Super bullet chess: LC0 network 31510 vs Fire 7.1

Discussion of computer chess matches and engine tournaments.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Hai
Posts: 455
Joined: Sun Aug 04, 2013 11:19 am

Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by Hai » Sat Nov 24, 2018 11:36 am

Super bullet chess: LC0 network 31510 vs Fire 7.1

40 games
ponder on
LC0 with super weak gpu 1 kN/s
Fire 7.1 with 4 cores @ 4.2 ghz
Time control: 1 min per game + 1 second per move.

Result:
+11 =15 -14
LC0 18.5 points
Fire 21.5 points
= LC0 is -26 elo weaker.

Hai
Posts: 455
Joined: Sun Aug 04, 2013 11:19 am

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by Hai » Sat Nov 24, 2018 12:52 pm

LC0 is a genius.
At the beginning of a game you should take 2 pawns and the 2 queens of the board and start an attack with your king in the middle of the board.

1.f4:

Hai
Posts: 455
Joined: Sun Aug 04, 2013 11:19 am

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by Hai » Sat Nov 24, 2018 12:59 pm

How to win with 1.h4 and kingside attack


Uri Blass
Posts: 8606
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by Uri Blass » Sat Nov 24, 2018 6:46 pm

Hai wrote:
Sat Nov 24, 2018 12:52 pm
LC0 is a genius.
At the beginning of a game you should take 2 pawns and the 2 queens of the board and start an attack with your king in the middle of the board.

1.f4:
I do not consider LC0 to be a genius
She clearly showed that she is a stupid engine in this game.
Other chess programs or humans who are not beginners do not need 80 moves to mate with KQBP vs K and can win faster.
LC0 is lucky that she has an increment of 1 second otherwise she probably could often lose on time.

User avatar
CMCanavessi
Posts: 844
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by CMCanavessi » Sat Nov 24, 2018 8:09 pm

And how many of those "chess programs or humans who are not beginners" can beat Fire 7.1?
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

F. Bluemers
Posts: 860
Joined: Thu Mar 09, 2006 10:21 pm
Location: Nederland
Contact:

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by F. Bluemers » Sun Nov 25, 2018 12:06 am

CMCanavessi wrote:
Sat Nov 24, 2018 8:09 pm
And how many of those "chess programs or humans who are not beginners" can beat Fire 7.1?
I think this is a bit unfair towards Uri. He sometimes seems to be trolling a bit , but he is spot on here.
The play in the endings is just extremely lousy.It can't hurt to point this out.
In the extreme case in otb meetings ab-searchers will just click a box when playing lc0 and have an easy win even though rating lists
would predict otherwise.
Best
Fonzy

User avatar
CMCanavessi
Posts: 844
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by CMCanavessi » Sun Nov 25, 2018 12:25 am

F. Bluemers wrote:
Sun Nov 25, 2018 12:06 am
CMCanavessi wrote:
Sat Nov 24, 2018 8:09 pm
And how many of those "chess programs or humans who are not beginners" can beat Fire 7.1?
I think this is a bit unfair towards Uri. He sometimes seems to be trolling a bit , but he is spot on here.
The play in the endings is just extremely lousy.It can't hurt to point this out.
In the extreme case in otb meetings ab-searchers will just click a box when playing lc0 and have an easy win even though rating lists
would predict otherwise.
Best
Fonzy
We already know that, currently, the endgame is the weakest part in lc0... yet, Leela still won the game. And besides, he said (quote) "She clearly showed that she is a stupid engine in this game". A stupid engine that beat the 4th or 5th strongest engine in the world?
"Game" is not the same as "endgame".
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

F. Bluemers
Posts: 860
Joined: Thu Mar 09, 2006 10:21 pm
Location: Nederland
Contact:

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by F. Bluemers » Sun Nov 25, 2018 1:07 am

CMCanavessi wrote:
Sun Nov 25, 2018 12:25 am
F. Bluemers wrote:
Sun Nov 25, 2018 12:06 am
CMCanavessi wrote:
Sat Nov 24, 2018 8:09 pm
And how many of those "chess programs or humans who are not beginners" can beat Fire 7.1?
I think this is a bit unfair towards Uri. He sometimes seems to be trolling a bit , but he is spot on here.
The play in the endings is just extremely lousy.It can't hurt to point this out.
In the extreme case in otb meetings ab-searchers will just click a box when playing lc0 and have an easy win even though rating lists
would predict otherwise.
Best
Fonzy
We already know that, currently, the endgame is the weakest part in lc0... yet, Leela still won the game. And besides, he said (quote) "She clearly showed that she is a stupid engine in this game". A stupid engine that beat the 4th or 5th strongest engine in the world?
"Game" is not the same as "endgame".
Are you taking Uri's critisim personal?
FWIW these endgames look like soccer games where all members of the opponent team got a red card and the ball is played all over the field except were it should land,eventually winning is not an excuse …
Why not try to fix it?
Best
Fonzy

carldaman
Posts: 1737
Joined: Sat Jun 02, 2012 12:13 am

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by carldaman » Sun Nov 25, 2018 1:18 am

I raised a similar point.
Yes, it can achieve a winning endgame in grandmasterly fashion, but just when things should be super easy they become almost impossibly hard, taking the longest possible time to win such a simple ending. Could be a faulty TB implementation? And why are the depths getting lower and not higher when there are only 3 (three) pieces left on the board?

What do the experts think?
Then nabildanial replied with:
It is an artifact or a side-effect to the current test30 "TB rescoring" implementation. She can convert those endgames quite well, at least better than previous test runs at the same LR, but it takes a little bit of pieces reshuffling and just before the games approaching the 50-move draw she'll start converting those endgames.
http://talkchess.com/forum3/viewtopic.php?f=2&t=68963

I do hope they make it better because the way it is now is like a bad joke (after coming this far, to the developers' credit).

User avatar
CMCanavessi
Posts: 844
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: Super bullet chess: LC0 network 31510 vs Fire 7.1

Post by CMCanavessi » Sun Nov 25, 2018 2:49 am

F. Bluemers wrote:
Sun Nov 25, 2018 1:07 am
CMCanavessi wrote:
Sun Nov 25, 2018 12:25 am
F. Bluemers wrote:
Sun Nov 25, 2018 12:06 am
CMCanavessi wrote:
Sat Nov 24, 2018 8:09 pm
And how many of those "chess programs or humans who are not beginners" can beat Fire 7.1?
I think this is a bit unfair towards Uri. He sometimes seems to be trolling a bit , but he is spot on here.
The play in the endings is just extremely lousy.It can't hurt to point this out.
In the extreme case in otb meetings ab-searchers will just click a box when playing lc0 and have an easy win even though rating lists
would predict otherwise.
Best
Fonzy
We already know that, currently, the endgame is the weakest part in lc0... yet, Leela still won the game. And besides, he said (quote) "She clearly showed that she is a stupid engine in this game". A stupid engine that beat the 4th or 5th strongest engine in the world?
"Game" is not the same as "endgame".
Are you taking Uri's critisim personal?
FWIW these endgames look like soccer games where all members of the opponent team got a red card and the ball is played all over the field except were it should land,eventually winning is not an excuse …
Why not try to fix it?
Best
Fonzy
Not personal at all, just pointing at an obvious trolling.
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

Post Reply