This paper examines the profitability and risk of a naive strategy of forward exchange speculation during pegged and flexible exchange rate systems. It suggests a new method of calculating these speculative returns and shows that the average return was much less than previously supposed and that the distribution of returns differed between systems and countries.