Identifying Strategies in Dominion Using Playtrace Clustering

We demonstrate the use of playtraces and playtrace clustering to identify strategies and card synergies in deck building card games, using Dominion as an example. We analyze playtraces generated from both online human play and a variety of AI agents, examining two types: card counts by round in a player's deck and N-Grams generated from player actions. Using both the $L_{k}$ -norm and Jensen–Shannon distance measures, in-conjunction with $K$ -Means, $K$ -Medoids and DBSCAN algorithms, we show that playtraces and distinct clusters can reveal both longer term strategies and card synergies. In addition, we use a restricted play framework to increase the variation in strategies and tactics explored by AI agents. Finally, we suggest that the game-agnostic, N-Gram-based approach may support strategy exploration in tabletop games more broadly.