Research & Papers

New auction theory unlocks 1-(1/e) welfare efficiency

Game theorists crack private position auctions with 2x better welfare guarantees

Deep Dive

Researchers Ruijie Wang (MIT) and Aviv Yaish (Technion) have published a seminal paper on *order auctions*—where two positions are auctioned to unit-demand bidders with heterogeneous preferences. Their model distinguishes between *specialists* (valuing only the top position) and *generalists* (indifferent between positions), revealing fundamental limits in standard first-price auctions.

The team proves that efficient allocation is unattainable without additional bidder communication. In deterministic one-round auctions, at least one *extra bit* beyond standard bids is required for efficiency. By allowing bidders to specify position preferences in winner-pays-bid formats, their design achieves a welfare guarantee of **1-(1/e) ≈ 63%**—a 26% improvement over the 50% baseline. The work has direct applications to blockchain transaction sequencing and priority service auctions, where sequencing decisions critically impact welfare.

Key Points
  • Model introduces 'specialists' (top-position-only valuers) and 'generalists' (position-indifferent bidders) in two-position auctions
  • Standard first-price auctions fail to achieve efficient allocation without extra bidder communication (proven via 1-bit lower bound)
  • Winner-pays-bid auctions with preference specification achieve 1-(1/e) ≈ 63% welfare guarantee vs. 50% baseline

Why It Matters

Revolutionizes blockchain sequencing and priority services by quantifying welfare gains from bidder preference revelation.

📬 Get the top 10 AI stories daily