We have posed a question whether the differences between various predictors affect one of the ultimate goals of prediction, generating synthetic amplitude variation with offset (AVO) gathers to serve as a calibration tool for interpreting the seismic amplitude for rock properties and conditions. We address this question by evaluating examples in which we test several such predictors at an interface between two elastic layers, at pseudowells, and at a real well with poor-quality S-wave velocity data. The answer based on the examples presented is that no matter which predictor is used, the seismic responses at a reservoir are qualitatively identical. The choice of a predictor does not affect our ability (or inability) to forecast the presence of hydrocarbons from seismic data. We also find that the amplitude versus angle responses due to different predictors consistently vary along the same pattern, no matter which predictor is used. Because our analysis uses a “by-example” approach, the conclusions are not entirely general. However, the method of comparing the AVO responses due to different predictors discussed here is. Hence, in a site-specific situation, we recommend using several relevant predictors to ascertain whether the choice significantly affects the synthetic AVO response and if this response is consistent with veritable seismic data.