A wave of recent research has studied the predictability of foreign currency returns. A wide variety of forecasting structures have been proposed, including signals such as carry, value, momentum, and the forward curve. Some of these have been explored individually, and others have been used in combination. In this paper we use new econometric tools for binary classication problems to evaluate the merits of a general model encompassing all these signals. We nd very strong evidence of forecastability using the full set of signals, both in sample and out-of-sample. The holds true for both an unweighted directional forecast and one weighted by returns. Our preferred model generates economically meaningful returns on a portfolio of nine major currencies versus the U.S. dollar, with favorably Sharpe and skewness characteristics. We also nd no relationship between our returns and a conventional set of so-called risk factors.