The application of a technical trading rule, which just provides long and short signals, requires the investor to decide upon the exposure to stake in each trade. Although this position sizing (or money management) crucially affects the risk and return characteristics, recent academic literature has largely ignored this effect, leaving reported results incomparable. This work systematically analyzes the impact of position sizing on timing strategies and clarifies the relation to the Kelly criterion, which proposes to bet relative fractions from the remaining gambling budget. Both erratic as well as different relative positions, i.e. fixed proportions of the remaining portfolio value, are compared for simple moving average trading rules. The simulation of parametrized return series allows systematically varying those asset price properties, which are most in uential on timing results: drift, volatility, and autocorrelation. The study reveals that the introduction of relative position sizing has a severe impact on trading results compared to erratic positions. In contrast to a standard Kelly framework, however, an optimal position size does not exist. Interestingly, smaller trading fractions deliver the highest risk-adjusted returns in most scenarios.
Kelly criterion money management parameterized simulation position sizing technical analysis technical trading timing strategy