One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

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.
Nay Lin Tun
Posts: 530
Joined: Mon Jan 16, 2012 5:34 am

One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Nay Lin Tun » Sat Oct 27, 2018 5:01 pm




AB engines do pseudo sacrifice, it means they always see material or other advantage back within their depth. However , Leela play with risks(winning probability), I am pretty sure that her MCTS wount see she would gain back a piece in next 25 moves, depth 50.

User avatar
Steve Maughan
Posts: 1061
Joined: Wed Mar 08, 2006 7:28 pm
Location: Florida, USA
Contact:

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Steve Maughan » Sat Oct 27, 2018 7:11 pm

Great game - thanks for posting. White missed a mate in 17 at move 45 (Rg8+)

Steve
http://www.chessprogramming.net - Maverick Chess Engine

Werewolf
Posts: 1198
Joined: Thu Sep 18, 2008 8:24 pm

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Werewolf » Sat Oct 27, 2018 8:37 pm

SF finds 12.gh6 quickly and with an advantage.

It has more trouble finding 9.g4 and 10.Rg1 - but I don't know if those moves are best.

chrisw
Posts: 2202
Joined: Tue Apr 03, 2012 2:28 pm

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by chrisw » Sat Oct 27, 2018 9:16 pm

Werewolf wrote:
Sat Oct 27, 2018 8:37 pm
SF finds 12.gh6 quickly and with an advantage.

It has more trouble finding 9.g4 and 10.Rg1 - but I don't know if those moves are best.
the concept "best move" has been consigned to the dustbin of history

Nay Lin Tun
Posts: 530
Joined: Mon Jan 16, 2012 5:34 am

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Nay Lin Tun » Sat Oct 27, 2018 11:16 pm

Werewolf wrote:
Sat Oct 27, 2018 8:37 pm
SF finds 12.gh6 quickly and with an advantage.

It has more trouble finding 9.g4 and 10.Rg1 - but I don't know if those moves are best.
She always love such 6.g4 !! (she does such attack whenever black plays h6??) .Here is vs Houdini 6.
Last edited by Nay Lin Tun on Sat Oct 27, 2018 11:22 pm, edited 1 time in total.

Nay Lin Tun
Posts: 530
Joined: Mon Jan 16, 2012 5:34 am

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Nay Lin Tun » Sat Oct 27, 2018 11:20 pm

Here is vs Texel! g4!! There could be one more (someone said 4 games)

Javier Ros
Posts: 183
Joined: Fri Oct 12, 2012 10:48 am
Location: Seville (SPAIN)
Full name: Javier Ros

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Javier Ros » Sun Oct 28, 2018 4:09 am

After seeing the piece sacrifice of lc0 I think its level of play is not far from Alphazero's.

And another g4, this time 11.g4 against Stockfish 9 at 5 minutes plus 3 sec.

Stockfish 9 with three logical cores of the i7 4790S, which allows calculating about 3.86 million nodes per second against 3300 of lc0 (GTX 1060 3Gb), with a ratio of:
LeR = 3300 * 875/3860000 = 0.75
Stockfish 9 played with opening book perfect2018 without tablebases and lc0 11261 without opening book but with syzygy of 6 pieces.

The love relationship between a chess engine tester and his computer can be summarized in one sentence:
Until heat do us part.

Javier Ros
Posts: 183
Joined: Fri Oct 12, 2012 10:48 am
Location: Seville (SPAIN)
Full name: Javier Ros

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Javier Ros » Sun Oct 28, 2018 4:22 am

In the following game against Stockfish 9, sacrifice of pawn followed by sacrifice of queen by lc0 11261, queen by rook and bishop, which has as compensation the devastating entry into action of rooks and pair of bishops.

Played at 25 minutes plus 10 secs, same hardware as above message LeR = 3300 * 875/3860000 = 0.75.

The love relationship between a chess engine tester and his computer can be summarized in one sentence:
Until heat do us part.

Javier Ros
Posts: 183
Joined: Fri Oct 12, 2012 10:48 am
Location: Seville (SPAIN)
Full name: Javier Ros

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by Javier Ros » Sun Oct 28, 2018 10:12 am

In the match AlphaZero-Stockfish 8, several positions of zugzwang appeared where lc0 paralyzed the rival.

The following zugzwang position has appeared in one of my tests. Played at 25 minutes plus 10 secs, same hardware as above messages but 6 logical cores LeR = 3300 * 875/6640000 = 0.43.



White moves.

The love relationship between a chess engine tester and his computer can be summarized in one sentence:
Until heat do us part.

MikeB
Posts: 3559
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: One piece sacrifice by Leela at move 12. (she gains material back in 50 ply later)

Post by MikeB » Sun Oct 28, 2018 6:43 pm

Nay Lin Tun wrote:
Sat Oct 27, 2018 5:01 pm



AB engines do pseudo sacrifice, it means they always see material or other advantage back within their depth. However , Leela play with risks(winning probability), I am pretty sure that her MCTS wount see she would gain back a piece in next 25 moves, depth 50.
That really is an impressive game by Lc0.

Post Reply