Author: Dann Corbit
Date: 14:54:34 06/15/04
Go up one level in this thread
On June 15, 2004 at 17:25:45, David Dahlem wrote: >On June 15, 2004 at 16:39:02, F. Huber wrote: >>On June 15, 2004 at 16:21:07, David Dahlem wrote: >>>On June 15, 2004 at 15:39:26, F. Huber wrote: >>>>On June 15, 2004 at 15:33:41, David Dahlem wrote: >>>>>One of the problems with the current method of testing engines with test suites >>>>>(e.g. WM-Test) is the problem of proving that the proposed solution move is >>>>>actually the best move, especially with positions of a positional nature. >>>>>Perhaps a new method would avoid this problem, namely a suite of mate positions, >>>>>with known, more easily proven solutions? Time to solution could be the criteria >>>>>by which engines are evaluated. >>>>> >>>>>Just an idea. Any thoughts? Would this work? >>>>> >>>>>Regards >>>>>Dave >>>> >>>>Hi Dave, >>>> >>>>a really very good idea! >>>> >>>>(Why? - Because here also ChestUCI could participate ;-)) >>>> >>>>Regards, >>>>Franz. >>> >>>Hi Franz >>> >>>The excellent ChestUCI mate solving could be used to "prove" the solutions, but >>>shouldn't be allowed to compete, since it would blow away the competition. :-) >> >>This will probably depend on the "goal" of this test: >>If it´s only to find _any_ mate (not the shortest one), then Chest would >>would certainly _not_ blow away anyone - >>in this case I would rather guess TheKing or Yace as winner! >> >>Regards, >>Franz. > >A "proven forced mate" problem would consist of a number of "best moves", so to >be considered solved, only the shortest mate should be found. Otherwise the >engine would be finding moves that are not the "best moves". Does that make >sense? :-) It is not impossible for there to be dozens of best moves, under this definition. Imagine a simple case, where six distinct moves lead to a mate. Imagine another situation, where there is a mate in 7 that is simple, follows forced lines, and totally obvious. There is also a mate in 5 that is fraught with peril. Is the mate in 5 better? In one sense it is, but in another sense it is worse. If a program sees a mate in 50 it will win. If a program sees a mate in 1 it will win. So from a game theoretic standpoint (as far as outcome) neither mate is superior to the other.
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.