This code is for the ACL 2022 paper ChatMatch: Evaluating Chatbots by Autonomous Chat Tournaments. First, we make each bot talk to each other using ParlAI (you can see the implementation details in https://github.com/facebookresearch/ParlAI). The python file here displays how we evaluate two bots' behavior in one game. Later we use TrueSkill Algorithm(https://trueskill.org/) to rank all the bots. (We will add more details later)