Page 19 of 19

Re: This is where top Engines fail compared to human........

Posted: Sat May 30, 2015 11:27 pm
by zullil
...where the Salov vs. Korchnoi, 1997 example is wrong, as proved in this thread!

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 8:40 am
by BBauer
This is no fortress, as has been shown
[D]8/5p2/4k2p/7R/6P1/5PK1/3q4/8 b - - 21 58
but after 58. ...f6 played by Korchnoi the defence on the 5.th rank is possible,
so this seams to be a fortresss
[D]r8/8/4kp1p/5R2/6P1/5PK1/3q4/8 b - - 1 59

Can you confirm?
Kind regards
Bernhard

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 8:42 am
by BBauer
Should be
[D]8/8/4kp1p/5R2/6P1/5PK1/3q4/8 b - - 1 59

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 10:06 am
by AdminX
BBauer wrote:Should be
[D]8/8/4kp1p/5R2/6P1/5PK1/3q4/8 b - - 1 59

8/8/4kp1p/5R2/6P1/5PK1/3q4/8 b - - 0 1

Analysis by Komodo 9.02 64-bit:

59...Qc1 60.Rh5 Qg1+ 61.Kf4 Qe1 62.Rb5 Qc1+ 63.Kg3 Qg1+ 64.Kf4 Qd4+ 65.Kg3 Qd1 66.Rh5 Qa1 67.Kf4 Qb1 68.Ke3 Qb8 69.Ke4 Qb4+ 70.Ke3 Qe1+ 71.Kf4 Kd6 72.Rb5 Qd2+ 73.Kg3 Qd3 74.Ra5 Ke6 75.Rh5 Qa6 76.Kg2 Qc6 77.Rh1 Qd5 78.Rh5 Qd2+ 79.Kg3 Qd1 80.Kf4 Qd6+ 81.Ke3 Qb6+ 82.Kf4 Qf2 83.Rb5 Qd2+ 84.Kg3 Qe3 85.Rb1 Kd6 86.Rb8 Qg1+ 87.Kf4 Kc6 88.Rc8+ Kd7
-+ (-250.00) Depth: 46 00:03:57 2394MN, tb=68116562

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 10:45 am
by whereagles
zullil wrote:
...where the Salov vs. Korchnoi, 1997 example is wrong, as proved in this thread!
Well, login to wiki and fix it :)

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 10:59 am
by BBauer
Thank you for your interest.
[D]1q6/8/4kp1p/7R/6P1/4KP2/8/8 w - -
In your line komodo played 69.Ke4 which was not forced.
SF proposes Kf2 which should hold the draw.
Kind regards
Bernhard

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 11:56 am
by zullil
BBauer wrote:Should be
[D]8/8/4kp1p/5R2/6P1/5PK1/3q4/8 b - - 1 59
Stockfish agrees with Ted's Komodo. No fortress:

Code: Select all

info depth 61 seldepth 109 multipv 1 score cp 12334 nodes 39613860187 nps 59365372 hashfull 359 tbhits 54489857 time 667289 pv d2d4 g3g2 d4b2 g2g3 b2b1 f5h5 b1g1 g3f4 e6d7 h5d5 d7c6 d5h5 g1f2 h5f5 f2h2 f4e3 c6d6 f5b5 d6e6 b5a5 h2g1 e3e2 g1b6 a5f5 b6a7 f5h5 a7d4 h5f5 d4c4 e2e3 c4c1 e3e4 c1d2 f5h5 e6e7 h5f5 e7f7 f5h5 f7g6 h5d5 d2e2 e4f4 e2c4 f4e3 c4d5 f3f4 d5e6 e3f3 e6g4

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 12:08 pm
by zullil
BBauer wrote:Thank you for your interest.
[D]1q6/8/4kp1p/7R/6P1/4KP2/8/8 w - -
In your line komodo played 69.Ke4 which was not forced.
SF proposes Kf2 which should hold the draw.
Kind regards
Bernhard
No.

Code: Select all

info depth 50 seldepth 75 multipv 1 score cp 12301 nodes 5828382568 nps 62384348 hashfull 5 tbhits 312394 time 93427 pv b8a7 f2g2 a7b6 h5f5 b6b2 g2g3 b2c1 f5h5 c1g1 g3f4 g1f2 h5b5 f2d4 f4g3 d4g1 g3f4 g1a7 f4g3 e6d6 g3g2 a7a6 b5h5 a6b7 h5a5 b7b4 a5f5 b4b2 g2g3 b2a1 g3g2 a1a2 g2g3 a2e2 f5h5 e2e3 h5f5 e3e1 g3f4 e1d2 f4g3 d2c3 f5h5 c3a1 h5f5 d6c6 g3f2 a1b2 f2e3 c6d6 f5h5 b2b3 e3f2 b3a2 f2e3 a2b2 e3f4 b2d4 f4g3 d4g1 g3f4 g1c1 f4g3 c1e1 g3g2 e1d2 g2g3 d2e2

Re: This is where top Engines fail compared to human........

Posted: Sun May 31, 2015 12:30 pm
by BBauer
I tried the position again and got a 123.77 score.
So you are right. No fortress.
Sorry.
Bernhard