The Dynamics of a "Selfish Mining" Infested Bitcoin Network: How the Presence of Adversaries Can Alter the Profitability Framework of Bitcoin Mining

Abstract

Bitcoin mining is a process that serves to both verify sets of transactions and slowly introduce new currency into the system. As a reward for performing this process, miners are paid in bitcoin for the blocks they mine. It was originally thought that there was no incentive in trying to subvert the mining protocol—in other words, there was no reason to believe that miners could be profitable by somehow cheating the system. As it turns out, a specific strategy called “selfish mining” was discovered to increase profitability for miners under certain conditions. This paper presents the selfish mining strategy, traverses a revenue model associated with the strategy, and then simulates the bitcoin network to see how this revenue model holds up under complicated network conditions. Specifically, the selfish mining revenue model typically assumes there is one selfish miner in the network—I simulate the more realistic case of there being many selfish miners in the network. We find that the revenue model can overestimate selfish miner revenues by up to 100% and underestimate them by up to 300% depending on network variables such as the number of selfish miners, the power of those miners, and network latency (the speed of block propagation from one miner to another).