Head and shoulders pattern recognition in stochastic processes

Technical analysis is the process of analyzing a security's historical prices in an effort to determine probable future prices. On the other hand, in the context of Markov Processes (e.g. Wiener, generalized Ito etc) that form the basis of financial markets theory the best estimation for future’s price is current price, and the path that a stock price followed in the past is irrelevant to the future. The second concept is consistent with the weak form of Efficient Market Hypothesis where the first is not. In this paper we examine if technical analysis can predict the unpredictable. Particularly we apply an automatic rule-based mechanism which can identify the Head and Shoulder pattern on a stochastic process, and we examine firstly if the aforementioned pattern occurs, and secondly when the pattern occurs, what is the price’s future behavior.