Reproduce “Efficiency of non-truthful auctions under auto-bidding” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Christopher Liaw, Aranyak Mehta, Andres Perlroth
This paper introduces and analyzes the randomized First Price Auction (rFPA) for auto-bidding environments, demonstrating that it achieves a Price of Anarchy (PoA) of 1.8 for two bidders, improving upon the PoA of 2 found in deterministic and truthful auctions. It also establishes an asymptotic PoA lower bound of 2 for a large number of bidders.
Key insights
- Any deterministic auction satisfying reasonable assumptions has a Price of Anarchy (PoA) of at least 2, even with only two bidders.
- Introducing randomization in allocation, particularly in a non-truthful First Price Auction (rFPA), can encourage more aggressive bidding and lead to efficiency improvements.
- The benefit of non-truthfulness is demonstrated by comparing rFPA (PoA 1.8) to its truthful randomized counterpart (PoA 1.98) for two bidders.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.