過去的連續性搜尋擇偶模型常只是假設在一個特殊的「一對多」 （一個擇偶者多位異性連續出場）的情況下進行，例如「古典的嫁妝最大化問題」（Mosteller, 1965; Ferguson,1989）。本文加入「婚姻賽局」模擬配對的概念，建構一個「多人對多人互選配對」的連續性搜尋模擬婚姻市場，根據Gale and Shapley（1962）的極大化配對規則，在當期中進行多對多的互選配對，並在跨期中進行連續性的配偶搜尋。本文在此婚姻市場下嘗試兩種常見的連續性搜尋模式的模擬。第一種為「騎驢找馬」，每位擇偶者均暫時先與眼前的一位異性交往，如果在下一期遇到更好的對象，就淘汰現在交往的對象。第二種為「劈腿」，每位擇偶者最多可暫時與兩位異性交往，但劈腿時有被抓到而失去所有交往異性的一個風險機率存在。本文以電腦模擬兩模式各100次，得出擇偶者的平均配對成功機率。研究結果顯示：雖然對個人而言，當自己採取劈腿策略而他人採取騎驢找馬策略時，劈腿策略是較優勢的擇偶策略選項。但是，當全社會的每一個擇偶者都採取劈腿策略（劈腿模式）時，擇偶者配對成功率卻較在騎驢找馬模式之下低，出現了劈腿的「合成謬誤」。本文引用賽局理論中的「囚犯困境」討論這種擇偶現象的可能性。

There have been two approaches to simulation mate-selection research. One has focused on one agent’s continuous searching in multiple periods (e.g., Mosteller, 1965; Ferguson, 1989). The other has focused on multi-agent matching in a marriage market in the same period (e.g., Gale and Shapley, 1962). The current study combines these two approaches in a new model. Using the new multi-agent continuous-searching simulation match model, this study simulates the following two mating strategies. A “one-mate dating strategy” (OMS) means dating with one mate while searching for another, better one. An “extra-dyadic dating strategy” (EDS) means secretly dating with two mates at the same time while searching for a better one. In a tenperiod limited searching situation, after 100 simulation iterations in the new model, the results show that the success rate of mating in an “OMS society” (with everyone applying OMS) is higher than in an “EDS society”. This phenomenon is called “a fallacy of composition in mate-selection”. EDS is a more unfaithful strategy than OMS, and EDS looks like a better strategy if others are applying OMS. However, in a marriage market when every individual applies EDS in mating, the results become worse. We suggest that this fallacy can be explained by the prisoner’s dilemma of game theory.