WinC 2023 Seniors
On May 7th, 2023, Casey Ford ’23, presented their thesis on Hyper Opponent Modeling. Below is Casey’s Thesis Abstract:
“In a multiplayer competitive game, each player develops a strategy to win. The game is stochastic (the initial conditions are randomly determined) and partially observable (you cannot see the initial conditions or the actions your opponents take), and involves players choosing and amassing game pieces which are hidden from the other players. It is advantageous to pick game pieces that work well together, and there are many different strategies for doing so. Once the game is done, there is a final contest with the game pieces collected by each individual player to determine the winner. As a player, it is useful to know what strategies your opponents executed during the first phase of the game in order to win in the final contest. Therefore, you may want a scheme to analyze the perceptions and signals you received throughout the game and use that to reason about what strategies your opponents might have taken. In such a competitive, multi-agent, stochastic, imperfect information, partially observable environment, (based off of a sub game called “Draft” of the popular trading card game Magic: The Gathering) how can we model the strategies of our opponents?”, (Ford, 2023).