Comment by stellalo

I guess you could attach the probability as weight to the event outcome, instead of taking a uniform mean. Not that without probabilities it’s “impossible”, it’s just a different metric