Author: Dann Corbit
Date: 17:12:36 09/17/02
Go up one level in this thread
On September 17, 2002 at 19:49:56, David Rasmussen wrote: >On September 17, 2002 at 19:08:09, Dann Corbit wrote: > >>On September 17, 2002 at 19:02:24, David Rasmussen wrote: >>[snip] >>>All of these alternative mates are not solutions, that is the point. I don't >>>think any program will have problems finding _some_ mate in these positions. The >>>hard thing is to find the right (shortest) mate in these last positions, with >>>normal playing techniques (null move etc.) . >> >>I disagree. Once a program has found a mate that is proven, the value of the >>solution is +1 from a game theoretic standpoint. The lenght to the mate is >>irrelevant, unless the program has neglected some checkmate rule. As far as >>getting the shortest possible checkmate, that is a job for a dedicated mate >>finder. >> > >That is a valid opinion (which I happen to agree with), but that is not the >point. I chose those positions specifically because they seem to highlight some >of the problems of some search-techniques, such as null-move. If one is happy >with any mate, one should disregard the positions from #931 and on. > >Even though mate finding is often not important in itself, I think it is >fascinating that some of the good programs manages to find these mates so >quickly, in normal playing mode, not a special mate search mode. What I am wondering is: "Is there really any problem at all?" There is a problem if an engine chooses a stalemate or loss instead of a checkmate. Perhaps a more interesting set of positions would be where a checkmate is deeply hidden by several consecutive zugzwang positions and the normal search will lead to a draw or a loss instead. When we are talking about a mate in 3 instead of a mate in 4, the only thing lost is some of the beauty. I will agree that it is sad to lose beauty. But that's why Heiner invented Chest -- so we won't have to suffer. ;-)
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.