime example of trading lemons for cherries.
I returned the favor and countered this competitor's trade by doing a similar maneuver; trading Matt Morris (and his unsustainable ERA south of 3.00 at the time) and Randy Winn (he of a 36% hit rate) for Rafael Furcal and Mike Wuertz. Now I just have to hope that Furcal reverts to form, and one could easily argue that Furcal is a lemon right now. So perhaps I should develop a taste for lemonade.
Patrick is a member of SABR's Statistical Analysis and Science of Baseball Committees and writes about fantasy baseball at The Fantasy Baseball Generals blog. He has achieved the dream of all of his MIT classmates. No, not making millions in the tech markets, but writing about baseball for free. Feel free to send along all insults and comments here.