Even if such observation of yours is an objective unbiased statistical estimation… statistical deviations as large as M are also expected on one out of N times, the trial factor correction is non trivial at all… even without multiverse assumptions