The analysis shown only considers ‘expected gain’. How about ‘expected losses’. How would that change the scenario?

Kind of reminds me of Kahneman & Tversky.

]]>However, by following the link supplied by Jung-Chin, you’ll find Amos Storkeys solution – which does make sense to me. But I can’t say for sure whether it is flawless… ]]>

A side observation is that your preference will depend on your utility function for money. If utility goes as the log of the amount of money, you will be indifferent between a doubling and halving of N. For most people, the amount in the envelopes will be a marginal contribution to their total wealth, and so the utility function will be more linear. Therefore, they should happily switch envelopes when promised a doubling or halving with equal probability.

Also, of course, you have to trust the host. If he would be more likely to offer you the swap when you would lose money, then it becomes a bad idea to swap.

]]>If you calculate the expected gains and losses in percentage terms, using an arc elasticity formula, does the paradox disappear?

]]>