Curiously, I was looking at “The Italian Game” Powerbook on Chessbase. Apparently it’s primarily derived from engine-to-engine games on the Chessbase server. I would assume other powerbooks would been have made from similar material.
Quite interesting. I would presume most of the engines would have an opening book of whatnot, but that, in this case the The Italian game Powerbook, had been produced by analyzing opening moves from the proper opening, in this case, Italian openings from engine-to-engine games.
I’m not surprised actually. Although according to Chessbase, the Italian game is in somewhat of a Renaissance at the moment, there would be no way an opening book could be properly made by only using high level human games. They used over 800,00 games. I’m pretty sure there hasn’t been 800,000+ grandmaster level human chess games in the Italian Opening. So, I’m actually glad Chessbase decided to use high level engine-to-engine games. I think they did add some OTB and correspondence human games, but a small fraction compared to the bulk of engine-to-engine games.
I suppose other powerbooks, or opening books for any other chess engine, would be better off using the bulk of their moves from engine-to-engine games. At least for specific openings. Apparently Fritz Powerbook 18 is still primarily culled from actual human play.
Although Google has all but proven A.I. can master the game of chess, even mortal engines like chess engines one could load onto a personal computer still can add to the body of chess knowledge.
There is a lot more machines can add to chess theory, and I suppose the Renaissance of Chess has many years to go before chess theory matures, even if more of the theory comes from machines and not humans.
Still, theory or not, I doubt A.I. could truly replace the creativity of humans, short of a 32 piece endgame tablebase. I think for many many years, the human element will continue to add to the theory of chess.
Just a few thoughts.