Man vs machine - how to beat chess engine

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

Moderators: hgm, Rebel, chrisw

zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Man vs machine - how to beat chess engine

Post by zullil »

Uri wrote:
zullil wrote:Even if you made each of your moves without computer assistance, there is the question of whether the engine's moves were actually made by the engine, and without the use of settings or modifications that might have weakened the engine.
Yes I analyzed the games and the engines really made the same positional mistakes so I have good reason to believe that his victories over these engines were really honest.

What I think this again proves that humans really are superior and much smarter than chess engines in their strategical understanding of the game. And chess engines really do not understand chess strategy.
Possible, I suppose. But the probability seems vanishingly small, especially at game-in-five-minutes.

I'd like to look at the moves supposedly made by an engine and see if those moves are reproducible by the engine.
User avatar
Dr.Wael Deeb
Posts: 9773
Joined: Wed Mar 08, 2006 8:44 pm
Location: Amman,Jordan

Re: Man vs machine - how to beat chess engine

Post by Dr.Wael Deeb »

zullil wrote:
Uri wrote:
zullil wrote:Even if you made each of your moves without computer assistance, there is the question of whether the engine's moves were actually made by the engine, and without the use of settings or modifications that might have weakened the engine.
Yes I analyzed the games and the engines really made the same positional mistakes so I have good reason to believe that his victories over these engines were really honest.

What I think this again proves that humans really are superior and much smarter than chess engines in their strategical understanding of the game. And chess engines really do not understand chess strategy.
Possible, I suppose. But the probability seems vanishingly small, especially at game-in-five-minutes.

I'd like to look at the moves supposedly made by an engine and see if those moves are reproducible by the engine.
Again +1....
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Man vs machine - how to beat chess engine

Post by zullil »

zullil wrote:
Uri wrote:
zullil wrote:Even if you made each of your moves without computer assistance, there is the question of whether the engine's moves were actually made by the engine, and without the use of settings or modifications that might have weakened the engine.
Yes I analyzed the games and the engines really made the same positional mistakes so I have good reason to believe that his victories over these engines were really honest.

What I think this again proves that humans really are superior and much smarter than chess engines in their strategical understanding of the game. And chess engines really do not understand chess strategy.
Possible, I suppose. But the probability seems vanishingly small, especially at game-in-five-minutes.

I'd like to look at the moves supposedly made by an engine and see if those moves are reproducible by the engine.
Here's just one example, from this game:

Code: Select all

[Event "5 min. game"] 
[Site "Microsoft"] 
[Date "2012.05.03"] 
[Round "?"] 
[White "Kvicala, Miroslav"] 
[Black "Critter 1.4 32-bit"] 
[Result "1-0"] 
[ECO "D05"] 
[WhiteElo "2028"] 
[PlyCount "75"] 

1. d4 d5 2. Nf3 Nf6 3. e3 e6 4. Bd3 Be7 5. O-O c5 6. c3 Nc6 7. Nbd2 c4 8. Bb1 
b5 9. a3 O-O 10. e4 Nd7 11. Re1 Re8 12. e5 Bb7 13. Nf1 Nf8 14. g4 a5 15. Ng3 a4 
16. g5 Na5 17. Bc2 Nb3 18. Bxb3 axb3 19. Bd2 Qd7 20. Qe2 Qc8 21. Nh5 Ng6 22. 
Nf6+ gxf6 23. gxf6 Bf8 24. Ng5 h6 25. Qh5 hxg5 26. Bxg5 Be7 27. Bh6 Bc5 28. Re3 
Re7 29. Rg3 Rd7 30. Kh1 Rxa3 31. bxa3 Bf8 32. Rag1 Bg7 33. Bxg7 Qf8 34. Bxf8 b4 
35. Rxg6+ fxg6 36. Rxg6+ Kxf8 37. Qh8+ Kf7 38. Rg7# 1-0 
[D]r1bqk2r/pp2bppp/2n1pn2/2pp4/3P4/2PBPN2/PP1N1PPP/R1BQ1RK1 b kq - 2 7

In the game, Critter supposedly played c4.

Here's what I see from Critter:

Code: Select all

Critter v1.4 64-bit SSE4, by Richard Vida
setoption name Threads value 3
setoption name Hash value 256
setboard r1bqk2r/pp2bppp/2n1pn2/2pp4/3P4/2PBPN2/PP1N1PPP/R1BQ1RK1 b kq - 2 7
info string Opening book: "book.cbk" - 57853 entries [native]
info string GTB Init OK (5 piece set)
info string Using 3 thread(s), 256 MB hash
go infinite
 2/ 2  00:00       163  163000  -0.09 O-O Ng5 Bd7 Ndf3
 3/ 4  00:00       824  824000  +0.07 O-O Ne5 Nxe5 dxe5
 4/ 6  00:00      3195 1065000  +0.12 O-O b3 cxd4 exd4 Qc7
 5/ 6  00:00      3976  994000  +0.10 O-O b3 cxd4 exd4 Bd6 h3
 6/ 8  00:00      5203 1040600  +0.07 O-O b3 cxd4 exd4 Bd6 Ng5 h6
 7/10  00:00      9673 1209125  +0.19 O-O b3 Bd7 Bb2 cxd4 exd4 Qc7 Ng5 h6
 8/11  00:00     14032 1403200  +0.19 O-O b3 Bd7 Bb2 cxd4 exd4 Qc7 Ng5 h6
 9/13  00:00     30713 1616473  +0.16 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 Qc7
10/18  00:00     71400 2040000  +0.19 O-O b3 Bd7 Bb2 cxd4 exd4 Qc7 Qe2 h6 Ne5 Nxe5 Qxe5 Qxe5 dxe5
11/18- 00:00    237068 2301631  +0.11 O-O b3 Bd7 Bb2 cxd4 exd4 Qc7 Qe2 h6 Ne5 Nxe5 Qxe5 Qxe5 dxe5
11/24  00:00    261804 2316849  +0.09 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Ba3 Bd6 Bxd6 Qxd6 Nf3 Bd7 Qc2 h6 a4 Rac8 h3 b6
12/24  00:00    392057 2435136  +0.06 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 Nf3 Bd6 bxa6 Qc7 Qc2 bxa6 Rab1
13/28  00:00    646388 2544834  +0.06 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 Nf3 Bd6 bxa6 Qc7 Qc2 e5 c4 e4
14/28  00:00   1382197 2737023  +0.05 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 bxa6 bxa6 Nf3 Bd6 c4 dxc4 Bxc4 Bb7 Bd3 Rc8 Rc1 Rxc1 Qxc1
15/30  00:00   2424998 3031247  +0.05 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 bxa6 bxa6 Nf3 Bd6 c4 dxc4 Bxc4 Bb7 Bd3 Rc8 Rc1 Rxc1 Qxc1
16/32  00:01   5523207 3343345  +0.03 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 bxa6 bxa6 Nf3 Bd6 c4 Rb8 Qc2 dxc4 Bxc4 Qc7 Rab1 Bb7 Bxf6 gxf6
17/34  00:03  11143535 3546637  +0.03 O-O dxc5 Bxc5 b4 Bd6 b5 Ne5 Nxe5 Bxe5 Bb2 a6 bxa6 bxa6 Nf3 Bd6 c4 Rb8 Qc2 dxc4 Bxc4 Qc7 Rab1 Bb7 Bxf6 gxf6
18/38  00:14  54912711 3816032  +0.00 b6 Ne5 Nxe5 dxe5 Nd7 f4 c4 Bc2 Qc7 Qh5 g6 Qh6 Bf8 Qg5 Be7 Qh6
19/38  00:21  82786139 3851234  +0.00 b6 Ne5 Nxe5 dxe5 Nd7 f4 c4 Bc2 Qc7 Qh5 g6 Qh6 Bf8 Qg5 Be7 Qh6
20/39  00:29 113938701 3896938  +0.00 b6 Ne5 Nxe5 dxe5 Nd7 f4 c4 Bc2 Qc7 Qh5 g6 Qh6 Bf8 Qg5 Be7 Qh6
21/44  01:04 258591214 3998565  -0.01 b6 Ne5 Nxe5 dxe5 Nd7 f4 c4 Bc2 Qc7 Qh5 g6 Qe2 O-O e4 Bb7 exd5 Bxd5 Nf3 Rad8 Be3 Nc5 Rad1 b5 Bf2 Qc6
22/50  03:22 820339440 4056327  -0.03 b6 e4 O-O e5 Nd7 Re1 Bb7 a3 Rc8 b4 Qc7 Qe2 a6 Nf1 h6 bxc5 bxc5 Rb1 c4 Bc2 Rb8 Ne3 Ba8
Robert Flesher
Posts: 1280
Joined: Tue Aug 18, 2009 3:06 am

Re: Man vs machine - and a bridge for sale

Post by Robert Flesher »

Dr.Wael Deeb wrote:
styx wrote:it's hard for me to believe, that you are able to defeat (or even draw) a 5 minutes game against a top engine - even with anti-engine-preparation. i'm really sorry about that. i want to believe it, but it's simply so unlikely. i would still be suspicious if you are a 2500+ elo player.

i've seen several postings of some games, where someone claimed to beat houdini. i'm still waiting for an official event where someone can replicate it without taking back a move.

i know that even the best engines sometimes can't find the right move in time, but even the ones they find are usually not the kind of moves, that lead to instant losses.
+1

+2
acase
Posts: 983
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Man vs machine - how to beat chess engine

Post by acase »

Kvicala,


I believe that you cherry-picked a few games that you won and left out the hundreds (if not thousands) that you lost. Also there is no proof that you didn't take back any moves either.
User avatar
M ANSARI
Posts: 3707
Joined: Thu Mar 16, 2006 7:10 pm

Re: Man vs machine - how to beat chess engine

Post by M ANSARI »

To be taken seriously, you unfortunately have to do something like go on Playchess and play an engine online at 3 0 or 5 0. Beating an engine at home where you might have the luxury of taking back moves or even have the engine analysis window open needs to be eliminated as a possibility. The reason for all the doubt is nothing personal, but very well known players such as Carlsen, Kramnik and Nakamura ... as well as countless other super strong chess professionals, can't do what you seem to be able to do so easily.
Uri
Posts: 473
Joined: Thu Dec 27, 2007 9:34 pm

Re: Man vs machine - how to beat chess engine

Post by Uri »

M ANSARI wrote:The reason for all the doubt is nothing personal, but very well known players such as Carlsen, Kramnik and Nakamura ... as well as countless other super strong chess professionals, can't do what you seem to be able to do so easily.
I'm sure that Nakamura could find a weakness or a limit in the program to get a win.

For example, I found out that the best chess engines still do not quite understand the king safety factor very well and often tend to underestimate attack on the king and I'm sure that Nakamura, being very experienced against chess engines, could exploit this weakness against the program and get a winning position.
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Man vs machine - how to beat chess engine

Post by overlord »

For sure, I lost most of games...it is clear but why to post games which engine won? There are no take backs. I will give one more example...it is 15 min. game (played on smartphone) and you can check it. There are so many human moves that any engine would never make...just see the game...

[Event "15 min. game"]
[Site "Microsoft"]
[Date "2012.06.07"]
[Round "?"]
[White "Kvicala, Miroslav"]
[Black "Komodo 3"]
[Result "1-0"]
[WhiteElo "2035"]

1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Qb6 6. a3 c4 7. Be2 Nge7 8. O-O Bd7
9. Qc2 Na5 10. Nbd2 Nf5 11. Rb1 Be7 12. Re1 h6 13. Nf1 Qb3 14. Qd2 O-O 15. Bd1
Qb6 16. Bc2 Nb3 17. Qd1 Rfe8 18. Ne3 Nxe3 19. Bxe3 Rec8 20. g4 Qd8 21. Kg2 b5
22. Rh1 a6 23. h4 Be8 24. Nd2 Nxd2 25. Qxd2 Bf8 26. h5 Bc6 27. g5 hxg5 28. Bxg5
Be7 29. Be3 a5 30. Qe2 b4 31. axb4 axb4 32. Qg4 Qe8 33. Bh6 Qf8 34. Rbg1 Kh8
35. Bc1 Rcb8 36. Bg6 Qe8 37. h6 fxg6 38. hxg7+ Kxg7 39. Qxe6 Bg5 40. Qh3 Qh8
41. Qg3 Be7 42. e6 Rc8 43. Bh6+ Qxh6 44. Rxh6 Be8 45. Kf3 g5 46. Rgh1 Bh5+ 47.
R1xh5 Rf8+ 48. Ke3 Rf4 49. Rh7+ Kg8 50. Rxg5+ Bxg5 51. Qxg5+ Kxh7 52. Qxf4 bxc3
53. Qf7+ Kh8 54. bxc3 Ra2 55. e7 Re2+ 56. Kf3 Rxf2+ 57. Kg4 Rxf7 58. e8=Q+ Kg7
59. Qe5+ Kg6 60. Qxd5 Rg7 61. Qg5+ Kf7 62. Qxg7+ Kxg7 63. Kf5 Kf7 64. Ke5 Ke7
65. Kd5 Kd7 66. Kxc4 Ke6 67. d5+ Kd6 68. Kd4 Ke7 69. Kc5 Kf7 70. Kc6 Kf6 71.
Kc7 Ke5 72. d6 Ke6 73. d7 Kd5 74. d8=Q+ Ke5 75. Qd4+ Kf5 76. Kd6 Kg6 77. Qe5
Kh6 78. Qg3 Kh5 79. Ke6 Kh6 80. Kf6 Kh5 81. Qg5# 1-0
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Man vs machine - how to beat chess engine

Post by overlord »

Just check this positions more times. You will find that at the same depth engines doesn´t play always the same move...
rbarreira
Posts: 900
Joined: Tue Apr 27, 2010 3:48 pm

Re: Man vs machine - how to beat chess engine

Post by rbarreira »

overlord wrote:For sure, I lost most of games...
What would you say is the ratio of wins/loss/draws you got?