The Boston mechanism (BOS) is widely used for the assignment of students to schools. Yet, BOS is highly manipulable and, therefore, may lead to Pareto inferior assignments. We propose a new indirect matching mechanism (\({{ NBOS }}\)) that is a slight variant of BOS. \({{ NBOS }}\) is less manipulable than BOS in two important ways: (i) students always have a best reply featuring truthful reported preference, (ii) rational students avoid particular re-rankings that are typical in BOS. Most importantly, the lower manipulability of \({{ NBOS }}\) helps reaching more efficient assignments than those reached by BOS. We show that each equilibrium of BOS is associated to a set of equilibria of \({{ NBOS }}\), all of which are Pareto superior. \({{ NBOS }}\) generalizes to a class of mechanisms whose members are even less manipulable than \({{ NBOS }}\).

波士顿机制（BOS）被广泛用于分配学生到学校。但是，BOS的操作性极强，因此可能导致帕累托的下等任务。我们提出了一种新的间接匹配机制（\（{{NBOS}} \）），它是BOS的一个细微变化。\（{{NBOS}} \）在两个重要方面比BOS难以操纵：（i）学生始终具有以真实的报道偏爱为特征的最佳答案，（ii）有理性的学生避免使用BOS中通常具有的特殊排名。最重要的是，\（{{NBOS}} \）的较低可操作性有助于实现比BOS更高的分配效率。我们证明BOS的每个平衡都与\（{{NBOS}} \）的一组平衡相关，所有这些平衡都是Pareto优越的。\（{{NBOS}} \）概括为一类机制，其成员比\（{{NBOS}} \）的可操纵性更低。