After Rybka 4 came out have we hit a WALL concerning software with tactics and postition? This excludes the opening library, endgame tablebases, and hardware. If there is room for improvement, then what would it be besides closed games?
I was a big fan of Mephisto Genius when it came out 17 years ago. The executable was 54,313 bytes for version number 2. That one was a senior master at the time and a grandmaster in tactics.
I am also a big fan of Fischer Random because obviously the opening library is very limited.
The horizon effect is still a problem, but I guess it will remain so for a long time to come.
Therefore, further development of long term planning could be a goal?
I'm not a programmer though.
jplchess wrote:After Rybka 4 came out have we hit a WALL concerning software with tactics and postition? This excludes the opening library, endgame tablebases, and hardware. If there is room for improvement, then what would it be besides closed games?
I was a big fan of Mephisto Genius when it came out 17 years ago. The executable was 54,313 bytes for version number 2. That one was a senior master at the time and a grandmaster in tactics.
I am also a big fan of Fischer Random because obviously the opening library is very limited.
Please make some constructive posts.
We haven't hit a wall, not by a fair margin. Rybka 4 displays numerous weaknesses still IMHO, whether in the opening/middlegame, to the endgame.
"Tactics are the bricks and sticks that make up a game, but positional play is the architectural blueprint."
You think that we hit a wall? Let me ask you who was number one before Rybka, number two? As long as there are chess programmers there will be the possibliity of someone improving the game. In my opinion only when chess has been somewhat solbed will we come close to a wall.
Every time someone creates a new super program many people seem to think it is the last and greatest only to have someone else make a better one.
Bill
jplchess wrote:After Rybka 4 came out have we hit a WALL concerning software with tactics and postition?
No, there still appears that there's a lot to improve upon, I'd guess.
Engines are not perfect at strategy I can assure you, because I have gone through validating 2000 positions for STS. It's toughest to evaluate the position. Strategy is what prevents computers from permanently solving chess, if there's ever such a thing. It's way too deep.
jplchess wrote:After Rybka 4 came out have we hit a WALL concerning software with tactics and postition?
No, there still appears that there's a lot to improve upon, I'd guess.
Engines are not perfect at strategy I can assure you, because I have gone through validating 2000 positions for STS. It's toughest to evaluate the position. Strategy is what prevents computers from permanently solving chess, if there's ever such a thing. It's way too deep.
Engines are near perfect in dynamic tactics.
I say not even close to pefect, but far better than fleshlings, like me and you
A lot of things can still be improved upon. R4 is obviously a Rybka with some stuff omitted to preserve some secrets, so we really don't know how far the engine has progressed. But I think better usage of hardware will improve engines by quite a bit. If you play an engine against itself with a time handicap, you will find that it can score much better with more time. Some new technologies such as Monte Carlo might come into effect in real time. Or maybe an End Game module. With more and more cores also being available on a simple home PC, the engines will start acting like clusters and will have cores to look at mundane things or dramatic sacs that would otherwise have been pruned. I think there is at least a good 200 to 300 ELO left before we start reaching a level of no more improvements ... at least at LTC's.
