I don't think economists really have such a narrow idea of utility. Usually it is applied in abstract models, where utility is predefined (like a lottery where

I don't think economists really have such a narrow idea of utility. Usually it is applied in abstract models, where utility is predefined (like a lottery where you can win money, and the amount you win is used as the utility - no need to consider how happy it might make the theoretical player). For the real world, I think evolutionary aspects have to be taken into account. A single actor may not act rational, but in the long term, the actors making the rational choice will prevail by the power of evolution. So rational choices will be made in the long run, even if it may not even be deliberation of the actors. Also, don't forget that happiness research itself may also be wrong. I think the classic about people without children supposedly being less happy than people with children points in that direction. By the standard research methodology, watching Netflix on the couch all the time would probably make people the most happy.