Is trading indicator performance robust? Evidence from scenario building

This paper challenges widely applied trading indicators with regard to their ability to generate a robust performance. In this study, we use a semiparametric scenario-building approach to simulate artificial price series based on characteristics of the observed price. In addition to testing the trading indicators on the observed price series and holding back some observed data for pro forma out-of-sample testing, our price simulations provide a backtesting environment to test trading strategies on artificially created prices. This provides an additional performance assessment by allowing us to test the trading indicators for robustness on a large set of artificially created price series with similar characteristics to the observed price series. We find that many trading indicators deliver robust results for certain performance metrics but are unable to deliver robust results and improvements across all reported performance metrics. In addition, most trading strategies influence the statistical moments of the return distribution. While they improve the skewness – and thereby increase the number of positive returns – in most cases, they also increase the kurtosis, introducing undesired additional observations in the tails of the return distributions.