无分布均衡的寻觅竞赛
Search Contests
AI总结:
本文研究了在固定成本下玩家连续多次抽取分布以获取最高值的寻觅竞赛,发现均衡特性与分布无关,总搜索支出等于奖金。研究还探讨了多奖和设计者层级竞争,并指出有效奖金需根据分布调整,重尾分布需更大奖金。
摘要:
我们研究了一种竞赛,其中N名玩家可以按需从分布中抽取多次,每次抽取固定成本且不可回溯,最高接受值获得奖金。在唯一对称均衡中,接受概率、预期搜索成本和玩家支付与底层分布无关。总搜索支出等于奖金(完全租金耗散)。这些无分布均衡性质扩展到多个奖金和设计者层级竞争。有效的奖金需根据分布调整:重尾分布需要更大的奖金。在有限抽取次数下,增加竞争者可提高质量门槛,当搜索成本较低时,这会逆转无限抽取情况的不利影响。选择奖金和领域规模的规划者总是偏好最小领域(N=2)无限抽取,但重尾分布和有限抽取次数偏好更大领域,因并行探索的广度补偿了个体搜索的有限深度。
英文摘要:
We study contests in which players search sequentially, drawing from a distribution at a cost per draw, and the highest accepted value wins the prize. We analyze unlimited and bounded search, with and without recall. Equilibria are distribution-free, depending only on the number of players, search depth, and the cost-prize ratio. With multiple prizes, equilibrium depends only on the average and the last-place prize. Entry has two effects, discouragement, which lowers the acceptance threshold, and selectivity, which raises it. With unlimited search, competition dissipates all rents, leaving only discouragement. With bounded search, rents survive, selectivity dominates at low cost, and the threshold is single-peaked in the number of players. In large fields discouragement prevails. A planner's optimal prize is invariant to location shifts and rises with tail thickness. Two players are optimal with unlimited search, while heavy tails favor larger fields when search is bounded.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。