I’ve been working on and off for years on a scrabble endgame solver; it uses all these techniques from chess like transposition tables, Negamax with alpha beta

I’ve been working on and off for years on a scrabble endgame solver; it uses all these techniques from chess like transposition tables, Negamax with alpha beta pruning, NegaScout, aspiration search and so on. There’s a French person who built his own endgame solver and this solver is significantly faster than mine, even with all of the optimizations that I’ve put into it. He is kind of secretive about it because it’s closed source and he makes some money on it, but we’ve talked a bit about it, compared some positions and we’ve determined that his move generation algorithm is actually not asoptimized as mine. But he can still solve the endgame faster despite seeing fewer positions, which implies to me that he’s doing a significantly better job of pruning the tree. But when we try to talk details, I asked him for example do you use minimax with alphabeta pruning and he told me like “I’m not sure if I am using minimax or what that is :(“ .. I ask him to describe what he does, he essentially describes minimax with pruning. I’ve sorta figured out that he must be doing some very intelligent version of an aspiration search. It’s really eye-opening because he doesn’t have any of this training. He’s never seen any related algorithms, he’s just figuring all this out on his own.