Our project focuses on developing software to find and teach a game theoretically optimal solution (or as close as is feasible) for a card game called “Pineapple Open Face Chinese Poker” (or Pineapple OFC for short). It is a 2-player zero sum game where each player’s goal is to make the strongest possible poker hand on each of the 3 rows such that each row’s hand is better than the ones above it. Solutions to this game using our algorithm has implications in other games of similar structure -- incomplete information, 2-player, zero sum -- as well as different fields such as the stock market.
The framework we are using to solve Pineapple OFC utilizes probability, game theory, and simulations to analyze the game tree. Our program also allows users to input situations and view the computer’s analysis of the best move available. The game is far too complex to solve through brute-force computation alone, so we analyze it in pieces. With this paradigm, we can make the game more manageable in size, and converge to an equilibrium strategy by analyzing ex-post every decision and its alternatives. In this way, we can develop a strategy superior to what human players can currently achieve.