Re: Bag of Tricks: Choosing actions via the matching law

[EMAIL PROTECTED] (Bryan) writes:

> So would I, but if I came to the game with no prior knowledge of the
> efficacy of the weapons, I would have to experiment - that is, use
> "trial-and-error" to build up a probability distribution estimate
> for the efficacy of each. There needs to be a balance between always
> choosing the weapon that appears to be the best, and sometimes using
> a weapon that appears not to be the best, just so that the second
> weapon can be accurately evaluated. It would appear that the formula
> strikes this balance between investing in learning and investing in
> using.

I don't think so, because it doesn't take into account the amount of
experience. It doesn not appear to be sensible to try X with the same
probability if it worked in 1 of 3 cases, or in 100000 of 300000 cases
(if there's also something that worked in 2/3 or 200000/300000 cases,
respectively).

-- 
        Falk