Author: Robert Hyatt
Date: 15:56:09 06/14/03
Go up one level in this thread
On June 14, 2003 at 08:18:14, Bo Persson wrote: >On June 13, 2003 at 17:20:04, Helmut Conrady wrote: > >>On June 13, 2003 at 14:41:04, Jeroen Noomen wrote: >> >>>On June 13, 2003 at 12:03:58, Michael Vox wrote: >>> >>>GM Shipov's conclusion is 'that the programmers still have a lot >>>of work to do'. Actually the chance of getting such a position >>>in a real game is around 0,00 >> >>I disagree. Look at the one with BPPP-B or BP-P for example. >> >>- even some of the positions cannot >>>be reached in a real game and are illegal- so I fail to see the >>>need to work on such positions... >>> >> >>IMO all positions are legal. > >But that wasn't the point. > >If a program is stupid enough to move all five pawn to the rook file, it doesn't >matter much if the score is wrong as well. If it would never get into the >position by itself, the score doesn't matter at all. > >If any program promoted to 8 same color bishops, I would erase it immediately >even if it scored it as a draw! > > >Some of these "problems" are like asking a real estate broker to estimate the >value of a house with one bedroom and 8 (same color) bathrooms. The answer is >pretty uninteresting. It is more interesting than you think. IE what if you have two connected passers, and a bishop, and your opponent has a king and knight, and offers the knight for free, to get both your pawns to the same (wrong) rook file. You just turned a win into a draw. I faced this in 1981 as I mentioned in another reply, luckily Cray Blitz knew about bishop + wrong rook pawn(s) and didn't take the piece and went on to win. Ken Thompson tried it on Belle and said "Belle doesn't understand this and would have to search deeply enough to see the draw, which is not possible in reasonable time..." I don't think you can ignore these kinds of positions, if they happen often enough to affect games. I've seen this on ICC way more than once or twice even though having 3-4-5 rook pawns is rare. But two is not that uncommon if a human is trying to swindle the computer out of an easy win.... > >> >>Helmut > > > >Bo Persson >bop2@telia.com
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.