Hi dears programmers,
I thought about something, an hypothese. Before i will ask something :
What is the difference between an opening book and an endgame database?
ie opening book isn't to have a good or perfect opening game and endgame database to have a perfect endgame?
If, and i suppose it, we generate a huge opening book for i suppose 75 plies, that i suppose again that it's generated by autoplay (computer vs computer himself) , with the idea to reach the 8 or 7 endgame database.
With this, can we speak of self-learning and can we reach a perfect game ?
IMO with such program you will never loose a game that does not mean that your game will be perfect. In order to be quite provocative suppose it is proven that after 31-27 16-21?? the black side can still reach the draw. How the program can be know that 16-21 is a very poor move if has learned that after 31-27 all moves (including 16-21) are drawing moves?
As another example suppose a world championship tournament with your programm and the top 11 GMI players. Because your program has not the capability to create great difficullties for the human opponent your program may obtain let's say 1 win and 10 draws but the winner of the tournament will typically obtain 3 wins and 8 draws.
A perfect game is a game in which your are never in a losing position but at the same time a game in which your opponent has in many occassions great difficulties in finding the good move. BTW letting your opponent with an obvious forcing move is not a good idea because there are no difficulties.
Thank very much, i have a question, the programmer has the ability to select all the opening that put the opponents in difficult position, ie the winning games generated by autoplay. And i suppose that the engine must be strong base to a good evaluation.
Suppose that after 1.31-27 16-21?? Your program focus only, after of course playing all the best way to convert this advantage in win, on the best moves that can win in this position, it will beat any program or grand master who will do this mistake. Only if in all possibilities this opening is draw.
The main idea of what i am saying is that if the program take after self-playing all the opening that reach to winning game it must be stronger and will only play these opening. I suppose that it will play a million of game, anything has a cost.
And when it has a good opening that will be very deep it will take advantage in all his game, because again it's focused on ONLY OPENING THAT REACH TO A WINNING GAME.