This paper analyzes the predictive performance of several forecasting techniques: (1) a leading series regression, (2) the independent components of these leading series, and (3) a diffusion index. The comparative predictive performance in forecasting the turning points of the FRB Index of Production was measured by means of lead-reliability functions. The use of lead-reliability functions requires rules for identifying true and false peaks. In applying these rules, it was shown that the estimated values of the lead-reliability relationship are sensitive to the rule which is used to identify the true turns. A comparison of the predictive performance of the three techniques shows that both the diffusion index and the independent components are better predictors than the regression.