The Financial Market provides perhaps the greatest challenge in the discipline of time-series prediction. While some make use of experience, talent or even superstition to try to beat the Market, this study built upon the synergy existing among soft computing techniques in order to predict the price of a financial asset one day ahead. More specifically, the classic method of Wang and Mendel (1992) was used to populate a fuzzy system with rules extracted from the data itself, being optimized by a genetic algorithm in two ways: (i) aiming to minimize the prediction error and (ii) aiming to maximize the expected return of a stock trading system. It was not possible to coadunate both objectives in one single system, but profitable results were achieved. The system chosen by the highest return in the validation set exceeded 25 per cent in out-of-sample profits. When statistically tested through Monte Carlo Permutation, it was estimated a probability of less than 4 per cent of achieving such profits by pure chance, validating the derived system. However, subsequent risk analysis demonstrated high volatility of the expected returns whereas a sensibility analysis carried out indicated that the method is highly sensible to the specific combinations of chromosomes used and, therefore, not reliable enough to be operated in real life. Nevertheless, possible improvements of the method are envisaged, such as changing the objective function and the target variable of the prediction. Only the continuity of the work, however, will show who will be more efficient: the Market or the algorithm proposed.