“Bouncing Back” from a Loss: No Statistical Artifact

Abstract

In 1991 Nahinsky demonstrated that victories were more likely to follow losses than to follow victories in World Series competition, consistent with a “bouncing back” effect. Zentall stated the result was a statistical artifact of the analysis and presented an analysis he claimed supported that assertion. This article shows that his conclusions are based upon a misconception about the assumptions underlying Nahinsky's analysis as well as faulty assumptions underlying his own. Other aspects of analyzing processes of this sort were also considered.