1. Introduction
In multi-shot measurements of pulse trains with pulse-shape instabilities, all ultrashort-pulse-characterization techniques are doomed to fail. Because the measurement averages over different events and can provide only a single result, any such result must necessarily be incorrect. The best that can be expected is a reasonable representation of the typical pulse intensity and phase vs. time and an indication of the stability of the pulse train or other indication of the reliability of the measurement.