Traditional Wisdom and Monte Carlo Tree Search Face-to-Face in the Card Game Scopone

TL;DR

This paper develops and compares rule-based, MCTS, and ISMCTS AI players for the Italian card game Scopone, demonstrating ISMCTS's effectiveness as a fair and challenging opponent, surpassing traditional strategies.

Contribution

It introduces the application of ISMCTS to Scopone, showing its superiority over rule-based strategies and its potential as a competitive AI in incomplete information settings.

Findings

MCTS outperforms all other strategies but requires complete information.

ISMCTS surpasses rule-based players and challenges human players.

Cheating MCTS performs best but is not fair.

Abstract

We present the design of a competitive artificial intelligence for Scopone, a popular Italian card game. We compare rule-based players using the most established strategies (one for beginners and two for advanced players) against players using Monte Carlo Tree Search (MCTS) and Information Set Monte Carlo Tree Search (ISMCTS) with different reward functions and simulation strategies. MCTS requires complete information about the game state and thus implements a cheating player while ISMCTS can deal with incomplete information and thus implements a fair player. Our results show that, as expected, the cheating MCTS outperforms all the other strategies; ISMCTS is stronger than all the rule-based players implementing well-known and most advanced strategies and it also turns out to be a challenging opponent for human players.