Trust in Bitcoin Exchange Networks

We consider two marketplaces to exchange bitcoins and dollars, Bitcoin OTC and Bitcoin Alpha. In these settings, trust is a fundamental necessity since any user can initiate a transaction, receive money from the other user and never send money back in the other currency. To achieve a web of trust, users rate each other after successful or unsuccessful trades. We analyze these datasets using the fairness and goodness method from [5]. These two measures can be used to predict the weight of a given edge (the rating that u gives to v) with good accuracy. We show that whenever the edge is reciprocated, social interaction becomes a better predictor than goodness, and users apply the talion law: "an eye for an eye". We analyze the relationship between trust in these networks with the exchange price of bitcoin and find that there is no evident correlation between these two variables.