A method for monitoring the effectiveness of VSR and for taking action to improve the effectiveness of VSRs, if they are determined to be ineffective, includes comparing the a VSR evoked electrogram to a template electrogram of a pure biventricular paced CRT beat. If the electrograms, or features thereof, are similar, the VSR is determined to be effective. If the VSR is determined to be ineffective, the AV delay of biventricular CRT is shortened in a step-wise fashion in an incremental manner.
