I conducted a three-person ultimatum experiment in which the stake is divided into two pots of different sizes. The responders receive either the same amount of money or the same share in their pots, and have full information about the distribution of the stake. I find that subjects care about comparisons between the responders and these comparisons significantly affect their emotions. The responders who are treated more poorly than their counterparts report high intensity of anger and envy, and by punishing the proposer with rejections cause a significant loss of effectiveness in the distribution of income. On the other hand, the responders who are treated better than their counterparts report high intensity of guilt, which suggests strong inequality aversion.

For more than half a century experimental results accumulate that disprove prognostic value of perfect equilibrium, a concept derived from the principle of rationalistic egoism. The article is a short survey of popular experimental games in which actors make decisions of monetary consequences, thus shaping the income distribution in a group (dyad, in particular). A brief account is given of experimental research on ultimatum and dictator game, dictator's dilemma, trust game, centipede, gift exchange and public good game. Special focus is on dictator's dilemma which tests for readiness to benefit one's partner at one's own expense. This elementary personal dilemma turns out to constitute an essential part of more complicated and better known two- and multi-personal games.