gomoku_rl.utils.elo module
- gomoku_rl.utils.elo.compute_elo_ratings(payoff: ndarray, average_rating: float = 1200) ndarray [source]
compute Elo ratings from the payoff matrix
- Parameters:
payoff (np.ndarray) – win rate matrix of shpe (n,n).
average_rating (float, optional) – I play Honor of Kings, so its default value is 1200.
- Returns:
(n,) estimated elo ratings
- Return type:
np.ndarray