International Joint Conference On Theoretical Computer Science – Frontier of Algorithmic Wisdom

August 15-19, 2022, City University of Hong Kong, Hong Kong

 

Invited Speakers

Track B


EIP-1559: Chaos and Efficiency in Ethereum's Transaction Fee Market

Stefanos Leonardos

King's College London

Abstract:
Ethereum Improvement Proposal (EIP) 1559 is a recently implemented upgrade that transformed Ethereum's transaction fee market. EIP-1559 uses an algorithmic update rule with a constant learning rate to estimate and broadcast a base fee that users need to pay to have their transactions included in the blockchain. The base fee reflects prevailing network conditions and aims to efficiently stabilize block occupancies at a target block size.
In this talk, we will try to understand the dynamics of the resulting transaction fee market using a diverse toolbox. We will stress test the system and study its behavior under a wide range of adversarial conditions that can emerge in practice. We will then compare our theoretical findings with empirical data and proceed to analyze alternative update rules that overcome undesirable properties of the current model. In the most important and surprising takeaway of our talk, we will show that these systems can be provable efficient despite being formally chaotic.

Bio:
Stefanos Leonardos recently joined King's College London as a Lecturer in Machine Learning at the Department of Informatics. Before coming to London, Stefanos spent 4 exciting years as a Research Fellow at Singapore University of Technology and Design working in game theory, game dynamics, multi-agent systems and their applications in artificial intelligence, machine learning and blockchain economies. Stefanos holds a PhD and a MSc in Game Theory and Operations Research from University of Athens and a Diploma in Mathematics from Technical University Darmstadt.