Does the frequency of events before and afterward settle down to a stable number as the time separation from the earthquake considered grows larger, as it did in the Pankow study? Maybe the fluctuations you are finding are not well resolved.