Tyrrell_McAllister comments on Buy Insurance -- Bet Against Yourself - Less Wrong

29 Post author: MBlume 26 November 2010 04:48AM

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

Comments (68)

You are viewing a single comment's thread. Show more comments above.

Comment author: Tyrrell_McAllister 28 November 2010 10:40:46PM *  1 point [-]

Decisions should be made not on absolute badness or goodness of any given outcome, but on comparative goodness of alternative outcomes following from alternative possible decisions.

On my reading, the comparisons are there, but they are implicit. Here are what I took to be the implicit assumptions. For every state S of affairs under consideration, if the money is worthless in S, then S is worse than the all-else-equal state S' where the money is not worthless. Similarly, in any state T where kratom is unregulated, that state is better than the all-else-equal state T' where kratom is regulated. (In particular, states are being considered at a low-enough resolution that you can just flip the "the money is worthless" or "kratom is unregulated" bits and get another state. Also, while some fanciful states might be made worse if kratom were unregulated and all else were equal, those states are not being considered because their probabilities are too low.)

So, according to the above implicit comparisons, losing money always makes things worse. It is therefore valid to say that losing money is bad. Similarly, it is valid to say that keeping kratom unregulated is good.

Given this, it is valid to make statements of the form, "If X happens, then A will happen (which is bad in the above sense), but B will happen (which is good in the above sense)."

Comment author: Vladimir_Nesov 29 November 2010 04:44:09PM *  0 points [-]

You are considering the wrong alternatives, things that can possibly happen independently on your actions, while the alternatives that should be considered result from your taking different actions. The alternatives under the factors you don't control are just internal structure of your model of uncertainty inside a particular possible decision.

Comment author: Tyrrell_McAllister 29 November 2010 06:09:54PM *  1 point [-]

I'm not sure what you mean by "thing that can possibly happen independently on your actions". But "the money becomes worthless" is not independent of your action — the money's becoming worthless requires that you choose to bet.