Page 1 of 1

"Complex Tactics" such as bishop versus knight

Posted: Fri Apr 08, 2011 3:04 am
by jplchess
I know that Bobby Fischer preferred bishop over knights. However, today's top softwares can figure out which one is preferable. I also know that pawn structure plays a huge role in it.

Since there are over a quarter of a million posts, please refer about bishop vs. knight. If you can, use the top 4 softwares in 2011 (Houdini, Rybka, Stockfish, and Critter), for information on these two completely different pieces with the similar amount of points (3 pts. vs. 3.5 pts. on average).

Fen diagrams on when the exchange took place a bishop took a knight or vice versa is what I want to see even if it was a draw. I extremely prefer long time controls at the very least one hour on a 1 GHZ hardware.

Another side note is to have a win/draw/loss on bishop versus knight just like 2 softwares are fighting each other. In other words, does the bishop win more often or the knight? Sounds interesting for a database of this sort.

Jonathan

Re: "Complex Tactics" such as bishop versus knight

Posted: Fri Apr 08, 2011 3:22 am
by Dann Corbit

Re: "Complex Tactics" such as bishop versus knight

Posted: Fri Apr 08, 2011 4:44 am
by jplchess
Thanks Dann for those 100 FEN diagrams. I believe that Rybka 3 was the leading software at the time in August 2009.

To elaborate on things I did not type here is that I would like to focus on complex tactics. Chess computers have understood easy tactics to the best of my knowledge for the last 20 years.

Other material imbalances are queen vs. 2 rooks, 2 minor pieces vs. a rook and a pawn, 2 or 3 pawns vs. a minor piece, etc ad nauseum. (Minor piece means bishop or knight.)

Houdini vs. Rybka (let the magician remove its slithered sleeve fish)
Jonathan Lee

Re: "Complex Tactics" such as bishop versus knight

Posted: Fri Apr 08, 2011 11:35 pm
by Tom Barrister
In Bishop vs Knight endgames, the Bishop generally fares better (or does so more often over a large number of positions) because most endgames tend to be more open then closed in nature.

Re: "Complex Tactics" such as bishop versus knight

Posted: Fri Apr 08, 2011 11:47 pm
by Dann Corbit
jplchess wrote:Thanks Dann for those 100 FEN diagrams. I believe that Rybka 3 was the leading software at the time in August 2009.

To elaborate on things I did not type here is that I would like to focus on complex tactics. Chess computers have understood easy tactics to the best of my knowledge for the last 20 years.

Other material imbalances are queen vs. 2 rooks, 2 minor pieces vs. a rook and a pawn, 2 or 3 pawns vs. a minor piece, etc ad nauseum. (Minor piece means bishop or knight.)

Houdini vs. Rybka (let the magician remove its slithered sleeve fish)
Jonathan Lee
The new programs understand material imbalances due to analysis by Larry Kaufman. The idea has been broadly copied now.

If you want to find positions that cause modern computer chess programs to spit out bolts and nuts in anguish, try fortress positions.

Re: "Complex Tactics" such as bishop versus knight

Posted: Sat Apr 09, 2011 12:07 am
by jplchess
Mephisto Genius (the German devil) in 1992 or 1993 knew how many pawns on each side to decide if the bishop or knight is more important.

Complex tactics means what to exchange especially in the middlegame.

I agree Dann fortress/closed positions are tough for the software to handle, but the opening library might minimize that problem. I used the word "might" very loosely.

On arena, I could have Houdini vs. Stockfish fight each other.

Jonathan Lee