Pricing vs Auctions
A large part of economic theory as well as algorithmic game theory focuses on designing an optimal auction, either with revenue or welfare objective. However, designing such auctions can be computationally hard in some cases. Sometimes a greater hindrance is the difficulty of implementing an optimal auction, especially for consumer goods. Buyers usually prefer posted pricing (when the seller offers a take-it-or-leave-it price) to auctions, and cannot be expected to participate in auctions to buy their daily consumptions! So pricing strategies are far more commonly used in practice compared to optimal auction designs. Unless some auction design promises far more revenue than any pricing strategy, a seller may prefer to use a pricing strategy. So it is important to understand the gap between the performance of auctions and pricing. In this talk, I will show that efficiently computable pricing strategies can perform almost as well as optimal auctions when there are many items to sell. Further, in some complex combinatorial auctions, simple prior-free pricing strategies provide the best revenue guarantees given by any known efficiently computable auction.