Nemorino-DeusX, TCEC 13

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

Moderators: hgm, Rebel, chrisw

jdart
Posts: 4367
Joined: Fri Mar 10, 2006 5:23 am
Location: http://www.arasanchess.org

Nemorino-DeusX, TCEC 13

Post by jdart »

DeusX went wrong right out of book with 4. .. Nxe4 and followed it up with 5. .. d5? It never recovered:

[pgn][Event "TCEC Season 13 - Division 3"] [Site "http://tcec.chessdom.com"] [Date "2018.08.15"] [Round "19.4"] [White "Nemorino 5.01"] [Black "DeusX 1.0"] [Result "1-0"] [ECO "C42"] [WhiteElo "3104"] [BlackElo "3200"] [PlyCount "74"] [EventDate "2018.??.??"] 1. e4 e5 2. Nf3 Nf6 3. Nxe5 Nxe4 4. Qe2 d5 5. d3 Bb4+ 6. c3 Bd6 7. Nf3 f5 8. Bg5 Qd7 9. dxe4 fxe4 10. Nd4 O-O 11. Nd2 Nc6 12. Nxc6 bxc6 13. O-O-O Rb8 14. Nb3 a5 15. Be3 a4 16. Nc5 Qe7 17. Nxa4 Be6 18. Qc2 c5 19. b3 c6 20. Be2 Ra8 21. Rd2 Rfb8 22. Kd1 c4 23. bxc4 c5 24. Ke1 Bd7 25. Nb2 Rxa2 26. Qb1 Ra3 27. Qc1 d4 28. cxd4 cxd4 29. Rxd4 Bb4+ 30. Kf1 Bc3 31. Bg5 Qe5 32. Rxd7 Rxb2 33. Rd5 Qc7 34. Rd8+ Kf7 35. Qd1 Qc6 36. Rd7+ Kg8 37. Rd6 Qc5 1-0 [/pgn]
Top
jdart
Posts: 4367
Joined: Fri Mar 10, 2006 5:23 am
Location: http://www.arasanchess.org

Re: Nemorino-DeusX, TCEC 13

Post by jdart »

It is worth noting, Bobcat also stumbled badly playing the Scandinavian (1. e4 d5 2. exd5 Qxd5 3. Nc3 Qe6+?) and lost.

--Jon
Top
tmokonen
Posts: 1296
Joined: Sun Mar 12, 2006 6:46 pm
Location: Kelowna
Full name: Tony Mokonen

Re: Nemorino-DeusX, TCEC 13

Post by tmokonen »

How does a program with an NN trained on master level games choose 3... Nxe4? Help this idiot understand.
Top
jp
Posts: 1470
Joined: Mon Apr 23, 2018 7:54 am

Re: Nemorino-DeusX, TCEC 13

Post by jp »

DeusX just a while ago made a shocking blunder on move 42 against Ethereal.
Top
MikeGL
Posts: 1010
Joined: Thu Sep 01, 2011 2:49 pm

Re: Nemorino-DeusX, TCEC 13

Post by MikeGL »

tmokonen wrote: ↑Thu Aug 16, 2018 6:33 am How does a program with an NN trained on master level games choose 3... Nxe4? Help this idiot understand.

I just checked with latest dev of SF9, but 3...Nxe4 always lands on
top 5 best move with only 0.10 difference with the other best moves.
So I think 3...Nxe4 is not that bad, but that game above looks amazing for
black with lots of gambits and free pieces given by NN in exchange for piece activity of black.

Edit: Sorry, didn't notice the score it was 1-0. White won against an NN engine. Too sleepy, I think i need some sleep now.
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.
Top
jp
Posts: 1470
Joined: Mon Apr 23, 2018 7:54 am

Re: Nemorino-DeusX, TCEC 13

Post by jp »

tmokonen wrote: ↑Thu Aug 16, 2018 6:33 am How does a program with an NN trained on master level games choose 3... Nxe4? Help this idiot understand.
No excuses for DeusX. But no master games have 3... Nxe4? so it doesn't see it punished. That may be part of the explanation.
Top
jdart
Posts: 4367
Joined: Fri Mar 10, 2006 5:23 am
Location: http://www.arasanchess.org

Re: Nemorino-DeusX, TCEC 13

Post by jdart »

MikeGL wrote: ↑Thu Aug 16, 2018 7:33 am
I just checked with latest dev of SF9, but 3...Nxe4 always lands on
top 5 best move with only 0.10 difference with the other best moves.
3.. Nxe4 maybe isn't so bad (but human practice does not encourage it). However if Black is going to play that it should not be followed up with .. d5. I think that is a second, serious error.

--Jon
Top

Return to “Computer Chess Club: General Topics”