This article assesses the economic significance of the non-linear predictability of EMS exchange rates. To that end, and using daily data for nine EMS currencies covering the 1 January 1978-31 December 1994 period, it considers nearest- neighbour non-linear predictors, transforming their forecasts into a technical trading rule, whose profitability has been evaluated against the traditional moving average trading rules, considering both interest rates and transaction costs. The results suggest that in most cases, a trading rule based on a non-linear predictor outperforms the moving average, both in terms of returns and in terms of the ideal profit and the Sharpe ratio profitability indicators.