An empirical analysis of chain reorganizations and double-spend attacks on proof-of-work cryptocurrencies

Abstract

Nakamoto consensus has powered Bitcoin and the cryptocurrency industry over the past 10 years, but its security properties when an adversary's economic incentives are taken into account remain poorly understood. Recently, reports of successful real-world attacks against some coins have served as a wake-up call for the industry to review each coins' consensus risk. This research contributes a new system for detecting transaction reordering events against live cryptocurrencies. We deployed the system on a spectrum of different cryptocurrencies and combined our results with historical market data to analyze how the properties of each coin affect its consensus risk and evaluate the effectiveness of existing theoretical models for quantifying the cost of attack. We also describe some of the significant attacks we detected, providing empirical evidence that launching an attack can be practical, and that counterattacking may be a viable strategy for victims to defend themselves from an economically rational adversary.