Abstract
This article suggests that the emphasis on forecast evaluations should be redirected from searching for “best” forecasters to finding those who are “better.” A new methodology is developed, and the results indicate that it is possible to identify “better” forecasters.