Currencies are used by many investors as a speculative or investment instrument. Their movements are complex and there are several fundamentals based models that suggest how to predict long and short term fluctuations. In this paper we investigate the predictability of an exchange rate with an entropy risk factor model. Considering that recent studies suggest that financial markets have complex systems characteristics, we explore the usefulness of sample entropy as a risk factor for currency fluctuations. The empirical testing of sample entropy is based on the data of the Chilean peso (CLP) exchange rate for the period of January 1, 2005 and November 25, 2016. We find evidence that Chilean peso market has enough market inefficiencies that can be profitably exploited by the sample entropy based algorithm developed in this paper. Although our results are based on out-of-sample performance of the active strategies, there could still be a chance that the level of effectiveness is sample specific.