@aryehazan @Sam_Schulman @ColinPClarke So this event is singular, and there's no way to learn from the past? I agree that reference classes are critical, and debateable, but arguing against a given reference class seems much less reasonable than proposing
“The reference class problem is your problem too.” https://t.co/VraIrYGTGM
@randomprocs @ETagliazucchi I would argue that the subjectivity applies equally to frequentist notions. Check out this excellent paper: https://t.co/Sx9K2ZsY6r
@tyrell_turing @danijarh Well the subjectivity applies to frequentists too, in a way. It's called the Reference Class Problem. I highly recommend this paper: https://t.co/Sx9K2ZbneT
This is why I think Alan Hájek's take on probability is the right one for scientists. Conditional probability should be the primary concept, not absolute probability. Conditional probability keeps us honest about how and what we are counting. https://t.co
@SimonDeDeo Haha that is exactly my position! Not a betting man if I can help it. Btw have you read this paper? Quite interesting. https://t.co/Sx9K2ZbneT
@ole_b_peters I am not sure if this is mainly an ergodicity problem or mainly a reference class problem ( https://t.co/mfeQXQGMZu ), but there is not a lot to be said for a reference class that doesn't even really exist.
RT @jvrbntz: The reference class problem is your problem too https://t.co/ZcHdqwW0hd https://t.co/lqJlbWFG8J
RT @jvrbntz: The reference class problem is your problem too https://t.co/iAZw0PmIW9 #InductiveReasoning #Statistics https://t.co/shFFjHXQ…
The reference class problem is your problem too https://t.co/iAZw0PmIW9 #InductiveReasoning #Statistics https://t.co/shFFjHXQIQ
The reference class problem is your problem too https://t.co/ZcHdqwW0hd https://t.co/lqJlbWFG8J