Manfred comments on Against improper priors - Less Wrong

2 Post author: DanielLC 26 July 2011 11:50PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (20)
Tags: bayesian prediction probability

Comments (20)

You are viewing a single comment's thread.

Comment author: Manfred 27 July 2011 01:41:32AM 3 points [-]

Dutch books also seems to be a popular way of showing what works and what doesn't, so here's a simple Dutch argument against improper priors: I have two real numbers: x and y. Suppose they have a uniform distribution. I offer you a bet at 2:1 odds that x has a higher magnitude. They're equally likely to be higher, so you take it. I then show you the value of x. I offer you a new bet at 2:1 odds that y has a higher magnitude. You know y almost definitely has a higher magnitude than that, so you take it again. No matter what happens, I win.

To fix this example, replace "real" with "positive real" and make the bets 2:4 and 100:1.

Still, an example that comes from using improper priors as probability distributions, which they are explicitly not, doesn't seem like a strong argument. Better to show that they can't come up in any interesting situations - this may be impossible, though.

Comment author: DanielLC 27 July 2011 03:33:39AM *  0 points [-]

To fix this example, replace "real" with "positive real"

I used "real" because with positive reals, you're more likely to use a logarithmic prior.

and make the bets 2:4 and 100:1.

Oops. Why did you say 2:4 instead of 1:2? Do you mean 2:1?

using improper priors as probability distributions, which they are explicitly not,

If they're not probability distributions, what are they?

Comment author: Manfred 27 July 2011 04:06:23AM 4 points [-]

Why did you say 2:4 instead of 1:2? Do you mean 2:1?

Just to emphasize that the victim should have more money riding on the first bet if they are to consistently lose money.

If they're not probability distributions, what are they?

Since they're invalid probability distributions but can be updated into a probability distribution given some evidence, you might think of these as representing states where you have some knowledge, but not enough to assign consistent probabilities. For example, if all you know is that X is a member of some infinite set, you cannot assign consistent probabilities, but you still have some knowledge, which might be represented as a uniform function.

Powered by Reddit Powered by Reddit