Abstract
We provide a dataset for our research article "Profitability, Efficiency and Inequality in Double Auction Markets with Snipers" [1]. This dataset [2] includes configuration files, raw output data, and replications of calculated metrics for our robot-populated market simulations. The raw data is subdivided into a hierarchy of folders corresponding to simulation treatment variables, in a 2 × 2 × 21 design for 84 treatments in total. Treatments variables include: (i) robot population ordering, either "primary" or "reverse"; (ii) two market schedules of agent's values and costs: equal-expected-profit "market 1" and unequal-expected-profit "market 2"; (iii) 21 robot populations identified by the number of Sniper Bots (0-20) on each side of the market. Each treatment directory contains a simulator input file and outputs for 10,000 periods of market data. The outputs include all acceptable buy and sell orders, all trades, profits for each agent, and market metrics such as efficiency-of-allocation, Gini coefficient, and price statistics. An additional public copy in Google Cloud is available for database query by users of Google BigQuery. The market simulator software is a private product created by Paul Brewer at Economic and Financial Technology Consulting LLC. Free open source modules are available for tech-savvy users at GitHub, NPM, and Docker Hub repositories and are sufficient to repeat the simulations. An easier-to-use paid market simulation product will eventually be available online from Econ1.Net. We provide instructions for repeating individual simulations using the free open source simulator and the free container tool Docker.