Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Still Missing the Point

Author: blass uri

Date: 00:19:48 05/17/00

Go up one level in this thread


On May 17, 2000 at 00:07:44, Robert Hyatt wrote:

>On May 16, 2000 at 22:56:56, Adrien Regimbald wrote:
>
>>Hello,
>>
>>>Your interpretation is that rule is _wrong_.  The rule you quoted was intended
>>>to handle one case: I have a King rook and pawn and a lot of time.  You
>>>have a king and rook and practically no time.  The position is dead drawn,
>>>so I just make moves trying to run you out of time.  You can invoke the above
>>>rule, ask an arbiter and have the game declared drawn.
>>
>>
>>The rule I quoted is intended to be used as it is written.  IE - if you are not
>>making a reasonable effort to win the game, or it is not possible for you to win
>>the game with reasonable play.  The situation that Tiviakov was in does fit this
>>situation - Fritz was not making any attempt in that game to win - Fritz was
>>simply trying to hold on.
>
>THis shows you know nothing about how a program plays chess.  The program
>_always_ is trying to win.  Even when down in material.  It is never just
>"trying to hold on".
>
>The rule has several requirements before it can be used.  One is "clearly
>winning".  In the final position, neither side is "clearly winning".  So it
>doesn't apply at all.
>
>>
>>
>>>You can _not_ invoke the above rule just because you have more material than
>>>I do, but hardly any time.  It doesn't work that way.  The above rule prevents
>>
>>
>>Actually, you can.  If you are <= 2 minutes on the clock, you may stop the clock
>>and call the arbiter over.  The arbiter then decides whether to accept your
>>claim or not.  (yes, there may be a penalty if your claim was unjustified)
>>
>
>And in this case, the arbiter would say "you are not clearly winning, your
>opponent has plenty of material to beat you if you make any sort of mistake.
>I am going to deduct 5 minutes from your clock and the game continues.  Oops.
>You are out of time.  I declare the game over.  You lose."
>
>
>
>
>
>>
>>>my trying to simply run you out of time.  It doesn't establish any threshold
>>>that says "if the side running out of time has two or more extra pawns, he
>>>may claim a draw."  The rule simply says that the side with more time _must_
>>>be trying to make progress to win, and not just be shuffling a piece waiting
>>>on the opponent's flag to fall.  If I can prove I have pushed a pawn every 10
>>>moves, that is _clearly_ making progress.
>>
>>
>>Fritz was simply trying to run Tiviakov out of time.  Fritz didn't have even a
>>remote chance of winning that game by any means other than flagging.  Fritz was
>>not making any efforts to win the game - it was simply trying to cling on and
>>responding to threats.
>
>what are you basing this on?  I have seen my program win _many_ games when
>the evaluation was as low as -6.0.  I don't consider -2 against a human as
>anything serious enough to think about resigning over.

-2.0 is not lost but I think that GM's can at least draw in the final position
of the game in 2 minutes/game.
There are cases when GM can lose position even when they have 6 pawns advnantage
but it is not the case.

I think that the only way that the GM can blunder is if he wants a win and if
you try to win the game there is more room for errors.
>
>
>
>>
>>Pushing a pawn every 10 moves is not a clear sign of progress.  Pushing a pawn
>>can be a good or a bad thing, depending on the position.  If things were as you
>>would have it, you could be in a dead lost position, but your opponent has very
>>little time left, and you can just push a pawn forward only to have it taken by
>>your opponent and somehow call that progress.
>
>
>wrong answer.  Pushing a pawn _is_ progress.  Because the move can't be undone.
>Hence the 50 move rule (this was why it was implemented as it was.)  If I
>push a pawn enough, I run out of pushes.  Then I either promote one or I stop
>making progress and the rule kicks in.  But not in this position.
>
>
>
>
>>
>>Please note that I didn't at any point in time say that "2 pawns up and down on
>>time == draw".  I simply believe that in the position where the operator offered
>>a draw Tiviakov would have a very strong case for a draw, and didn't need the
>>draw offer.
>
>He had _no_ case for a draw claim.  Even claiming a draw there would be an
>instant loss because the arbiter would have to penalize his clock time.  And
>any penalty would result in a time loss.
>
>
>
>>
>>
>>>Yes he did.  He can't claim a draw just because he is a pawn up and about to
>>>lose on time.  That isn't what the rule says.  It says the side with more time
>>>must not be trying to make any progress, before the rule can be used.
>>
>>
>>You keep putting words in my mouth!  I didn't say that Tiviakov can just say
>>"I'm up a pawn or two and down on time, therefore I have a draw".  I'm saying
>>that he may claim a draw, and would most likely get it.  The justification for
>>this is simple - Fritz was making no attempt to win the game on the board, and
>>didn't have any realistic chance..
>
>
>I just got an email reply from an international arbiter I have known for 20+
>years.  His response was the same as mine.

I think that the decision may be dependent on the arbiter.

  This draw offer would be instantly
>rejected and the player would be penalized for stopping the clock to make the
>claim.  He said the criteria must be "easily won" by the side with no time left,
>and a single pawn would not be nearly enough unless it was a simple KP vs K that
>was known to be a draw.  And even then he would make the player demonstrate for
>a few moves that he knew how to draw it.  or win it if he could, in order to
>claim the draw.

I did not have to demonstrate that I know the way to draw the game in the KB vs
KNP endgame.

It is a draw but the claim that it is a draw by tablebases is not enough to
accept a draw claim and I know that in KR vs KRB this claim will not be accepted
by the side without the B even if tablebases say that it is a draw because there
is a real practical chance for mistakes.

The rules are not clear and the decision may be dependent on the arbiter.

Uri



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.