Tamanho da fonte: 

We apply technical analysis (TA) in the Brazilian stock market using four chart patterns widely employed by traders. Tests combine a large sample (up to 191 assets) and a long time interval (2004-2014). By doing so, we aim to test the quality of chart patterns for price prediction, therefore endorsing them as useful tools for trading. Results highlight the quality of Inverse Head-and-Shoulders (IHS) and Double Top (DTOP) patterns to systematically deliver positive returns, with higher returns being obtained using longer trading windows. In addition, Monte Carlo Simulations (MC) present a clear tendency of TA to overcome the natural path of the stock prices. MC analysis (based on volatility modelled with GARCH) shows that IHS and DTOP generate returns above the 50th percentile of MC simulated returns in at least 60% of the trading operations. These results show chart patterns capacity to generate positive returns in the long run.