Page 7 of 30

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 6:21 pm
by hgm
S.Taylor wrote:Even after the 72 moves (or how ever many moves it stipulates) perfectly played by both sides?
No, after 50 moves of perfect play by both sides.
S.Taylor wrote:It can't if the game was already over, which it normally is, as soon as there are only 7 or 6 pieces left on the board, because the end game TB's already knows the true result with perfect play on both sides.
You still don't get it. The TBs know nothing. They are full of shit. They are known to be full of shit. They indicate positions as won that in truth are dead draws.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 7:04 pm
by Evert
S.Taylor wrote: It can't if the game was already over, which it normally is, as soon as there are only 7 or 6 pieces left on the board, because the end game TB's already knows the true result with perfect play on both sides.
There are a number of well defined ends to a game of chess. Reaching a tablebase position is not one of them.

You can terminate the game in a tablebase position under the assumption that the engines will play optimally (which they might not, if they don't use a tablebase themselves). But then you depend on the tablebase to give you accurate information. In this situation this is simply not the case: the game is a draw under the 50 move rule (as reported by the syzygy tables), but the Gaviota tables do not consider the 50 move rule and so give misleading information.

This is really not difficult. Is the 50 move rule used? Yes, it is. Then using tablebases that don't implement this rule to adjudicate games is wrong. Particularly if there are tablebases that do implement the 50 move rule.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 8:10 pm
by mwyoung
hgm wrote:
mwyoung wrote:This is what happens when you have rules to adjudicate games, and not let the programs play out the moves.
There is of course always some unfairness in adjudication, in that you assume perfect play for engines that might neot be able to deliver that. But you could argue that with participants like these sub-optimal play would be very rare anyway.

But this is much worse. It is whet happens when you have ERRONEOUS rules to adjudicate games.The rules assume the programs are stupid and will blunder,and decide the result accordingly.

This shows how much better the WinBoard approach to tablebase adjudication is: rather than assigning a result, it would just force the engines to move practically instantly (from their own tablebase), to see what they would make of it. That achieves the desired result (namely almost instant termination of the game) without introducing artifacts.
I find this whole controversy rather silly. Because TCEC does much worse things to changes results of their tournament by design. I am sure this was an over site, but using dubious openings to force more decisive games is by design. And skews results much more than this adjudication. But let's get all bent out of shape about an oversight by TCEC.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 8:36 pm
by hgm
How can the choice of openings skew results? The engines always get to play both sides of each opening, not?

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 8:52 pm
by syzygy
mwyoung wrote:I find this whole controversy rather silly. Because TCEC does much worse things to changes results of their tournament by design. I am sure this was an over site, but using dubious openings to force more decisive games is by design. And skews results much more than this adjudication. But let's get all bent out of shape about an oversight by TCEC.
The thing is that it would have been so easy to correct the mistake. What went wrong did not affect game play at all. It only affected the points that were awarded after the game finished. So no need to replay or anything, just award the points correctly.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 9:33 pm
by mwyoung
hgm wrote:How can the choice of openings skew results? The engines always get to play both sides of each opening, not?
Because I said dubious opens. As in not sound. Depending on the strength difference of the two engines playing the opening. This will change results. This is not a criticism of TCEC. I understand why they want more decisive games.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 9:36 pm
by Dan Cooper
None of the TCEC openings I've observed so far are dubious or crazily skewed. They are complex with no opportunity to trade into a boring endgame straight out of the gate.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 9:52 pm
by hgm
mwyoung wrote:Because I said dubious opens. As in not sound.
I still don't see why you think that would skew the results, if each engine has to play both sides. In the worst case the start position is a certain win, and then they would just get a 1-1 on that line, and it is as if the game was never played, and there were just fewer games.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 9:53 pm
by Evert
mwyoung wrote: I find this whole controversy rather silly. Because TCEC does much worse things to changes results of their tournament by design. I am sure this was an over site, but using dubious openings to force more decisive games is by design. And skews results much more than this adjudication. But let's get all bent out of shape about an oversight by TCEC.
Personally, I don't care about TCEC one way or the other. I just get annoyed (in general) by wilful ignorance.
In this particular case: with the 50 move rule in play and under perfect play by both engines (guaranteed by the use of tablebases) the relevant position is a draw. There's simply no room for discussion.

Re: Cursed win at TCEC

Posted: Wed Nov 16, 2016 10:01 pm
by Rochester
Evert wrote:In this particular case: with the 50 move rule in play and under perfect play by both engines (guaranteed by the use of tablebases) the relevant position is a draw. There's simply no room for discussion.
I say "yes". The same they deny the climate change also deny 50 mover and voted the Zaphod Beeblebrox for the president.