Unstable Throughput: When the Difficulty Algorithm Breaks

Sam M. Werner, Dragos I. Ilie, Iain Stewart, William J. Knottenbelt

Published 2020-06-04Version 1

Difficulty algorithms are a fundamental component of Proof-of-Work blockchains, aimed at maintaining stable block production times by dynamically adjusting the network difficulty in response to the miners' constantly changing computational power. Targeting stable block times is critical, as this ensures consistent transaction throughput. Some blockchains need difficulty algorithms that react quickly to severe hash rate fluctuations. However, without careful design this could create vulnerabilities that incentivize miners to engage in coin-hopping strategies which yield an unreliable system due to unstable processing of transactions. We provide an empirical analysis of how Bitcoin Cash exhibits cyclicality in block solve times as a consequence of a positive feedback loop in its difficulty algorithm design. Additionally, we examine the extent to which miners' behavior contributes towards this phenomenon over time. In response, we mathematically derive a difficulty algorithm based on a negative exponential filter that prohibits the formation of positive feedback loops and exhibits additional desirable properties, such as history agnosticism. We compare the described algorithm to that of Bitcoin Cash in a simulated mining environment and verify that the former would eliminate the severe oscillations in block solve times. Lastly, we outline how this model can more generally replace difficulty algorithms in other Proof-of-Work blockchains.