Comparison of diverse and random picking (three-way ANOVA with sigma restricted parameterization) in the case of fragment-like molecules. The SRD values in the case of standardization are quite different compared to the others. (This effect seems to be less pronounced for intentionally diverse molecules). Weighted means were used for the creation of the plot. The vertical bars denote 0.95 confidence intervals. (Manhattan and Soergel coefficients were omitted for clarity).