Market analysts and central banks often use the implied volatility of FX options as an indicator of expected exchange rate uncertainty. The aim of our study is to investigate the limits of this statistic. We present some key factors that may deviate the value of implied volatility from the exchange rate variability expected by the market. These biasing factors are linked to the simplifying assumptions of the Black-Scholes option pricing model. Our empirical results show that forint/euro implied volatilities carry useful information about future exchange rate uncertainty when the forecast horizon is shorter than one month. However, implied volatility provides a biased estimate, and does not encompass the information included in other (GARCH, ARMA) predictors of volatility calculated from historical exchange rate data. These results are in line with the findings of similar analyses of other currency pairs.