Is Black's compensation after 16. g3 really sufficient?

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
AdminX
Posts: 5188
Joined: Mon Mar 13, 2006 1:34 pm
Location: Acworth, GA
Contact:

Is Black's compensation after 16. g3 really sufficient?

Post by AdminX » Tue Sep 24, 2019 8:51 am



The position is from the 'Man vs Machine 5th Game' played in 2003 Kasparov vs Deep Junior.

Image
Direct Link: https://i.ibb.co/64hfWYw/2019-09-24-4-23-46.jpg

I took a short look at this with Komodo MCTS, and think 16. g3 proves that Deep Junior's 10. ... Bxh2?! was dubious. Looking down the rabbit hole, I see positions like this.





Of course this is the famous position that started it all.

"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers

Dann Corbit
Posts: 10189
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Dann Corbit » Tue Sep 24, 2019 8:59 am

In did not analyze with a computer, but I like g3 in that position. What would you play, and why do you like it better? Maybe I am naiive but g3 looks like a pretty obvious good.move to me. I imagine a deep search might find something much better but it is very unlikely that I would find it otb.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

User avatar
AdminX
Posts: 5188
Joined: Mon Mar 13, 2006 1:34 pm
Location: Acworth, GA
Contact:

Re: Is Black's compensation after 16. g3 really sufficient?

Post by AdminX » Tue Sep 24, 2019 9:35 am

Dann Corbit wrote:
Tue Sep 24, 2019 8:59 am
In did not analyze with a computer, but I like g3 in that position. What would you play, and why do you like it better? Maybe I am naiive but g3 looks like a pretty obvious good.move to me. I imagine a deep search might find something much better but it is very unlikely that I would find it otb.
I just happened to be revisiting some old Junior games with Chessbase and was going over the notes to this game. I thought with advances in MCTS engines it would be interesting to go over this with Komodo or Leela. So I fired up Komodo MCTS and liked to improvements it found over the game notes. I would not have given it a go before as I found the line Gulko gave with 18. Ke1 Qh3 19. Nd1 left me feeling somewhat suspect. I think the key position that needs deeper thought may be here.



I would not play this with blitz time controls, but in LTC games I would now give it a go. In the game notes g3 was given as interesting, however I am of the opinion that it feels more like the only move to play for the win as whites king now seems to avoid the danger zones.
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers

User avatar
Nordlandia
Posts: 2476
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Nordlandia » Tue Sep 24, 2019 9:41 am

According latest stockfish dev at C=0. The sacrifice is dubious. +1.5 +/− edge for white.

i7-5960X 4.5GHz | 7-threads used.

User avatar
Eelco de Groot
Posts: 4168
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Eelco de Groot » Tue Sep 24, 2019 4:04 pm

Stockfish with a small change predicts 19.Nd1 and then 19...Nh2 even from move 11 after 2 minutes correctly predicting the lines after the Bishop sac as played in the match but deviating with 16.g3. So not so much 19.f5 from the diagram.





38/99 2:15 +1.36 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Bxh7+ Kh8 21.Rf2 Qh5 22.Bd3 Nf3+
23.Rxf3 Qxf3 24.Bc1 (204.334.744) 1502

39/61 2:25 +1.25-- 11.Kxh2 Ng4+ (217.472.092) 1499

39/61 2:30 +1.14-- 11.Kxh2 Ng4+ (225.027.068) 1497

39/63 2:37 +1.25++ 11.Kxh2 (236.409.626) 1498

39/74 3:06 +0.99-- 11.Kxh2 Ng4+ (277.531.971) 1490

39/74 3:26 +1.23++ 11.Kxh2 (307.279.343) 1485

39/74 3:43 +1.18 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (331.861.560) 1481

40/69 4:06 +1.29++ 11.Kxh2 (366.221.763) 1485

40/77 4:17 +1.40++ 11.Kxh2 (383.482.677) 1486

40/77 4:22 +1.39 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (389.743.851) 1486

41/66 4:42 +1.28-- 11.Kxh2 Ng4+ (418.530.005) 1483

41/66 4:46 +1.17-- 11.Kxh2 Ng4+ (424.874.460) 1483

41/70 5:04 +1.14 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (451.715.057) 1481

42/76 5:41 +1.24++ 11.Kxh2 (507.678.399) 1485

42/76 5:59 +1.18 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (533.034.215) 1480

43/80 6:56 +1.29++ 11.Kxh2 (618.907.331) 1486

43/80 7:18 +1.40++ 11.Kxh2 (652.154.999) 1488

43/80 7:25 +1.30 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (662.523.346) 1488

44/71 8:31 +1.41++ 11.Kxh2 (761.350.139) 1487

44/71 10:21 +1.28 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (925.010.200) 1488

45/80 13:00 +1.39++ 11.Kxh2 (1.161.324.689) 1488

45/80 13:57 +1.50++ 11.Kxh2 (1.247.590.155) 1488

45/80 14:14 +1.33 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (1.270.938.422) 1487

46/69 17:28 +1.23-- 11.Kxh2 Ng4+ (1.555.884.543) 1483

46/78 17:48 +1.33++ 11.Kxh2 (1.584.391.123) 1483

46/82 18:56 +1.49++ 11.Kxh2 (1.687.188.404) 1484

46/88 19:36 +1.40 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (1.746.094.053) 1484

47/73 22:53 +1.51++ 11.Kxh2 (2.040.046.973) 1485

47/82 23:12 +1.29-- 11.Kxh2 Ng4+ (2.069.080.588) 1485

47/84 23:19 +1.45++ 11.Kxh2 (2.078.821.690) 1485

47/84 27:05 +1.60 11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5
14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+
17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2
20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1
23.Kxg1 Bxf5 24.Bxf5 (2.417.651.915) 1486


After 19.f5, from the diagram in the post above, can follow, this is Kaissa not really Stockfish



Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

41 194:28 +1.43 1...Nd7 2.e4 dxe4 3.Nxe4 Nb6 4.Nf4 Qxg3+
5.Kd1 Ne3+ 6.Bxe3 Qxe3 7.Qe2 Nc4
8.Qf2 Bxf5 9.Nf6+ gxf6 10.Bxc4 Bg4+
11.Kc2 Rad8 12.Qxe3 Rxe3 13.Rg1 f5
14.Bd3 Kf8 (18.602.583.000) 1594

41 194:28 +1.44 1...Rxe3 2.Nxd5 Re8 3.Ndf4 Qxg3+
4.Kd1 Nf2+ 5.Rxf2 Qxf2 6.Kc1 Qf1+
7.Qd1 Qxd1+ 8.Kxd1 Na6 9.Kc2 Nc7
10.Rg1 Kh8 11.Nh5 Rxe2 12.Bxe2 Bxf5+
13.Kc1 Ne6 14.Be3 f6 (18.602.583.000) 1594

41 194:28 +1.85 1...Qg2 2.Nf4 Qxg3+ 3.Kd1 Nd7 4.Nce2 Nxe3+
5.Bxe3 Qxe3 6.Qd2 Nf6 7.Qxe3 Rxe3
8.Kd2 Re4 9.Bxe4 Nxe4+ 10.Ke1 Bxf5
11.Nxd5 Bh3 12.Rh1 Bg4 13.Ndc3 Re8
14.Nxe4 Rxe4 (18.602.583.000) 1594

41 194:28 +2.07 1...b6 2.e4 Nd7 3.Bf4 Nh2 4.Rf2 Ng4
5.e5 c5 6.Bb5 cxd4 7.Nxd4 Ba6 8.Bxa6 Ndxe5
9.Be2 Nxf2 10.Kxf2 Qh2+ 11.Ke3 Rac8
12.Bxe5 Qxg3+ 13.Kd2 Qxe5 14.Qd3 Qf4+ (18.602.583.000) 1594

41 194:28 +2.29 1...Na6 2.Nf4 Qxg3+ 3.Kd1 Nb8 4.Qc1 Nd7
5.Nce2 Nxe3+ 6.Bxe3 Qxe3 7.Qxe3 Rxe3
8.Kd2 Re7 9.Rg1 Nf6 10.Raf1 Bd7
11.Ng3 Rf8 12.Ngh5 Nxh5 13.Nxh5 f6
14.Rg3 Rff7 (18.602.583.000) 1594

But the scores are not really going up enough to say that White would be winning.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

User avatar
Eelco de Groot
Posts: 4168
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Eelco de Groot » Wed Sep 25, 2019 12:07 pm

I let the search continue a little bit. For Stockfish I think the main line after Deep Junior's B sac is still a draw. White about 1½ pawn up but this advantage isn't growing anymore. Of course it is possible that Stockfish misses something. I would like to test the small change to Stock but I can't use the repo anymore because it is on a Windows XP machine, which is no longer supported. It helps a tiny bit with Zugzwang and the Mate in 5. Maybe I can open a new repository but I think I would need a new account?




Engine: Stockfish with small change (512 MB)

Code: Select all

48/87  38:31  +1.51    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (3.439.099.092) 1488 

49/89  52:34  +1.61++  11.Kxh2 (4.727.839.060) 1498 

49/89  54:48  +1.40--  11.Kxh2 Ng4+ (4.931.981.905) 1499 

49/89  54:58  +1.56++  11.Kxh2 (4.947.329.112) 1499 

49/89  55:35  +1.61    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (5.002.453.787) 1499 

50/89  67:22  +1.50--  11.Kxh2 Ng4+ (6.068.373.495) 1500 

50/89  72:52  +1.39--  11.Kxh2 Ng4+ (6.574.211.040) 1503 

50/89  73:15  +1.50++  11.Kxh2 (6.607.349.081) 1503 

50/89  75:03  +1.45    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (6.767.767.089) 1502 

51/75  80:21  +1.34--  11.Kxh2 Ng4+ (7.237.503.722) 1501 

51/83  80:46  +1.45++  11.Kxh2 (7.274.492.566) 1501 

51/88  81:58  +1.60++  11.Kxh2 (7.381.185.057) 1500 

51/88  88:21  +1.67    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (7.966.099.510) 1502 

52/90  90:21  +1.56--  11.Kxh2 Ng4+ (8.146.883.863) 1502 

52/90  92:54  +1.45--  11.Kxh2 Ng4+ (8.377.542.068) 1502 

52/90  97:10  +1.56++  11.Kxh2 (8.772.569.591) 1504 

52/90  103:12 +1.30--  11.Kxh2 Ng4+ (9.313.728.887) 1504 

52/90  104:25 +1.53++  11.Kxh2 (9.420.140.876) 1503 

52/90  111:34 +1.53    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (10.065.529.117) 1503 

53/93  143:22 +1.42--  11.Kxh2 Ng4+ (12.948.786.472) 1505 

53/93  150:25 +1.31--  11.Kxh2 Ng4+ (13.568.863.601) 1503 

53/93  153:04 +1.42++  11.Kxh2 (13.808.200.435) 1503 

53/93  161:45 +1.45    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (14.570.997.864) 1501 

54/78  191:04 +1.56++  11.Kxh2 (17.249.382.210) 1504 

54/94  195:08 +1.34--  11.Kxh2 Ng4+ (17.605.110.442) 1503 

54/94  195:37 +1.50++  11.Kxh2 (17.648.939.229) 1503 

54/94  212:58 +1.64    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (19.213.645.390) 1503 

55/85  257:07 +1.53--  11.Kxh2 Ng4+ (23.138.797.156) 1499 

55/85  260:40 +1.42--  11.Kxh2 Ng4+ (23.454.040.249) 1499 

55/85  264:11 +1.53++  11.Kxh2 (23.760.133.894) 1498 

55/85  279:38 +1.54    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (25.156.696.053) 1499 

56/80  291:37 +1.44--  11.Kxh2 Ng4+ (26.192.854.430) 1496 

56/82  295:48 +1.33--  11.Kxh2 Ng4+ (26.556.289.830) 1496 

56/82  309:55 +1.44++  11.Kxh2 (27.826.295.722) 1496 

56/90  331:02 +1.47    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (29.723.211.080) 1496 

57/90  512:01 +1.37--  11.Kxh2 Ng4+ (46.093.627.830) 1500 

57/90  518:11 +1.47++  11.Kxh2 (46.651.213.198) 1500 

57/90  549:46 +1.63++  11.Kxh2 (49.548.598.246) 1502 

57/99  567:36 +1.44    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (51.169.077.895) 1502 

58/91  615:52 +1.54++  11.Kxh2 (55.659.477.410) 1506 

58/91  713:58 +1.65++  11.Kxh2 (64.687.754.062) 1510 

58/91  729:30 +1.58    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (66.080.428.262) 1509 

59/102 803:50 +1.47--  11.Kxh2 Ng4+ (72.772.984.283) 1508 

59/102 817:27 +1.58++  11.Kxh2 (73.992.956.605) 1508 

59/108 1040:31+1.74++  11.Kxh2 (94.286.584.607) 1510 

59/108 1085:07+1.37--  11.Kxh2 Ng4+ (98.289.599.965) 1509 

59/108 1092:17+1.66++  11.Kxh2 (98.931.829.257) 1509 

59/108 1108:49+1.50    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (100.370.841.437) 1508 

60/99  1187:13+1.39--  11.Kxh2 Ng4+ (107.323.021.394) 1506 

60/99  1202:20+1.50++  11.Kxh2 (108.564.407.924) 1504 
After 19. f5 {!?} also the score for 19...Rxe3 is not so good anymore at very long time, so it could indeed be an improvement over 19. Nd1



Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

44 674:28 +1.48 1...Rxe3 2.Nxd5 Re8 3.Ndf4 Qxg3+
4.Kd1 Nf2+ 5.Rxf2 Qxf2 6.Kc1 Qf1+
7.Qd1 Qxd1+ 8.Kxd1 Nd7 9.Kc2 Nf6
10.Ng3 g6 11.Rh1 Ng4 12.Rg1 a5
13.Ngh5 Ne3+ 14.Kb3 Bxf5 (61.622.918.102) 1522

44 674:28 +1.78 1...Qg2 2.Nf4 Qxg3+ 3.Kd1 Nd7 4.Nce2 Nxe3+
5.Bxe3 Qxe3 6.Qd2 Nf6 7.Qxe3 Rxe3
8.Kd2 Re4 9.Bxe4 Nxe4+ 10.Kc1 Bxf5
11.Nxd5 Bh3 12.Rh1 Bg4 13.Ndc3 Re8
14.Kc2 Ng5 (61.622.918.102) 1522

44 674:28 +1.84 1...Nd7 2.e4 Nh2 3.Rf4 Qg2 4.Qb1 Nf6
5.Kd1 Nh5 6.Rh4 Nxg3 7.Rxh2 Qxh2
8.Nxg3 Qxg3 9.Kc2 dxe4 10.Nxe4 Qh4
11.Qh1 Qxh1 12.Rxh1 Bxf5 13.Nf6+ gxf6
14.Bxf5 Re2 (61.622.918.102) 1522

44 674:28 +2.39 1...Na6 2.Nd1 Qg2 3.Rg1 Qh2 4.Kf1 Qh3+
5.Rg2 Qh1+ 6.Ng1 Nh2+ 7.Rxh2 Qxh2
8.g4 h5 9.g5 c5 10.dxc5 b6 11.cxb6 axb6
12.b4 Bd7 13.Be1 Qxc2 14.Bxc2 Rac8 (61.622.918.102) 1522

44 674:28 +2.70 1...b6 2.e4 Nh2 3.Rg1 Nd7 4.Nf4 Qg4
5.Be2 Nf3+ 6.Kf2 dxe4 7.Nxe4 Nxd4
8.Bxg4 Nxc2 9.Rac1 Rxe4 10.Rxc2 Nf6
11.Bh3 Re5 12.Kg2 Bxf5 13.Bxf5 Rxf5
14.Rxc6 g5 (61.622.918.102) 1522
________________________________________________________________
.
.
.

47 1377:12+1.68 1...Rxe3 2.Nxd5 Re8 3.Ndf4 Qxg3+
4.Kd1 Nf2+ 5.Rxf2 Qxf2 6.Kc1 Qf1+
7.Qd1 Qxd1+ 8.Kxd1 c5 9.Nd5 Na6
10.Rc1 cxd4 11.b4 Re5 12.Nef4 Kh8
13.b5 Bxf5 14.bxa6 Bg4+ (126.633.595.890) 1532

47 1377:12+1.81 1...Nd7 2.e4 Nh2 3.Rf4 Qg2 4.Qb1 Nf6
5.Kd1 Nh5 6.Rh4 Nxg3 7.Rxh2 Qxh2
8.Nxg3 Qxg3 9.Kc2 dxe4 10.Nxe4 Qh4
11.Qh1 Qxh1 12.Rxh1 Bxf5 13.Nf6+ gxf6
14.Bxf5 Re2 (126.633.595.890) 1532

47 1377:12+1.99 1...Qg2 2.Nf4 Qxg3+ 3.Kd1 Nd7 4.Nce2 Nxe3+
5.Bxe3 Qxe3 6.Qd2 Nf6 7.Qxe3 Rxe3
8.Kd2 Re4 9.Bxe4 Nxe4+ 10.Ke3 Bxf5
11.Nxd5 Re8 12.Rxf5 Nd6+ 13.Re5 Nc4+
14.Kf4 Nxe5 (126.633.595.890) 1532

47 1377:12+2.51 1...b6 2.e4 Nh2 3.Rg1 Nd7 4.Nf4 Qg4
5.Be2 Nf3+ 6.Kf2 dxe4 7.Nxe4 Nxd4
8.Bxg4 Nxc2 9.Rac1 Rxe4 10.Rxc2 Nf6
11.Bh3 Re5 12.Kg2 Ne4 13.Bc1 c5
14.Re1 Bb7 (126.633.595.890) 1532

46 1377:12+2.68 1...Na6 2.Nf4 Qxg3+ 3.Kd1 Qxe3
4.Bxe3 Nxe3+ 5.Kd2 Nxc2 6.Bxc2 Nb8
7.Rh1 Nd7 8.Rag1 Nf6 9.Nce2 Bd7
10.Bd3 Rad8 11.Nh5 Nxh5 12.Rxh5 f6
13.Rgh1 h6 14.R5h3 Kh7 (126.633.595.890) 1532
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

jp
Posts: 839
Joined: Mon Apr 23, 2018 5:54 am

Re: Is Black's compensation after 16. g3 really sufficient?

Post by jp » Thu Sep 26, 2019 10:06 am

Eelco de Groot wrote:
Wed Sep 25, 2019 12:07 pm
Engine: Stockfish with small change (512 MB)

Code: Select all

59/102 803:50 +1.47--  11.Kxh2 Ng4+ (72.772.984.283) 1508 

59/102 817:27 +1.58++  11.Kxh2 (73.992.956.605) 1508 

59/108 1040:31+1.74++  11.Kxh2 (94.286.584.607) 1510 

59/108 1085:07+1.37--  11.Kxh2 Ng4+ (98.289.599.965) 1509 

59/108 1092:17+1.66++  11.Kxh2 (98.931.829.257) 1509 

59/108 1108:49+1.50    11.Kxh2 Ng4+ 12.Kg3 Qg5 13.f4 Qh5 
                       14.Bd2 Qh2+ 15.Kf3 Qh4 16.g3 Nh2+ 
                       17.Kf2 Ng4+ 18.Ke1 Qh3 19.Nd1 Nh2 
                       20.Rg1 g6 21.f5 Nf3+ 22.Kf2 Nxg1 
                       23.Kxg1 Bxf5 24.Bxf5 (100.370.841.437) 1508 

60/99  1187:13+1.39--  11.Kxh2 Ng4+ (107.323.021.394) 1506 

60/99  1202:20+1.50++  11.Kxh2 (108.564.407.924) 1504 
What do the "--" and "++" mean? e.g. eval +1.50 (no sign after); +1.39--; +1.50++.

Paloma
Posts: 859
Joined: Thu Dec 25, 2008 8:07 pm

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Paloma » Thu Sep 26, 2019 12:57 pm

fail low and fail high

User avatar
Eelco de Groot
Posts: 4168
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Is Black's compensation after 16. g3 really sufficient?

Post by Eelco de Groot » Thu Sep 26, 2019 1:13 pm

Hi guys, yes that is right, Shredder's way of indicating a "jump out of the search window" by a Fail Low or a Fail High. Other GUIs do that a bit differently sometimes. Stockfish for a while gave really long PVs even in case of a FL or FH but the moves are often very imprecise in such cases (*), even if they can be correctly retrieved from the search. Gary Linscott did some great work on this in the past.

(*) because of the narrow null window search, although even a PV search nowadays does enormous pruning so the PV if deep should not be paid too much attention to. Crystal tries to get better PV by doing less pruning in the PV nodes but it is not always meaningful, if there are just many good moves available. Or the engine makes just one mistake in a critical sidebranch null window search :shock: PV is worth not much anymore either then.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

BrendanJNorman
Posts: 1475
Joined: Sun Feb 07, 2016 11:43 pm
Full name: Brendan J Norman

Re: Is Black's compensation after 16. g3 really sufficient?

Post by BrendanJNorman » Thu Sep 26, 2019 3:23 pm

Even if the sac is dubious...this only illustrates that DJ is a great engine.

The recent flood of 3300+ engines who all play like Stockfish (and have ppl running off to do sim tests!) gets a bit boring imo .

But if an engine can play interesting chess and dubiously sac a piece to draw with perhaps the strongest human player in history...that makes it something special.

Probably just a matter of taste.

Post Reply