I think this misses out on the fact that utility is always indirect—there is a function from world-state to utility that each rational agent has, so there can never be a lottery that directly awards utility. Meaning you can model the utility valuation linearly, but the mapping of resources to utility with logarithmic declining marginal utility.
I think this misses out on the fact that utility is always indirect—there is a function from world-state to utility that each rational agent has, so there can never be a lottery that directly awards utility. Meaning you can model the utility valuation linearly, but the mapping of resources to utility with logarithmic declining marginal utility.