Show HN: Searching for optimal strategies in a puzzle game with chance (Drop7)
Summary
Show HN presents Drop7 research focused on optimizing strategies in a probabilistic puzzle game. The piece catalogs a wide range of AI approaches (expectimax, q-learning, n-tuple networks, NNUE, policy gradients, evolutionary methods, Monte Carlo tree search, rollouts, and more), reports on empirical results and thresholds, and emphasizes reproducible research with public datasets and benchmarks.