Small probabilities trend towards incomputability in tail events. The rate of error for the measure becomes so explosive that in that arena individual parameter

Small probabilities trend towards incomputability in tail events. The rate of error for the measure becomes so explosive that in that arena individual parameters seem to be more relevant, among many other things. We'd be better off with PG's intuition than this.