With the increasing popularity of e-commerce systems, commercial transactions become more and more frequent. Such transactions are not direct but mediated by the supporting online platforms, that is the payment and delivery of the good (or the use of the service) are not at the same time. In the current practice, the seller delivers the good only after receiving the proof of payment from the buyer. In this context, the buyer is in a position of weakness with respect to the seller, especially in the case of a failure of the transaction. Reputation is an aggregate composite of all previous transactions over the life of the entity, a historical notion, and requires consistency of the entity’s actions over a prolonged time [1]. Reputation includes not only the direct experiences of the buyer but also any other form of communication – reviews, scores – that provides information about the seller [2]. Reputation can play an important role to reduce the risks of the buyer in the current e-commerce environment. In [3], the authors showed that positive online review scores can positively influence the firm financial performance while the heterogeneity of different product classes moderates the relationship between review score and performance. Furthermore, in [4], the authors reported that a limited number of fake reviews can determine a consistent reduction of the reputation of a competitor. In order to limit the impact of malicious behaviors, online reputation management systems (RMS) have been developed over the years. RMS is a system that maintains the beliefs or the opinions that are generally held about someone or something. Such a RMS can provide a solution to guarantee the reliability of the transactions that take place in an e-commerce system [5–۷]. Several RMS are proposed in the literature: those systems are based on different methodologies, such as artificial intelligence, multi-agent systems, cognitive science, game theory, and the social and organi zational sciences [8]. In computer science, particular attention has been dedicated to the analysis of the RMSs operating on a peer-topeer systems [9–۱۳]. The interaction between the seller and the buyer, which is the basic element of a RMS, is a classical field of application of the Game Theory (GT) methodologies, which allow modeling the rational behavior of the individuals [14]. On the contrary, the use of a GT approach in this context seems quite limited despite of its potential (see, e.g., [15–۱۷]). This is probably due to the resulting solution complexity of the GT approach. Such a complexity does not depend on the complexity of each single transaction: actually, the strategic interaction model of a single transaction between seller and buyer is extremely simple as the buyer has to decide whether to buy or not, while the alternative of the seller are to fully comply with the request or not. On the contrary, the complexity relies on the fact that the reputation is the result of (i) A number of repeated transactions between pairs of sellers and buyers, not necessarily the same, and (ii) the sharing with other sellers and buyers of the outcomes of the transactions. Note that the sharing of the outcomes of the transactions represents the learning effect that is typical of repeated games.