标题： 最后通牒游戏：遗憾还是公平？ 显示英文标题

标题： Ultimatum game: regret or fairness?

摘要： 在最后通牒游戏中，挑战在于解释为什么接受者会拒绝非零报价，从而违背了经典理性。 公平及相关概念迄今为止一直是主要的解释。 我们通过以下原则来解释这种拒绝行为：如果接受者对失去报价的后悔小于提议者对没有提供最佳选项的后悔，那么该报价将被拒绝。 这一原则被归类为理性的惩罚行为，并取代了实验上被证伪的经典理性（子博弈完美纳什均衡），该理性导致接受任何非零报价。 该原则通过概率彩票的传递性后悔理论来实现。 期望效用的实现是这一理论的一个极限情况。 我们证明，通常归因于公平和意图识别的几个实验结果可以通过理性惩罚给出另一种解释；例如。 “公平”与“超公平”之间的比较，以及提高赌注下的行为等。 因此，我们还提出了可以区分这两种情况（公平与基于后悔的惩罚）的实验。 它们假设提议者和接受者的效用不同。 我们专注于游戏的迷你最后通牒版本，并且还展示了它如何从游戏的更一般设置中产生。 显示英文摘要

摘要： In the ultimatum game, the challenge is to explain why responders reject non-zero offers thereby defying classical rationality. Fairness and related notions have been the main explanations so far. We explain this rejection behavior via the following principle: if the responder regrets less about losing the offer than the proposer regrets not offering the best option, the offer is rejected. This principle qualifies as a rational punishing behavior and it replaces the experimentally falsified classical rationality (the subgame perfect Nash equilibrium) that leads to accepting any non-zero offer. The principle is implemented via the transitive regret theory for probabilistic lotteries. The expected utility implementation is a limiting case of this. We show that several experimental results normally prescribed to fairness and intent-recognition can be given an alternative explanation via rational punishment; e.g. the comparison between "fair" and "superfair", the behavior under raising the stakes etc. Hence we also propose experiments that can distinguish these two scenarios (fairness versus regret-based punishment). They assume different utilities for the proposer and responder. We focus on the mini-ultimatum version of the game and also show how it can emerge from a more general setup of the game.