Un approccio sistematico all'analisi tecnica: sviluppo del modello e implementazione

The paper takes a quantitative and analytical approach to Technical Analysis, a disciple that has developed increasingly from the '70, with the improvement in computer power. The discipline has not received yet the same level of acceptance from the academic world, as more traditional approaches to the financial markets, and to the study of financial assets, such as fundamental analysis. The main reason for the lack of recognition is to blame to the subjectivity of the matter, especially in the analysis of geometric figures given by the price fluctuations, which often differ from one analyst to another one. So the need to apply stricter and replicable quantitative rules, to perform an objective analysis of data, and to come to definite conclusions. We obtained the data for some equity indices, government bonds and commodities in the last 20 years, and built a specific code on Matlab to detect the presence of technical analysis figures. Afterwards we analysed the data on the number of figures detected and data from random series specially built with the parameters of historical data under observation (mean return, volatility), to understand the nature of the phenomenon. Lastly, we filtered the results and performed an analysis directly on the graph of the asset under observation to determine if the figures have forecasting ability for the future movements of prices. The work is intended to be dynamic and to evolve beyond the scope of this Thesis. Our starting point is the work of Andrew Lo, Harry Mamaysky and Jiang Wang described in ¿Foundations of Technical Analysis: Computational Algorithms, Statistical Inference, and Empirical Implementation¿.