A new Scrabble engine uses probability-based move selection to compete at championship levels. The system prioritizes board control and tile equity over simple high-scoring plays. This approach outperforms traditional greedy algorithms in long-term game strategy. It demonstrates how targeted probabilistic modeling solves specific combinatorial challenges better than general-purpose search methods.