Artificially Intelligent Marketplaces

31 Pages Posted: 25 May 2022

See all articles by Ruiqi Lin

Ruiqi Lin

affiliation not provided to SSRN

Pavel Kireyev


Date Written: May 25, 2022


Peer-to-peer marketplaces have exploded in popularity in the past decade. Companies and individuals have deployed bots into these marketplaces to generate profits by purchasing and reselling items. However, the profitability of such bots remains uninvestigated. We develop a framework for back-testing trading bots in peer-to-peer marketplaces. The framework infers outcomes for counterfactual actions taken by the bot and uses data on what actually happened in the marketplace to assess the bot’s profitability out-of-sample. Hyperparameters allow developers to fine-tune bot strategies and high-dimensional variable selection models help isolate the effects of bot actions from a large number of confounders. We apply the framework to data from the CryptoPunks NFT marketplace which uses a combination of bidding and buy-it-now prices. We show that a strategy that “spams” low bids to a large number of listings can be profitable with a return-on-investment of 10.24% over 4 months. Bots that buy at buy-it-now prices struggle to generate significant profits. Developers can also deploy seller-rewarding bots that pay for themselves and stimulate the market by improving seller revenues. Our framework helps bot developers test different strategies and points to the feasibility and limitations of artificially intelligent peer-to-peer marketplaces in which bots participate together with humans.

Keywords: Marketplaces; Platforms; Artificial Intelligence; Bots; Back-Testing; Pricing; Nonfungible Tokens; Machine Learning

Suggested Citation

Lin, Ruiqi and Kireyev, Pavel, Artificially Intelligent Marketplaces (May 25, 2022). INSEAD Working Paper No. 2022/26/MKT, Available at SSRN: or

Ruiqi Lin

affiliation not provided to SSRN

Pavel Kireyev (Contact Author)

INSEAD ( email )

Boulevard de Constance
77305 Fontainebleau Cedex


Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics