|Author||Yoad Lewenberg, Yoram Bachrach, Yonatan Sompolinsky, Aviv Zohar, Jeffrey S. Rosenschein|
|Publisher||Proceedings of the 14th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2015), Bordini, Elkind, Weiss, Yolum (eds.), May 4–8, 2015, Istanbul, Turkey|
|Link||View Research Paper|
Bitcoin is an innovative decentralized cryptocurrency whose core security relies on a “proof of work” procedure, which requires network participants to repeatedly compute hashes on inputs from a large search space. Finding one of the rare inputs that generates an extremely low hash value is considered a successful attempt, allowing miners to approve new transactions and, in return, to collect rewards in bitcoins. This reward allocation, which provides the incentive for miners to participate, is a random process with a large variance. Miners who desire a steady income thus often participate in mining pools that divide among their members the earned rewards, and reduce this variance. Mining pools are slightly better at coordinating participants due to lowerlatency communication, a fact which implies that they manage to collect slightly higher rewards. We examine dynamics of pooled mining and the rewards that pools manage to collect, and use cooperative game theoretic tools to analyze how pool members may share these rewards. We show that for some network parameters, especially under high transaction loads, it is difficult or even impossible to distribute rewards in a stable way: some participants are always incentivized to switch between pools.