Are Two (Samples) Really Better Than One? On the Non-Asymptotic Performance of Empirical Revenue Maximization

The literature on "mechanism design from samples," which has flourished in recent years at the interface of economics and computer science, offers a bridge between the classic computer-science approach of worst-case analysis (corresponding to "no samples") and the classic economic approach of average-case analysis for a given Bayesian prior (conceptually corresponding to the number of samples tending to infinity). Nonetheless, the two directions studied so far are two extreme and almost diametrically opposed directions: that of asymptotic results where the number of samples grows large, and that where only a single sample is available. In this paper, we take a first step toward understanding the middle ground that bridges these two approaches: that of a fixed number of samples greater than one. In a variety of contexts, we ask what is possibly the most fundamental question in this direction: "are two samples really better than one sample?". We present a few surprising negative results, and complement them with our main result: showing that the worst-case, over all regular distributions, expected-revenue guarantee of the Empirical Revenue Maximization algorithm given two samples is greater than that of this algorithm given one sample. The proof is technically challenging, and provides the first result that shows that some deterministic mechanism constructed using two samples can guarantee more than one half of the optimal revenue.

[1]  Nicole Immorlica,et al.  Randomization Beats Second Price as a Prior-Independent Auction , 2015, EC.

[2]  Yoav Shoham,et al.  Truth revelation in approximately efficient combinatorial auctions , 2002, EC '99.

[3]  Richard Cole,et al.  The sample complexity of revenue maximization , 2014, STOC.

[4]  Jason D. Hartline.  Mechanism Design and Approximation , 2014 .

[5]  Yannai A. Gonczarowski,et al.  Efficient empirical revenue maximization in single-parameter auction environments , 2016, STOC.

[6]  Tim Roughgarden,et al.  On the Pseudo-Dimension of Nearly Optimal Auctions , 2015, NIPS.

[7]  Andrew V. Goldberg,et al.  Competitive auctions and digital goods , 2001, SODA '01.

[8]  Maria-Florina Balcan,et al.  A General Theory of Sample Complexity for Multi-Item Profit Maximization , 2017, EC.

[9]  Tim Roughgarden,et al.  Simple versus optimal mechanisms , 2009, SECO.

[10]  Nikhil R. Devanur,et al.  The sample complexity of auctions with side information , 2015, STOC.

[11]  Vianney Perchet,et al.  Online learning in repeated auctions , 2015, COLT.

[12]  Maria-Florina Balcan,et al.  Sample Complexity of Automated Mechanism Design , 2016, NIPS.

[13]  Silvio Micali,et al.  Parametric digital auctions , 2013, ITCS '13.

[14]  Tim Roughgarden,et al.  Learning Simple Auctions , 2016, COLT.

[15]  Noga Alon,et al.  Submultiplicative Glivenko-Cantelli and Uniform Convergence of Revenues , 2017, NIPS.

[16]  Jeremy I. Bulow,et al.  Auctions versus Negotiations , 1996 .

[17]  S. Matthew Weinberg,et al.  Optimal and Efficient Parametric Auctions , 2013, SODA.

[18]  Claudio Gentile,et al.  Regret Minimization for Reserve Prices in Second-Price Auctions , 2013, IEEE Transactions on Information Theory.

[19]  Tim Roughgarden,et al.  Revenue maximization with a single sample , 2015, Games Econ. Behav..

[20]  Tim Roughgarden,et al.  Making the Most of Your Samples , 2018, SIAM J. Comput..

[21]  Roger B. Myerson,et al.  Optimal Auction Design , 1981, Math. Oper. Res..