A note on stochastic survival probabilities and their calibration