From the little I know about these kind of experiments, they're a good start and useful groundwork, but in my opinion too simple to reveal much about the real thing as a whole. I don't think real insight will be gained until we start to model and analyse large(ish) complex, dynamic systems. I'd be interested to hear if anyone knows of any such attempts.
There has been some analysis of eBay, MMORPGs and related online markets. I know of one study that compared eBay's rules that encourage late bidding to Amazon's rules that manage to get people to bid early. (Amazon extends the deadline after late bids so that at least 20 minutes remain for other bidders to react. This removes the advantages of late bidding.)
Check "The Origin of Wealth" One of the conclusions is precisely that economy is a very complex dynamic system and advance techniques and models need to be built to understand it. Also, our capacity to make predictions is very limited. Free markets can be efficient but sometimes they are not the best mechanism.
