Page 1 of 3
playing like a human should do it
Posted: Fri Sep 04, 2020 1:10 pm
by mclane
[pgn]
[Event "Xeon5690"]
[Site "DESKTOP-UL2RIJT"]
[Date "2020.09.04"]
[Round "1"]
[White "Deep Rybka 4.1 x64 [t6]"]
[Black "Lc0 0.26.2 cuda RTX2070 [j92-100]"]
[WhiteElo "2200"]
[BlackElo "2200"]
[ECO "C41"]
[Result "0-1"]
1. e4 c6 2. d4 h6 3. Nf3 g5 4. h3 d6 5. Bd3 Nf6 6. O-O Bg7
7. Nc3 g4 {-0.52/10 4} 8. hxg4 {+0.57/15 12} Bxg4 {-0.52/12
0} 9. Be2 {+0.54/15 11} Nbd7 {-0.48/11 13} 10. a4 {+0.49/15
10} Qa5 {-0.43/12 6} 11. Qd3 {+0.49/13 0} Rg8 {-0.34/9 11}
12. b4 {+0.37/15 23} Qh5 {-0.21/10 7} 13. Rb1 {+0.29/14 0}
e5 {-0.18/10 14} 14. dxe5 {+0.56/14 9} Nxe5 {-0.25/17 11}
15. Nxe5 {+0.41/17 0} dxe5 {-0.24/15 0} 16. Rd1 {+0.50/16
8} Bf8 {-0.02/16 8} 17. Nb5 {+0.57/14 9} Be7 {+0.61/17 8}
18. Nc7+ {+1.68/15 23} Kf8 {+0.69/1 0} 19. Bxg4 {+1.68/15
0} Rxg4 {+0.23/25 11} 20. Nxa8 {0.00/12 4} Rh4 {+0.11/13 0}
21. f3 {0.00/15 6} Rh2 {+0.05/27 5} 22. Be3 {0.00/15 0} Qh4
{+0.82/14 15} 23. Kf1 {0.00/16 0} Nh5 {+1.02/15 6} 24. Bf2
{0.00/15 0} Qg5 {+3.80/13 14} 25. g4 {-2.74/14 2} Nf4
{+4.35/13 6} 26. Qd2 {-4.42/14 1} h5 {+5.44/11 13} 27. Bg3
{-3.49/10 7} Rxd2 {+16.39/7 10} 28. Rxd2 {-4.60/14 0 Weiss
gibt auf} 0-1
[/pgn]
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 1:30 pm
by towforce
Impressive! Looks nothing like a computer game to me!
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 3:00 pm
by mclane
Yes, the machine who taught itself how to do chess plays more human then any human effort to produce a chess engine,
This is indeed disapointing and shows the situation we have,
The normal AB engines create ELO, but they play a machine chess.
The only engine that plays human alike chess on this level is in the moment LC0.
This is IMO disapointing for the human programmers,
Only a few concentrated on playing style and knowledge,
David Broughton, Thomas nitsche, Ed Schröder, richard Lang, Marty hirsch, Mark uniacke, Steen Surballe (Gandalf), Chris Whittington, christophe theron (gambit tiger)
and many many many others concentrated mainly on ELO and how to beat the opponent.
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 3:27 pm
by JohnW
Can you give an example of a move(s) in this game that only a human would play? Not trying to argue, I just can't tell the difference.
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 3:35 pm
by Dr.Wael Deeb
JohnW wrote: ↑Fri Sep 04, 2020 3:27 pm
Can you give an example of a move(s) in this game that only a human would play? Not trying to argue, I just can't tell the difference.
Hard to estimate ....
Which human exactly as they also differ much in style and tactical abilities ....
Cheers
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 3:38 pm
by chrisw
JohnW wrote: ↑Fri Sep 04, 2020 3:27 pm
Can you give an example of a move(s) in this game that only a human would play? Not trying to argue, I just can't tell the difference.
Conventionally it would be only some moves and only some humans. Moves that deliberately take the game into uncharted territory where outcomes become difficult to calculate. No, I won’t define uncharted nor deliberately nor difficult to calculate.
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 4:48 pm
by mclane
Forgot Julio kaplan, and I sure many others. All those programmers who tried to make computerchess not only successful but also on the base of knowledge and human alike.
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 9:37 pm
by MikeGL
Beautiful game. Thanks for sharing.
It is a very deep tactic in my opinion starting with Bf8! provoking Nb5 -> Nc7 fork and winning material for AB, but Lc0 saw that N fork too hence more tactical.
I remember Rybka 4 as a tactical beast during its heyday, even winning almost all titles in prestigious engine tourneys in a row. But not now.
Modern engines, NN specifically, already made a huge paradigm shift on this field even getting criticisms and resistance from majority of
chess enthusiasts on this forum calling it (A0) a hoax, scam, etc. I remember very few members here defended Googles DeepMinds project, A0, as legit.
NN is just too different from hardware being used (GPU) up to the software side. Rybka is too old and not updated but if its recoded to parse NNUE it will
surely be boosted too just like the rest.
Chess is 99% tactics -Richard Teichmann
I think I agree with that GM.
mclane wrote: ↑Fri Sep 04, 2020 1:10 pm
[pgn]
[Event "Xeon5690"]
[Site "DESKTOP-UL2RIJT"]
[Date "2020.09.04"]
[Round "1"]
[White "Deep Rybka 4.1 x64 [t6]"]
[Black "Lc0 0.26.2 cuda RTX2070 [j92-100]"]
[WhiteElo "2200"]
[BlackElo "2200"]
[ECO "C41"]
[Result "0-1"]
1. e4 c6 2. d4 h6 3. Nf3 g5 4. h3 d6 5. Bd3 Nf6 6. O-O Bg7
7. Nc3 g4 {-0.52/10 4} 8. hxg4 {+0.57/15 12} Bxg4 {-0.52/12
0} 9. Be2 {+0.54/15 11} Nbd7 {-0.48/11 13} 10. a4 {+0.49/15
10} Qa5 {-0.43/12 6} 11. Qd3 {+0.49/13 0} Rg8 {-0.34/9 11}
12. b4 {+0.37/15 23} Qh5 {-0.21/10 7} 13. Rb1 {+0.29/14 0}
e5 {-0.18/10 14} 14. dxe5 {+0.56/14 9} Nxe5 {-0.25/17 11}
15. Nxe5 {+0.41/17 0} dxe5 {-0.24/15 0} 16. Rd1 {+0.50/16
8} Bf8 {-0.02/16 8} 17. Nb5 {+0.57/14 9} Be7 {+0.61/17 8}
18. Nc7+ {+1.68/15 23} Kf8 {+0.69/1 0} 19. Bxg4 {+1.68/15
0} Rxg4 {+0.23/25 11} 20. Nxa8 {0.00/12 4} Rh4 {+0.11/13 0}
21. f3 {0.00/15 6} Rh2 {+0.05/27 5} 22. Be3 {0.00/15 0} Qh4
{+0.82/14 15} 23. Kf1 {0.00/16 0} Nh5 {+1.02/15 6} 24. Bf2
{0.00/15 0} Qg5 {+3.80/13 14} 25. g4 {-2.74/14 2} Nf4
{+4.35/13 6} 26. Qd2 {-4.42/14 1} h5 {+5.44/11 13} 27. Bg3
{-3.49/10 7} Rxd2 {+16.39/7 10} 28. Rxd2 {-4.60/14 0 Weiss
gibt auf} 0-1
[/pgn]
Re: playing like a human should do it
Posted: Fri Sep 04, 2020 11:21 pm
by syzygy
mclane wrote: ↑Fri Sep 04, 2020 3:00 pm
Yes, the machine who taught itself how to do chess plays more human then any human effort to produce a chess engine,
This is indeed disapointing and shows the situation we have,
Well, who taught the machine how to teach itself to play chess?
Re: playing like a human should do it
Posted: Sat Sep 05, 2020 3:34 am
by zullil
MikeGL wrote: ↑Fri Sep 04, 2020 9:37 pm
Beautiful game. Thanks for sharing.
It is a very deep tactic in my opinion starting with Bf8! provoking Nb5 -> Nc7 fork
A/B engines see Bf8 almost instantly, even without NNUE.
All I see in this game is a bad blunder (Nb5) by an old engine. That move is the most human-like of all ... a mistake.
