back


29913WCN
2020-02-20 20:53:10
Stockfish 11 Plays The (A.R.B.C.S) - Kings & Pawns Game - A.R.B :)


Below is Stockfish's game Who would have guessed 1. b4 as it's first move ? and Why ? I am sure most Humans would play 1. e4 or 1. d4 first ? and then march the King forward in the centre ?

I would love to see some Expert Analysis showing Stockfish's reasons for it's strange play!


https://groups.google.com/forum/?nomobile=true#!topic/fishcooking/mgz2-3XYI8Y





WCN
29912queen13
2020-02-20 19:14:33
by the way, if anyone missed the transition, the site that shows lc0 progress has been moved

Here:

https://training.lczero.org/
29911queen13
2020-02-19 17:23:25
What happened at stockfish - Lc0 in cccc?

I saw that many games have been drawn but now that I see again they have already moved on to a gpu tournament which brings me to the next question: Why Komodo participates in a gpu only tournament? Why not the other engines as well?
29910WCN
2020-02-18 19:42:46
Stockfish 11 Plays The (A.R.B.C.S) - Kings & Pawns Game - A.R.B :)

https://groups.google.com/forum/?nomobile=true#!topic/fishcooking/mgz2-3XYI8Y





WCN
29909queen13
2020-02-16 21:52:15
This has already happened before.
I don't know how accurate it is.
29908HomoSapiens
2020-02-16 13:41:20
https://www.chess.com/computer-chess-championship
CCC has officially estimate Lc0 a higher ELO than Stockfish. Holy shit!
29907Peter
2020-02-08 11:46:10
@stoeler1 (2020-02-01 09:21:17)
SF has to reach at least depth 31:
"nice -n -19 ./stockfish_x64_modern.exe bench 1024 16 31 Rxf5.epd depth > ~/Rxe5.log 2<&1"

results in (only the last few lines of logfile):

info depth 31 seldepth 50 multipv 1 score cp 1031 nodes 125060534 nps 31501393 hashfull 286 tbhits 0 time 3970 pv e5f5 e6f5 e2e8 h5h4 f2g2 g7h7 g2h3 h7g7 e8b8 g7h7 h3h4 h7g7 h4g3 g7h7 g3h3 h7g7 h3h4 d5d4 c3d4 g7h7 h4h3 h7g7 b8d8 g7h7 d8d6 h7g8 d4d5 g8f8 d5c6 a6c6 d6c6 f8e7 c6a6 e7d7 h3g3 d7e8 a6a7 f7f6 g5f6
bestmove e5f5 ponder e6f5

===========================
Total time (ms) : 3970
Nodes searched : 125061226
Nodes/second : 31501568
29906queen13
2020-02-06 10:59:10
I don't think that you need to add one for comparing stockfish 11 with more recent dev versions.

It's true that stockfish plays better with them though.

You can use google to find out how to download them, from a quick google search I found this site: https://syzygy-tables.info/

Which has some sites that you can download them from as well as instructions for how to best use them. (Make sure to keep the WDL files on an SSD if you can.)

Unfortunately, I can't tell what the matter is with the lower nps. I hope someone will get to the bottom of this.
29905Kenneth Lovering
2020-02-06 09:07:23
@Queen13
Thank you for your response of 1/30. All parameters are as required and/or as you suggested/assumed. The only thing is that you mentioned to make sure that all the versions are having access to the same endgame tablebases and with the same corresponding settings for them.
I don't have an endgame tablebase, didn't know you needed to add one! Where would I get that? Thanks
29904Mrert
2020-02-04 21:15:58
The first move out of the book was move 13 and was evaluated as 0.00
So the book played a very drawish opening, especially since SF is very optimistic in its base settings. 0.00 really means SF thinks it's completely drawish
Try without book.
29903queen13
2020-02-04 11:45:07
It was playing with a book. Are you sure this is a good book?

Try Brainfish or just bare stockfish.

Also, rarely the magnus App might get a draw once in a while...
29902Mediocre
2020-02-04 02:26:51
The latest development of Stockfish almost used up its time running on a PC while the Magnus Carlsen app played moves almost instantly. All odds were in favor of Stockfish but it never capitalized and played for draw. Latest Stockfish had everything, power, RAM, endgame tablebases, latest opening book but it only managed a draw on an iphone app? This is very disappointing folks.

[Event "Blitz 485min+32sec-5min+2sec"]
[White "Stockfish 020220 64 BMI2"]
[Black "Play Magnus App iphone"]
[Result "1/2-1/2"]
[PlyCount "153"]
[TimeControl "300+2"]

{512MB, Hiarcs14zgBook.ctg, Core i5, 5&6 pc EGTB} 1. e4 {B 0} c5 {5} 2. Nf3 {B 0}
Nc6 {4} 3. Bb5 {B 0} g6 {4} 4. O-O {B 0} Bg7 {4} 5. c3 {B 0} Nf6 {4} 6. Qe2 {
B 0} O-O {5} 7. d4 {B 0} cxd4 {4} 8. cxd4 {B 0} d5 {4} 9. e5 {B 0} Ne4 {4} 10.
Be3 {B 0} f6 {5} 11. exf6 {B 0} exf6 {4} 12. Nc3 {B 0} Be6 {7} 13. Rfe1 {
0.00/28 27} Re8 {11} 14. Bxc6 {0.07/28 0} bxc6 {8} 15. Qc2 {0.00/30 0} Nxc3 {
8 (Rc8)} 16. Qxc3 {0.19/24 4} Qc7 {6 (Qb6)} 17. Nd2 {0.23/28 11} Bf8 {7 (Rac8)}
18. Nb3 {0.30/28 14} Bd6 {7} 19. h3 {0.38/26 2} Kg7 {8 (h5)} 20. Nc5 {0.31/29
17} Bf5 {8 (Bf7)} 21. Bd2 {0.24/26 10} h5 {8} 22. Qf3 {0.33/29 7} Qb6 {9 (Bh2+)
} 23. b4 {0.75/25 8} Qb5 {7 (Qc7)} 24. g4 {0.68/28 21} hxg4 {6} 25. hxg4 {
0.98/26 2} Be4 {6} 26. Nxe4 {0.81/26 4} dxe4 {6} 27. Rxe4 {0.85/28 0} Qd5 {11}
28. Re3 {0.97/28 0} Qxf3 {6} 29. Rxf3 {0.60/29 14} Re4 {8} 30. g5 {0.62/27 0}
fxg5 {10} 31. Ra3 {0.62/27 21} Kf7 {7 (Re2)} 32. Be3 {0.38/26 15} Bxb4 {13} 33.
Ra6 {0.20/28 8} Re6 {6} 34. Rd1 {0.13/29 7} Be7 {8} 35. Re1 {0.20/30 17} Bd8 {
7 (Bb4)} 36. a4 {0.56/24 11} Bb6 {8 (Bf6)} 37. Rd1 {0.45/24 12} Rd6 {7} 38. a5
{0.40/26 7} Bd8 {7 (Bc7)} 39. Kg2 {0.62/26 5} Ke6 {6} 40. Rc1 {0.26/26 5} Kd7 {
8} 41. Kh3 {0.26/27 1} Bf6 {7} 42. Rc5 {0.26/29 5} Be7 {7} 43. Kg3 {0.25/27 6}
Re6 {6} 44. Rc4 {0.25/28 0} Bd6+ {7 (Bf6)} 45. Kg4 {1.06/20 4} Be7 {7} 46. Kh3
{0.25/28 24} Bf6 {7} 47. Rc2 {0.25/30 0} Bg7 {7 (Rd6)} 48. Kg4 {0.25/31 11} Bf6
{5} 49. Rc1 {0.25/29 0} Rd6 {6 (Re4+)} 50. Kg3 {0.25/29 6} Bh8 {6 (Re6)} 51. f3
{0.45/21 3} Re6 {7 (Bf6)} 52. Bf2 {0.23/22 8} Rd6 {6 (Rf6)} 53. Rc5 {0.31/23 5}
Bf6 {7} 54. Be3 {0.11/26 14} Be7 {6} 55. Bf2 {0.00/26 0} Rf6 {6 (Bf6)} 56. Rc1
{0.00/25 7} Bd6+ {8 (c5)} 57. Kg2 {0.00/27 4} Bf4 {7} 58. Rc2 {0.00/28 0} Rd6 {
7 (Re6)} 59. Rc5 {0.00/25 3} Kc7 {5} 60. d5 {0.00/28 0} Rxd5 {7} 61. Raxc6+ {
0.00/28 0} Kd7 {5} 62. Rc7+ {0.00/29 0} Bxc7 {6} 63. Rxd5+ {0.00/30 0} Ke7 {
8 (Kc6)} 64. Rxg5 {0.00/23 5} a6 {7 (Kd7)} 65. Be1 {0.00/26 4} Kf6 {9 (Rg8)}
66. Rc5 {0.00/20 4} Bf4 {6 (Be5)} 67. Rc6+ {0.00/24 10} Kf5 {7} 68. Rc5+ {
0.00/27 0} Kf6 {5 (Ke6)} 69. Rc6+ {0.00/27 5} Kf5 {5} 70. Bc3 {0.00/30 3} Ra7 {
8} 71. Rf6+ {0.00/31 0} Kg5 {6} 72. Rf8 {0.00/31 0} Rd7 {8} 73. Bf6+ {0.00/30 0
} Kf5 {9} 74. Ba1+ {0.00/30 0} Kg5 {5} 75. Bf6+ {0.00/31 5} Kf5 {7} 76. Bd4+ {
0.00/32 0} Kg5 {6} 77. Bf6+ {0.00/32 7} 1/2-1/2
29901queen13
2020-02-01 10:13:24
Most likely this is how David knows, thanks
29900stoelerl
2020-02-01 09:21:17
Best move is re5f5, but every release of stockfish played e2a2?

[Event "?"]
[Site "?"]
[Date "2016.02.17"]
[Round "?"]
[White "Player"]
[Black "Player"]
[Result "1-0"]
[FEN "8/p4pk1/r1p1p1p1/P1PpRnPp/1P3P2/2P5/4RK2/8 w - - "]
[SetUp "1"]
29899Mrert
2020-01-31 22:17:01
Well logos (λόγος) as the expression of a rational discourse is quite an important concept in the history of the western world as greek philosophers are corner stones of most european societies.
29898queen13
2020-01-31 18:40:57
David
Do you know ancient greek? How do you know that intelligence = λόγος in ancient greek?
29897David
2020-01-31 08:42:15
Queen
Perhaps the word "conscience" is more appropriate if we want to define the difference between man and machine.
"AI" is an expression to be traced back to the concept of energy as a capacity for work, like speaking, choosing, etc.
Intelligence = λόγος ( ancient greek) = to say. to choose.
29896queen13
2020-01-30 20:35:47
@David

It doesn't matter as the statement is incorrect.

It's possible that humans will create a machine that appears to be smart when in reality it does no thinking and only appears to be thinking.

That wouldn't prove that humans are a machine or that intelligence doesn't exist.

It's a good idea to define terms...

A machine can be said to be smart if it can do very smart things.

But then that machine could be doing no real thinking and only giving the appearance of such.

So, is it a smart machine or is it dumb but just amazing?(Kind of like stockfish is for chess. Would you say stockfish is smart?)

I don't know what you think about this. But even though I know what smart means it would be impossible to know what you would think in this case...
29895queen13
2020-01-30 20:29:35
Ok, if I understand correctly you downloaded the DVs and stockfish 11 from here and used the "modern" versions.

I assume that you downloaded DVs from here after stockfish 11 and they are for some reason quite faster. (I saw some benches for DVs after sf11 and if I understand correctly some versions after stockfish 11 are slower but you say that stockfish 11 is slower...)

I assume that you compared them on the same position.

I don't know what the issue is, you have probably checked already what I am thinking...

Make sure you set the threads parameter the same as well as the hash memory.

Make sure that all the versions are having access to the same endgame tablebases and with the same corresponding settings for them.

I hope this helps and that a better explanation will be posted.
29894Kenneth Lovering
2020-01-30 18:26:35
I downloaded stockfish 11 from here
29893queen13
2020-01-30 12:06:31
It's the one that says for Haswell cpus here but if you download stockfish from
https://stockfishchess.org/download/
then you will get 4 executables.

Did you download stockfish 11 from here?
29892Kenneth Lovering
2020-01-30 02:05:02
@queen13
I get my DVs from http://abrok.eu/stockfish/
There is no option for the x64 bmi2 exe at that location.
29891queen13
2020-01-29 20:58:24
try the x64 bmi2 exe.
If it works it is better to use this one.
If it doesn't then you are using the correct version.

Where do you get the DVs from?
29890Kenneth Lovering
2020-01-29 14:39:55
@queen13
I use the Windows x64 for modern computers version. I don't know what you mean by BMI.
29889David
2020-01-29 12:40:20
:)queen13
Yeah, I have to to define what smart is first...
because you don't know it...

next page >
single page