>In the Prisoner’s Dilemma, a player starts off with a small set of sample strategies to be tested against the opponent. Each strategy is a set of stimulus-response rules, where past history is the stimulus that determines which play is to be taken in response. Adaptation then involves (1) assignment of ratings to each of the strategies on the basis of experience, and (2) invention of new strategies to replace those that end up with low ratings. >The rating of a strategy is merely the average of the payoffs it receives when it is used against the opponent. >The genetic algorhythm uses these ratings as fitnesses, and generates new strategies accordingly. >We can anticipate which building blocks (schema) are will be used by the genetic algorhythm in the prisoner’s dilemma, because we know that tit-for-tat is a favorable strategy. >According to the schema theorem for genetic algorhythms, such building blocks should appear ever more frequently as new strategies (strings) are generated. >Moreover, as building blocks at different positions become more common, crossover can combine them, providing offspring with still more building blocks. >In economics, credit assignment and the genetic algorhythm provide each agent with experience-based rules for buying, selling, and holding stock. An agent might develop rule of this form. IF (the price is 40), THEN (sell), and IF (the prices is below 40), THEN (buy). The markets smooth out as in classical economics. >Then, one of the agents finds a rule that exploits the market’s “inertia”, making money by selling a bit “late” in a rising market. Other agents begin to anticipate trends, and the whole learning process yields a market which makes the trend projections self-fulfilling – for a while. >Over time, after enough self-fulfilling prophesies, the behavior becomes more exaggerated, leading to a bubble, and eventually to a crash. The whole process seems quite normal and not the least surprising, in this framework. >When we “dissect” the agents, we find sets of rules that mimic, in this simple letup, the well known market strategies such as “Chartism”.