Some comments on the above:remember that some years before many people were talking about Chaos Theory Applications in Asset Pricing. Now none. It seems that we just expected too much..Yet there are applications and tools that sound strange for people with a rather mainstream Economics/Finance background. Lets mention some: Genetic Algorithms, Neural Networks, Fuzzy Sets……..Don’t omit or neglect that the greatest innovations in the Field of Economics/Finance have been made through arbitrage among sciences. Examples? Lets mention the two most prestigiousMarkowitz took some statistical concepts combined with an Operational Research tool (Mathematical Programming) and a Computing tool(Simplex Algorithm) and applied them to the field of Portfolio Selection.(By the way I have noticed that this forum neglects OR applications in Finance, with the exception of simulation methods)Black & Scholes applied the heat equation (from Physics) in the pricing of options.Of course many other attempts of this “arbitrage” have taken place. Some have failed mainly because of inadequate problem specification. Being an economist is not just matter of knowledge. It is probably a “state of mind”, a type of though, a sort of a specific approach to a problem. An economist, a mathematician, a lawyer, and a computer scientist may approach the same problem each one with a different point of view.Thus, the main cause of failure I think is the lack of economic thought from people who attempt to make such types “arbitrage”

If we believe that the price of a traded security is determined by the interaction of the demand and supply curves of all participants in the market, then naturally we will want to understand how those curves evolve and what their dynamics are. Our investigation will lead us to realize that each individual has unique reasons for his demand/supply decisions,which are driven by yet more "atomic" components, and so on.It seems unlikely that there is any level of detail which can not be reduced still further. So what shall we do in this case? Perhaps we will postulate some set of factors which individual supply and demand curves change in response to changes in these factors. But that leads us back to looking at prices, which respond to the same set of factors, and various physical models of price behavior.I guess what I don't see is why modeling the dynamics of prices should lead to worse approximations than modeling the dyanmics of individual supply and demand curves.

MobPsycho,You are absolutely right. There is not much to do for a physicist in finance.Here is an example:The American Finance Association 2002 Presidential Address* Rational Asset Priceshttp://www.afajof.org/presadd.shtml*Note that the President started his academic career as a physics student at Oxford University (BA, Physics, 1972).

MP - let's say that I decide to only interact with a limited set of market participants. namely, the ones that I can get reasonably good information about their demand and supply curves. I have no doubt that having reliable information of this type on my counterparties would allow me to make better trading decisions. However, it seems I would be giving up one of the biggest advantages that capital markets have to offer. Namely, I would be restricting myself according to the geographic friction, to use your term, to a small subset of the market. While I might be able eliminate some uncertainty withiin this sub-market, there is no guarantee that I could not transact much more profitably with someone outside my sub-market.Instead of direct access to every buyer and seller, I now have n degrees of separation between myself and the counterparty to my trade. Is it worth it?