Cursed win at TCEC

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
syzygy
Posts: 4258
Joined: Tue Feb 28, 2012 10:56 pm

Re: fortress_draw_rule

Post by syzygy » Sun Nov 20, 2016 1:12 pm

Lyudmil Tsvetkov wrote:no one knows how the game would have ended, as top engine analysis is imperfect, even in relatively simple positions.
I am afraid you have far too little knowledge of the engines and how they deal with tablebases.

It is 100% certain that the game would have ended in a draw. Thanks to its tablebases, H5 knew how to keep the mate beyond the 50-move horizon.

Try it out for yourself:
https://syzygy-tables.info/?fen=K5Q1/8/ ... _b_-_-_0_1

You are white and may play whatever moves you like. Black always plays the move shown at the top.

You won't get DTZ down by more than 1 ply on each move and black won't let DTZ go down by more than 1 ply on each move. Since you start at DTZ=124, you need more than 60 moves to get to a capture or mate. The 50-move rule kicks in after 50 moves. QED.

Now that you have learned something new, you might want to reconsider your earlier statements in this thread.

User avatar
MikeB
Posts: 2530
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Cursed win at TCEC

Post by MikeB » Sun Nov 20, 2016 1:55 pm

syzygy wrote:
MikeB wrote:Actually Stockfish has that knowledge, all its need is that little check box to be unchecked to false. If the operator knew the 50 move rule was not going to follow FIDE 50 move rule, the setting should, have been false. So either way , it is operator error.
But then you get the problem that 50-move rule draws that happen with 6 pieces still on the board are thought by the engine to be a win. But these 6-piece positions are not TB-adjudicated by TCEC, so they have to be won over the board and that will not work (even with the 50-move rule disabled the losing engine will make sure to keep the draw).

So what TCEC would need are 5-piece tables that ignore the 50-move rule, and custom-built 6-piece tables that do take into account the 50-move rule as long as the 50 moves happen before a conversion to a 5-piece position. This is not a very satisfactory situation.

The same applies to ICCF but with 5 replaced by 6: they need 6-piece tables that ignore the 50-move rule and custom-built 7-piece tables that take into account the 50-move rule as long as the 50 moves happen before a conversion to a 6-piece position.
Good point , it does sound a whole lot simpler just to follow the FIDE rules that govern the way most people play chess...why start start adjudicating draws (what people would consider to be a draw) as wins in the first place?
Using a faulty automated adjudicator is a little absurd in the first place.

whereagles
Posts: 487
Joined: Thu Nov 13, 2014 11:03 am

Re: fortress_draw_rule

Post by whereagles » Sun Nov 20, 2016 2:10 pm

Lyudmil Tsvetkov wrote:guess the starting position is simply won for white, so the Dragon system might be altogether a bust, unlike other kingside fianchettoe lines
Well, this particular dragon yugo attack line is know to be better for white. The lines with black replying h5 to h4 are more balanced.

Ralph Stoesser
Posts: 408
Joined: Sat Mar 06, 2010 8:28 am

Re: Cursed win at TCEC

Post by Ralph Stoesser » Sun Nov 20, 2016 3:31 pm

MikeB wrote:
syzygy wrote:
MikeB wrote:Actually Stockfish has that knowledge, all its need is that little check box to be unchecked to false. If the operator knew the 50 move rule was not going to follow FIDE 50 move rule, the setting should, have been false. So either way , it is operator error.
But then you get the problem that 50-move rule draws that happen with 6 pieces still on the board are thought by the engine to be a win. But these 6-piece positions are not TB-adjudicated by TCEC, so they have to be won over the board and that will not work (even with the 50-move rule disabled the losing engine will make sure to keep the draw).

So what TCEC would need are 5-piece tables that ignore the 50-move rule, and custom-built 6-piece tables that do take into account the 50-move rule as long as the 50 moves happen before a conversion to a 5-piece position. This is not a very satisfactory situation.

The same applies to ICCF but with 5 replaced by 6: they need 6-piece tables that ignore the 50-move rule and custom-built 7-piece tables that take into account the 50-move rule as long as the 50 moves happen before a conversion to a 6-piece position.
Good point , it does sound a whole lot simpler just to follow the FIDE rules that govern the way most people play chess...why start start adjudicating draws (what people would consider to be a draw) as wins in the first place?
Using a faulty automated adjudicator is a little absurd in the first place.
Just wait a little until cutechess can handle 6 piece tb adjucation and that problem will disappear. There is already a pull request under review.

It may be technically easier to follow the 50 move draw rule in all cases, but many people feel that would be against the main objectives of the game of chess (See article 1.1 - 1.3 FIDE rules for reference). If one side can prove to have an unavoidable checkmate, the game should not end with a draw. It simply feels wrong. Checkmates should take precedence.

syzygy
Posts: 4258
Joined: Tue Feb 28, 2012 10:56 pm

Re: Cursed win at TCEC

Post by syzygy » Sun Nov 20, 2016 3:45 pm

Ralph Stoesser wrote:If one side can prove to have an unavoidable checkmate, the game should not end with a draw. It simply feels wrong. Checkmates should take precedence.
But you seem to be alone in this: http://talkchess.com/forum/viewtopic.php?t=62175

User avatar
Laskos
Posts: 8060
Joined: Wed Jul 26, 2006 8:21 pm

Re: Cursed win at TCEC

Post by Laskos » Sun Nov 20, 2016 4:13 pm

Some people commenting in this thread must apply for jobs as lawyers or jurors in some countryside court. This passion for details of legalistic matters and the deepness of these details make me nostalgic of Rybka/Vasik threads.

Ralph Stoesser
Posts: 408
Joined: Sat Mar 06, 2010 8:28 am

Re: Cursed win at TCEC

Post by Ralph Stoesser » Sun Nov 20, 2016 4:48 pm

syzygy wrote:
Ralph Stoesser wrote:If one side can prove to have an unavoidable checkmate, the game should not end with a draw. It simply feels wrong. Checkmates should take precedence.
But you seem to be alone in this: http://talkchess.com/forum/viewtopic.php?t=62175
That one is about the clock, right? No flag fell issues in what we talk about here, so no need to mention that the clock should take priority in such a case. That of course feels absolutely right for most people, because we all know that our time on earth is limited. When our time is over we are not able to complete tasks in the future. That's somehow logical....

FIDE is not the ruler for engine competitions. FIDE rules are made for humans exclusively. No piece of software will ever be able to follow article 4.1, because software neither have two hands nor one. So we are somewhat free to do what we want with the FIDE rule set. That seems to be a new message for some people, and for sure it is a good message, so I think it's worth to mention it again here.

Ralph Stoesser
Posts: 408
Joined: Sat Mar 06, 2010 8:28 am

Re: Cursed win at TCEC

Post by Ralph Stoesser » Sun Nov 20, 2016 4:50 pm

Laskos wrote:Some people commenting in this thread must apply for jobs as lawyers or jurors in some countryside court. This passion for details of legalistic matters and the deepness of these details make me nostalgic of Rybka/Vasik threads.
:lol:

User avatar
MikeB
Posts: 2530
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Cursed win at TCEC

Post by MikeB » Sun Nov 20, 2016 5:00 pm

Ralph Stoesser wrote:
MikeB wrote:
syzygy wrote:
MikeB wrote:Actually Stockfish has that knowledge, all its need is that little check box to be unchecked to false. If the operator knew the 50 move rule was not going to follow FIDE 50 move rule, the setting should, have been false. So either way , it is operator error.
But then you get the problem that 50-move rule draws that happen with 6 pieces still on the board are thought by the engine to be a win. But these 6-piece positions are not TB-adjudicated by TCEC, so they have to be won over the board and that will not work (even with the 50-move rule disabled the losing engine will make sure to keep the draw).

So what TCEC would need are 5-piece tables that ignore the 50-move rule, and custom-built 6-piece tables that do take into account the 50-move rule as long as the 50 moves happen before a conversion to a 5-piece position. This is not a very satisfactory situation.

The same applies to ICCF but with 5 replaced by 6: they need 6-piece tables that ignore the 50-move rule and custom-built 7-piece tables that take into account the 50-move rule as long as the 50 moves happen before a conversion to a 6-piece position.
Good point , it does sound a whole lot simpler just to follow the FIDE rules that govern the way most people play chess...why start start adjudicating draws (what people would consider to be a draw) as wins in the first place?
Using a faulty automated adjudicator is a little absurd in the first place.
Just wait a little until cutechess can handle 6 piece tb adjucation and that problem will disappear. There is already a pull request under review.

It may be technically easier to follow the 50 move draw rule in all cases, but many people feel that would be against the main objectives of the game of chess (See article 1.1 - 1.3 FIDE rules for reference). If one side can prove to have an unavoidable checkmate, the game should not end with a draw. It simply feels wrong. Checkmates should take precedence.
I'm fine with that as long as that is well communicated in advance and understood, and obviously the appropriate settings , i.e, "setoption name Syzygy50MoveRule value false" are used and it is obvious they were not.

syzygy
Posts: 4258
Joined: Tue Feb 28, 2012 10:56 pm

Re: Cursed win at TCEC

Post by syzygy » Sun Nov 20, 2016 5:34 pm

MikeB wrote:
Ralph Stoesser wrote:It may be technically easier to follow the 50 move draw rule in all cases, but many people feel that would be against the main objectives of the game of chess (See article 1.1 - 1.3 FIDE rules for reference). If one side can prove to have an unavoidable checkmate, the game should not end with a draw. It simply feels wrong. Checkmates should take precedence.
I'm fine with that as long as that is well communicated in advance and understood, and obviously the appropriate settings , i.e, "setoption name Syzygy50MoveRule value false" are used and it is obvious they were not.
I am looking forward to the discussion after a game has ended in a 50-move draw with 6 pieces left on the board which is shown to be a "win" by TBs. I guess such a draw will have to be manually corrected to a win?

I am also looking forward to the discussion after a game has ended in a 50-move draw with 7 pieces left on the board which is shown to be a "win" by the Lomonosov TBs. I guess such a draw will have to be manually corrected as well?

I am even more looking forward to the discussion after a game has ended in a 50-move draw with 10 pieces left on the board which is shown to be a "win" by running finalgen for two weeks. I guess such a draw will have to be manually corrected too?

The "unavoidable checkmate" concept is simply broken.

Post Reply