Little research has been done into implied correlations, and the small literature grows even smaller when referring to currency options. The existing literature has established that implied correlation is a good if not the best forecaster of future realized correlation, and that this ability to forecast is not necessarily universal. This paper will establish that the forecasting ability of implied correlations in currency options varies across currency pairs, thus proving that not all implied correlations are created equal. Using two different proxies for the quality of the forecaster, the paper attempts to explain which characteristics of an option on a currency pair affect the variation in forecasting ability.