This analysis assumes (inmplicitly) that the Underlying environment is constant, unchanging. The climate system is not in a steady state right now! This means to me that the probabilites derived from past history, implicitly assuming a steady state (unchanging) climate, have a systemic error of unknown size that should at least be mentioned, especially given the tiny data set we are extrapolating from.