A new Scrabble engine uses probability-based move selection to compete at a championship level. The system prioritizes board control and future tile availability over immediate point maximization. This approach demonstrates how targeted probabilistic modeling outperforms greedy algorithms in strategic board games. It offers a blueprint for optimizing decision-making in constrained state spaces.