Cluster Pays

Cluster games award wins when there are sufficiently many neighboring like-symbols. Neighbours must share the same reel or row, where diagonal connections do not count towards the cluster size. A minimum of 5 like-symbols is typical, though this can be defined in GameConfig class. Since it is possible for up to num_rows * num_columns winning symbols to occur, it is common to define a particular payout range. For example 5 kind pays p1 6-7 kind to pay p2, 8-10 kind to pay p3 and 12+ symbols pay p4 etc... Instead of manually including all possible pay combinations in config.paytable there is a convert_range_table() function in the Config class which takes in a symbol range, name and payout amount which is used to generate all config.paytable entries. This pay group should be of the format:

    paygroup = {
        ((min_combination[int], max_combination[int]), name[str]) : payout[float],
        ... 
    }

Ranges defined in min_combination and max_combination are inclusive, so for example if the 5-kind payout for symbol H1 pays 10x, this would be written as: ((5,5),H1): 10.

Often (though not always) cluster pays games include a tumbling mechanic. Within the cluster sample game for example, while there are still winning combinations, the board is tumbled, wins are evaluated for the new board, the wallet manager is updated and relevant events are emitted:

    while self.win_data["totalWin"] > 0 and not (self.wincap_triggered):
        self.tumble_game_board()
        self.win_data = self.get_cluster_data(record_wins=True)
        self.win_manager.update_spinwin(self.win_data["totalWin"])
        self.emit_tumble_win_events()

Clusters are found using a Breath First Search (BFS) algorithm. Wild attributes can be set (wild is the default value). Wild symbols can contribute to multiple clusters, including those formed by different symbols.