This dissertation presents an alternative tool for selecting buying and selling moments for stocks by investors who demonstrate a level of risk-aversion. This procedure, inspired by statistical process control chart, adopts the exceedance of certain limits by data as indicative signals of favourable moments for buying and selling. The monitored data are the residuals resulting from the application of time series models to the logarithms of daily returns of closing quotations of the assets taken in consideration (the Ibovespa index, ten brazilian blue chip stocks and the Dow Jones index, from July 1994 to April 2003). Based on the rationale that a sudden trend or shift on the data would make the prediction errors (residuals of the series) increase beyond certain levels, the comparison of these residuals with pre-defined limits can be used as signals of these trends and shifts. In the case under analysis, the choice of values for the limits was done experimentally, with 420 combinations (pairs) of values being tested, one value to signal buying moments and another to point out selling moments. The average returns and standard deviations resulting from the simulated operations with each pair of limits were used as basic data for an statistical analysis that highlighted the most efficient pairs. After choosing the best values, the result of their use were tested on the last semester`s data, verifying that two of the combinations (pairs) produced the most consistent results. The comparison between the results obtained by the use of the proposed procedure (with each of the two pairs of selected values) and the results provided by other investment options (for example, stock funds) revealed that, for the test period, using the proposed procedure, it would have been possible to achieve superior rentabilities comparing with the other options. These results demonstrate the potential of this procedure. However, being this work an empirical study, it is advisable to carry on a deeper investigation before recomending its use. The main issues to be brought about by this research are indicated as directions for further research.