At 08:22 AM 5/2/01 -0600, Robert Feldman wrote: -snip- And at 07:51 AM 5/2/01 -0700, Marvin wrote: Sigh. You guys are taking the position of lots of economists when discussing real-world markets, saying that the theory supports the well-behavedness of the market and neglecting the fact that the assumptions that were necessary to prove the theory do not apply to the real problem. In the auction institution that eBay uses, which is technically a demand-side bidding, raise-only, interactive auction with sealed bid caps and fixed increments, for the clearing price to reflect the actual market price you need many assumptions, not the least important of which is something called "rational bidding behavior". That means that the buyers have done their homework, examined their costs and benefits and determined an optimal strategy (i.e., a "maximum price" in the case of single lot auctions as in most of eBay auctions) that they are willing to pay. However, it takes only _one_ irrational buyer to bias the clearing price by the minimum price increment, and it takes only _two_ irrational buyers to transform the auction into a game whose payoff (or rather, whose negative payoff) will be bounded only by how stupid the buyers are. Bidding contests are proof of irrational behavior in eBay and it happens all the time. There is a definite psychological component that fuels the escalation of the price: "I want to win even if it makes no sense financially". There is an inherent time dimension in the problem because when people lose the high bidder position they convince themselves to bid even more although they had decided on a max price earlier. Because of this time dimension, the longer the price stays high, the likelier it is to go up further. It is simple psychology, and undoubtedly irrational. But that's how it is, and there is nothing that you, as a buyer, can do to change the behavior of such bidders. However, you can mitigate to some extent the effect that such buyers have on the clearing price by bidding late in the auction, giving them less time to convince themselves to raise their bids. Thus sniping is born. Sniping is a sensible strategy in an eBay auction with irrational bidders. It has the effect of transforming the auction institution into something of a sealed bid auction, which can really mitigate the effects of irrational bidding. Furthermore, because irrational bidding tends to be specific to a particular auction (it doesn't necessarily carry over to set the price of subsequent auctions, because of the "bidder remorse" issue), repeated auctions will tend to stabilize closer to the fair market value. The fact that sniping is a standard mechanism used by lots of buyers on eBay just proves what everyone doing experimental economics has observed again and again: if there is one way to game the system, it will be found, and, once it is found, the strategy will spread and become commonplace. At our group, we have used experimental economics tools to analyze the behavior of open energy markets. We have tried almost every auction institution (first accepted offer, first rejected offer, discriminatory, Vickrey, variants of these, you name it). Even though the experiments are carefully setup to prove or disprove a given hypothesis, we are always amazed by the range of unexpected, highly creative behaviors that we had not thought of. If you care to read about our group's work, here's our web site http://www.pserc.wisc.edu/ and our web-based experimental economics market simulator http://www.pserc.cornell.edu/powerweb/ Carlos. -------------------------------------------------------------- Carlos E. Murillo-Sanchez carlos_murillo(a)nospammers.ieee.org