## My touchstone for philosophies of probability

A running theme on this site has been
prediction tournaments.
The conceptual insight is that, in a prediction tournament, one can estimate different
people's *relative* abilities
to assess probabilities of unique future real-world events, even though it is
impossible to know the true probabilities.
This is based on what philosophers might call the "naive" philosophy that such an event has an unknown
"true probabilities" based on given available information.
To me, this example provides a touchstone to reject extreme philosophies of probability.
An extreme Bayesian might deny that "true probabilities" ever exist, and an extreme frequentist
might say they only make sense for repeatable events.
But if you deny that "true probabilities" exist for unique events, then the empirical observation
from prediction tournaments that some
people are consistently better at prediction than others is hard to incorporate into your philosophy.