We demonstrate how simple differences in dual momentum implementations can lead to annual performance differences up to thousands of basis points.