Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Milos
Posts: 3563
Joined: Wed Nov 25, 2009 12:47 am

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Milos » Sat May 23, 2020 11:24 pm

Milos wrote:
Sat May 23, 2020 10:53 pm
The problem is with CCC, they are a joke, they play stupid bullet tournaments and then assume as some novelty some doubtful move by Lc0. If this was LTC game Lc0 would have been crushed.
Why is 8.d5 a blunder? Well before it white has an active advantage of about 0.8. After it comes continuation: 8...cxb5, 9.d6 h6 black has at least a draw (white's best play leads to perpetual). Losing 0.8 advantage in one move is a blunder. Period.
Last edited by Milos on Sat May 23, 2020 11:40 pm, edited 1 time in total.

Milos
Posts: 3563
Joined: Wed Nov 25, 2009 12:47 am

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Milos » Sat May 23, 2020 11:29 pm

Nay Lin Tun wrote:
Sat May 23, 2020 11:16 pm
It is like trolling. They made 30 minutes video but didnot put pgn for another 5 secs. :lol: :mrgreen:

Agadmator channel has pgn.
That IM Daniel Rensch is a clueless tool. He can't use his own proper computer analysis and is just blabbering stuff based on mediocre chess.com engine analysis.

Leo
Posts: 899
Joined: Fri Sep 16, 2016 4:55 pm
Location: USA/Minnesota
Full name: Leo Anger

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Leo » Sun May 24, 2020 12:05 am

Sure was fun to watch.
Advanced Micro Devices fan.

MikeGL
Posts: 920
Joined: Thu Sep 01, 2011 12:49 pm

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by MikeGL » Sun May 24, 2020 1:42 am

MonteCarlo wrote:
Sat May 23, 2020 10:58 pm
Yeah, different event and different game.

Here's the game from PGN (full PGN so far for the event can be found at https://cccfiles.chess.com/archive/tour ... 105870.pgn):



Cheers!

Impressive game. Just like watching one of those Morphy classics. But this is amazing because black is SF known for its brutal accuracy, but was still crushed with passive 2R+B vs Q in mid game, but white has activity, mobility.
I told my wife that a husband is like a fine wine; he gets better with age. The next day, she locked me in the cellar.

User avatar
Ovyron
Posts: 4002
Joined: Tue Jul 03, 2007 2:30 am

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Ovyron » Sun May 24, 2020 8:30 am

Leo wrote:
Sun May 24, 2020 12:05 am
Sure was fun to watch.
Yes! This is what computer chess is about!

Dann Corbit
Posts: 10835
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Dann Corbit » Mon May 25, 2020 2:08 am

MonteCarlo wrote:
Sat May 23, 2020 10:58 pm
Here's the game from PGN (full PGN so far for the event can be found at https://cccfiles.chess.com/archive/tour ... 105870.pgn):
I don't suppose that there is some place that says:
For tournament <X> the PGN is found at <locx>
For tournament <Y> the PGN is found at <locy>
For tournament <Z> the PGN is found at <locz>

Or are we just supposed to know where to look?
TCEC has never been hard to find the PGN files.

The computer chess content for CCC is wonderful.
The web design is a bit hard for some of us to follow.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

MonteCarlo
Posts: 98
Joined: Sun Dec 25, 2016 3:59 pm

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by MonteCarlo » Mon May 25, 2020 3:41 pm

Yes, there is a way.

On the main page you can select events with the dropdown circled on the right, and then click on the download link circled next to the board:

https://ibb.co/b7spt0g

Then in the download pop-up:

https://ibb.co/QcFN8YL

Cheers!

OliverBr
Posts: 253
Joined: Tue Dec 18, 2007 8:38 pm
Location: Munich, Germany
Contact:

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by OliverBr » Sat May 30, 2020 3:54 pm

Milos wrote:
Sat May 23, 2020 11:24 pm
Milos wrote:
Sat May 23, 2020 10:53 pm
The problem is with CCC, they are a joke, they play stupid bullet tournaments and then assume as some novelty some doubtful move by Lc0. If this was LTC game Lc0 would have been crushed.
Why is 8.d5 a blunder? Well before it white has an active advantage of about 0.8. After it comes continuation: 8...cxb5, 9.d6 h6 black has at least a draw (white's best play leads to perpetual). Losing 0.8 advantage in one move is a blunder. Period.
Leela hasn't been crushed by anybody or anything for a long, long time. She only loses (no crush) a match when she was force in a really bad opening.

Which program did you use for your analysis that sees a draw after 8. d5?

Cornfed
Posts: 18
Joined: Sun Apr 26, 2020 9:40 pm
Full name: Brian D. Smith

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Cornfed » Sat May 30, 2020 6:43 pm

Observation: Commentators are now fond of saying "that's an idea borrowed from Alpha Zero"...heard that in response to white playing a2-a4 recently on chess24. They are NOT saying the engine played that in this position before of course.

Really though, it's more correct to say that an engine is 'borrowing ideas' (not strickly 'borrowing perhaps) from humans. What I mean is that chess players simply 'know' to play a4 in certain positions...intuition based on seeing a lot of games in vaguely similar situations.

One sees similar things in Komodo MTCS...kind of 'positional' play not 100% based on calculation of 'best replies'. More akin to 'feel' in human chess. The same goes for certain piece sacks - you don't see any more than 'vague compensation' sometimes....and that makes chess interesting and not accounting.

Milos
Posts: 3563
Joined: Wed Nov 25, 2009 12:47 am

Re: Leela Chess Zero Makes Mind-Blowing Piece Sacrifice!

Post by Milos » Sat May 30, 2020 11:29 pm

OliverBr wrote:
Sat May 30, 2020 3:54 pm
Milos wrote:
Sat May 23, 2020 11:24 pm
Milos wrote:
Sat May 23, 2020 10:53 pm
The problem is with CCC, they are a joke, they play stupid bullet tournaments and then assume as some novelty some doubtful move by Lc0. If this was LTC game Lc0 would have been crushed.
Why is 8.d5 a blunder? Well before it white has an active advantage of about 0.8. After it comes continuation: 8...cxb5, 9.d6 h6 black has at least a draw (white's best play leads to perpetual). Losing 0.8 advantage in one move is a blunder. Period.
Leela hasn't been crushed by anybody or anything for a long, long time. She only loses (no crush) a match when she was force in a really bad opening.

Which program did you use for your analysis that sees a draw after 8. d5?
You seems to be a fanboy with not much chess analysis experience. You see, the question what program I used totally misses the point. Everyone who knows anything about computer chess uses SF for primary analysis, because it's by far the best program for analyzing games. Lc0 is totally useless for that matter. Ofc one can use MCTS based programs like Lc0 or Komodo MCTS to find ideas for different replies, but that's all.
So to go back to the position after I wrote up the drawing reply by black (9...h6, and as a matter of fact Nc6 or Ng6 also draw), I can write the strongest continuation, but I don't think that would matter because you probably wouldn't know what to do with it, would you?
If you wish you can enlighten us and write the winning line for white after 9...h6. ;)

Post Reply