Outskirt Forum offers you a new engine which one would you like?

[M.Z]

Outskirt Forum offers you a new engine!

[kramnik]

I have a question on this topic:

since the most of the new engines are based on Stockfish, is there a substantial moves choice difference in determined positions amongst these engines?

I guess that Cfish - for example - and Stockfish chose the same moves in 99% of cases, but I could be wrong...

[SkyNet]

I guess that Cfish - for example - and Stockfish chose the same moves in 99% of cases, but I could be wrong...

Cfish is 100% Stockfish, the difference is that Cfish is written in C language, while SF in C++.

[kramnik]

Ok, this is the definition. But in terms of choice, did you notice a substantial difference between Stockfish and the Stockfish based engines on high depth? (>= 35 plies) For example CorChess, Swordfish, BrainLearn and others.

Plus, let's not forget that Cfish is not properly Stockfish because it uses a now outdated algorythm and a now outdated nnue network.

I guess that Cfish - for example - and Stockfish chose the same moves in 99% of cases, but I could be wrong...

Cfish is 100% Stockfish, the difference is that Cfish is written in C language, while SF in C++.

[camran72]

New engine with improved learning algorithm with a cloud experience file or forum shared experience file

Outskirt Forum offers you a new engine!

[SkyNet]

Ok, this is the definition. But in terms of choice, did you notice a substantial difference between Stockfish and the Stockfish based engines on high depth? (>= 35 plies) For example CorChess, Swordfish, BrainLearn and others.

Plus, let's not forget that Cfish is not properly Stockfish because it uses a now outdated algorythm and a now outdated nnue network.

There is a bit difference in CorChess vs SF analysis (don't know about the other), but it is because of the little bit different algorithm which CorChess is using. Eventually, at very high depths both engines find identical moves (mostly ).

[M.Z]

This is a private reply

The content of this post is classified.

[kramnik]

Then I can't see the reason to put efforts on new lookalike engines. This is the main reason why I'm not updating my engines often as I did in the past: the progresses tend to nothing and we get only draw games.

Ok, this is the definition. But in terms of choice, did you notice a substantial difference between Stockfish and the Stockfish based engines on high depth? (>= 35 plies) For example CorChess, Swordfish, BrainLearn and others.

Plus, let's not forget that Cfish is not properly Stockfish because it uses a now outdated algorythm and a now outdated nnue network.

There is a bit difference in CorChess vs SF analysis (don't know about the other), but it is because of the little bit different algorithm which CorChess is using. Eventually, at very high depths both engines find identical moves (mostly ).

[SkyNet]

Then I can't see the reason to put efforts on new lookalike engines. This is the main reason why I'm not updating my engines often as I did in the past: the progresses tend to nothing and we get only draw games.

They are not exactly "similar", I mean that the style of play at (greater depths) can be similar, but if you compare their ability to solve tactical problems, then you'll see that Corches (and Sugar Iccf) are much better, they can solve positions where SF is blind. We should not confuse "pure power" of SF which we using for book tests/tuning with engines which we use for our (human's) practice/analysis - where Sugar Iccf is "must to have" engine.

[kramnik]

Can CorChess solve this position?

[bbvideo]https://youtu.be/5say2stC7eQ[/bbvideo]

I consider it the real trial.

6 years old video, Stockfish fails even today...

Then I can't see the reason to put efforts on new lookalike engines. This is the main reason why I'm not updating my engines often as I did in the past: the progresses tend to nothing and we get only draw games.

They are not exactly "similar", I mean that the style of play at (greater depths) can be similar, but if you compare their ability to solve tactical problems, then you'll see that Corches (and Sugar Iccf) are much better, they can solve positions where SF is blind. We should not confuse "pure power" of SF which we using for book tests/tuning with engines which we use for our (human's) practice/analysis - where Sugar Iccf is "must to have" engine.

[kramnik]

Ok, just tested now: CorChess cannot solve this position, but Crystal does, and in a matter of seconds!

Now I know what to vote.

I hope MZ will take this position as reference too for problems solving.

Can CorChess solve this position?

[bbvideo]https://youtu.be/5say2stC7eQ[/bbvideo]

I consider it the real trial.

6 years old video, Stockfish fails even today...

Then I can't see the reason to put efforts on new lookalike engines. This is the main reason why I'm not updating my engines often as I did in the past: the progresses tend to nothing and we get only draw games.

They are not exactly "similar", I mean that the style of play at (greater depths) can be similar, but if you compare their ability to solve tactical problems, then you'll see that Corches (and Sugar Iccf) are much better, they can solve positions where SF is blind. We should not confuse "pure power" of SF which we using for book tests/tuning with engines which we use for our (human's) practice/analysis - where Sugar Iccf is "must to have" engine.

[Peter Grayson]

Can CorChess solve this position?

[SetUp "1"] [FEN "8/3P3k/n2K3p/2p3n1/1b4N1/2p1p1P1/8/3B4 w - - 0 50"] { 8/3P3k/n2K3p/2p3n1/1b4N1/2p1p1P1/8/3B4 w - - 0 50 } *

I consider it the real trial.

6 years old video, Stockfish fails even today...

Depends on hardware and time given ...
4Gb hash for all engines. i7-10700 CPU 16 threads.
Analysis by Stockfish dev-20230422-ba06c480 :

50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kb4 Bf6 56.Nd5 Kf7 57.Nxc3 Ke6 58.Ne2 Ke5 59.Nf4 Ne4 60.Kc4 Nxg3 61.Nd3+ Kd6 62.Nf4 Bg5 63.Ng2 Ke5 64.Bg6 Ne2 65.Be8 Nf6 66.Bc6 Nf4 67.Ne1 N4d5 68.Nd3+ Kd6 69.Bb7 h5 70.Kd4 Be3+ 71.Kc4 h4 72.Ne1 h3 73.Nf3 Bb6
-+ (-3.18) Depth: 41/61 00:01:12 919MN, tb=2580586
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Kxc5 Nxd7+ 54.Kb4 Bf6 55.Bc2+ Kg8 56.Nd1 Kf7 57.Nxc3 Ke6 58.Ne2 Ke5 59.Ba4 Nb6 60.Be8 Ne4 61.Kb3 Nd5 62.Kc2 Bd8 63.Nc1 Nxg3 64.Nd3+ Kf5 65.Bc6 Ne3+ 66.Kd2 Bb6 67.Be8 Nd5 68.Ke1 Nf6 69.Bf7 Nfe4 70.Be8 Nd6 71.Bc6 h5 72.Bd7+ Ke4 73.Nf2+ Bxf2+ 74.Kxf2 Kf4 75.Kg2
-+ (-3.24) Depth: 42/66 00:01:32 1171MN, tb=3243793
50.Nf6+
-+ (-2.94 ++) Depth: 43/67 00:02:18 1719MN, tb=5063652
50.Nf6+
-+ (-2.62 ++) Depth: 43/67 00:02:18 1719MN, tb=5064262
50.Nf6+
+- (2.70 ++) Depth: 43/67 00:02:19 1723MN, tb=5086954
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Kc6 Ba3 55.Qd1+ Kxg3 56.Qg1+ Kf3 57.Bd1+ Ke4 58.Qh1+ Ke5 59.Qd5+ Kf6 60.Qd8+ Ke5 61.Qd6+ Ke4 62.Qd5+ Kf4 63.Qc4+ Ne4 64.Qxa6 c2 65.Bxc2 Bb2 66.Bd1 Nf2 67.Qe2 Nxd1 68.Qxd1 Kg5 69.Kd5 Bd4 70.Ke4 Kg6 71.Qg4+ Kf7
+- (3.16) Depth: 43/67 00:02:20 1737MN, tb=5144281
50.Nf6+
+- (3.02 ++) Depth: 44/49 00:02:21 1757MN, tb=5242176
50.Nf6+
+- (3.23 ++) Depth: 44/49 00:02:25 1798MN, tb=5436598
50.Nf6+
+- (3.50 ++) Depth: 44/49 00:02:27 1831MN, tb=5625249
50.Nf6+
+- (3.84 ++) Depth: 44/49 00:02:32 1896MN, tb=5990277
50.Nf6+
+- (4.27 ++) Depth: 44/49 00:02:55 2213MN, tb=7956931
50.Nf6+
+- (4.82 ++) Depth: 44/67 00:03:50 2959MN, tb=12675594


A little faster than
Analysis by CorChess 3 dev-20230423-7a447ace :

50.Nxe3 Ba5 51.Kc6 Bd8 52.Kb5 Nb8 53.Bg4 Kg6 54.Kxc5 Ne4+ 55.Kd4 Nf2 56.Bf5+ Kg5 57.Be6 Nd1 58.Nc2 Kf6 59.Bh3 Ba5 60.Kc4 Nf2 61.Kb5 Bd8 62.Bg2 Nxd7 63.Kb4 Bb6 64.Kxc3 Kf5 65.Bb7 Ne4+ 66.Kb2 Nd6 67.Ba8 Nf6 68.g4+ Nxg4 69.Ne1 Ne5 70.Kb1 Ne4 71.Ng2 h5 72.Kc2
-+ (-3.55) Depth: 36/47 00:02:25 1802MN, tb=5750119
50.Nf6+
-+ (-3.49 ++) Depth: 37/58 00:03:07 2291MN, tb=6941978
50.Nf6+
-+ (-3.46 ++) Depth: 37/58 00:03:08 2291MN, tb=6942780

50.Nf6+
+- (3.99 ++) Depth: 39/49 00:03:41 2667MN, tb=8408596
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Qf6 Nb8 55.Qf4+ Kh3 56.Qxe3 Kg2 57.Qe2+ Kxg3 58.Qe3+ Kg2 59.Kc7 Na6+ 60.Kb7 Ba3 61.Kxa6 Bb2 62.Qf4 Ba3 63.Be4+ Kh3 64.Qe3+ Kg4 65.Bc2 Bb2 66.Bd1+ Kf5 67.Kb5 Kf6 68.Qf4+ Ke7 69.Kc6 c2 70.Bxc2 Bd4
+- (4.31) Depth: 39/52 00:03:54 2821MN, tb=9401252
50.Nf6+
+- (4.39 ++) Depth: 40/25 00:04:08 3014MN, tb=10860532

But more tactical focused engines fair better ...

Analysis by Deep Rybka 4.1 SSE42 x64:

50.d8Q Nf7+[] 51.Kd7 Nxd8[] 52.Kxd8 e2 53.Bxe2 c4 54.Bxc4 Ba5+ 55.Ke7 Nc5 56.Ne3[] Kg6 57.Kd6 Ne4+ 58.Kc6 Kg5 59.Bb3 Nxg3 60.Kb5 Bc7 61.Nd5 Be5 62.Kc4 Ne2 63.Ba4 h5 64.Kd3
-+ (-2.13) Depth: 21 00:00:15 9154kN, tb=171
50.Nf6+ Kg7[] 51.Nh5+
-+ (-2.01 ++) Depth: 21 00:00:16 9645kN, tb=179
50.Nf6+ Kg7[] 51.Nh5+
-+ (-1.86 ++) Depth: 21 00:00:16 9767kN, tb=180
50.Nf6+ Kg7[] 51.Nh5+[]
-/+ (-1.46 ++) Depth: 21 00:00:17 10171kN, tb=180
50.Nf6+ Kg7[] 51.Nh5+[] Kg6 52.Bc2+[] Kxh5 53.d8Q Kg4 54.Qc8+ Kf3 55.Qf5+ Kxg3 56.Qe5+ Kf2 57.Qh2+ Ke1 58.Qh1+ Ke2 59.Qh2+ Ke1 60.Qh1+ Ke2 61.Qh2+ Ke1 62.Qh1+ Ke2 63.Qh2+ Ke1 64.Qh1+ Ke2 65.Qh2+
=/+ (-0.66 ++) Depth: 21 00:00:18 10950kN, tb=184
50.Nf6+ Kg7[] 51.Nh5+[]
+/- (0.94 ++) Depth: 21 00:00:19 12576kN, tb=188
50.Nf6+ Kg7 51.Nh5+[] Kg6 52.Bc2+[] Kxh5 53.d8Q c4+ 54.Kd5 Kg4 55.Kxc4 Kxg3 56.Qd3 Bc5 57.Qxc3 Kg2 58.Qd3 Ba7 59.Qe2+ Kg3 60.Qe1+ Kg2 61.Bd3 Nc5 62.Bf1+ Kf3 63.Be2+ Kg2 64.Qa5 Kf2 65.Qxa7
+- (4.14 ++) Depth: 21 00:01:53 61361kN, tb=2936

Analysis by Solista Attack v2-bmi2:
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kc4 Bf6 56.Nd5 Kf7 57.Nxc3 Ne5+ 58.Kd5 Ng4 59.Nd1 Bd8 60.Kd4 Bc7 61.Kd3 Bxg3 62.Ne3 Ne5+ 63.Ke2 h5 64.Kf1 Kf6 65.Nd5+ Ke6 66.Ne3 Nef3 67.Nf5 Be5 68.Kf2 Nh2 69.Ne3 Ng4+ 70.Ke2 Kf6 71.Nd5+ Kg7 72.Bf5 Nf7 73.Kf1
-+ (-4.71) Depth: 32/50 00:00:13 208MN, tb=625100
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kc4 Bf6 56.Nd5 Kf7 57.Nxc3 Ne5+ 58.Kd5 Ng4 59.Nd1 Bd8 60.Kc4 Bc7 61.Kd3 Bxg3 62.Ne3 Ne5+ 63.Ke2 h5 64.Kf1 Nef3 65.Bd1 Nd2+ 66.Ke2 Nde4 67.Kf1 h4 68.Nf5 Kf6 69.Nxh4 Bxh4 70.Bg4 Ke5 71.Bd7 Nc3 72.Bc6 Kf5 73.Be8 Ke5
-+ (-4.86) Depth: 33/53 00:00:19 285MN, tb=884374
50.Nf6+
-+ (-4.43 ++) Depth: 34/59 00:00:23 347MN, tb=1075540
50.Nf6+
-+ (-4.06 ++) Depth: 34/59 00:00:23 347MN, tb=1076512
50.Nf6+
-+ (-3.58 ++) Depth: 34/59 00:00:23 348MN, tb=1079114
50.Nf6+
-+ (-2.99 ++) Depth: 34/59 00:00:23 349MN, tb=1080440
50.Nf6+
-+ (-2.23 ++) Depth: 34/59 00:00:23 349MN, tb=1080669
50.Nf6+
-/+ (-1.28 ++) Depth: 34/59 00:00:23 349MN, tb=1082388
50.Nf6+
= (-0.08 ++) Depth: 34/59 00:00:23 349MN, tb=1082560

50.Nf6+
+- (6.14 ++) Depth: 35/65 00:00:31 464MN, tb=1621417
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Qe8 c4+ 55.Kc6 Nc5 56.Qxe3 Nce6 57.Kd5 Bc5 58.Qxc3 Bf2 59.Kxc4 Nd4 60.Bd1+ Kh3 61.Qd3 Kg2 62.Qg6 Nh3 63.Bg4 Ng1 64.Qxh6 Nc2 65.Be6 Ne3+ 66.Kd3 Kxg3 67.Qg5+ Kf3 68.Qg6 Kf4 69.Qf6+ Kg3 70.Qe5+ Kf3 71.Qh5+ Kf4 72.Qf7+ Kg3
+- (6.12) Depth: 35/65 00:00:33 489MN, tb=1774542
50.Nf6+
+- (6.58 ++) Depth: 36/43 00:00:39 587MN, tb=2347309
50.Nf6+
+- (7.14 ++) Depth: 36/53 00:00:46 695MN, tb=3011624

Peter

[kramnik]

I was testing it with my phone. Stockfish didn't succeed while Crystal succeded almost immediately. The last time I tried a few years ago on my main PC, Stockfish wasn't able to solve it (and the sharper Cfish neither), while Komodo did it in few time.

The correct move is indeed 50. Nf6+. I am very surprised that Stockfish is now able to solve it... Maybe I should give another try on the big boy with the latest beta release.

Depends on hardware and time given ...
4Gb hash for all engines. i7-10700 CPU 16 threads.
Analysis by Stockfish dev-20230422-ba06c480 :

50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kb4 Bf6 56.Nd5 Kf7 57.Nxc3 Ke6 58.Ne2 Ke5 59.Nf4 Ne4 60.Kc4 Nxg3 61.Nd3+ Kd6 62.Nf4 Bg5 63.Ng2 Ke5 64.Bg6 Ne2 65.Be8 Nf6 66.Bc6 Nf4 67.Ne1 N4d5 68.Nd3+ Kd6 69.Bb7 h5 70.Kd4 Be3+ 71.Kc4 h4 72.Ne1 h3 73.Nf3 Bb6
-+ (-3.18) Depth: 41/61 00:01:12 919MN, tb=2580586
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Kxc5 Nxd7+ 54.Kb4 Bf6 55.Bc2+ Kg8 56.Nd1 Kf7 57.Nxc3 Ke6 58.Ne2 Ke5 59.Ba4 Nb6 60.Be8 Ne4 61.Kb3 Nd5 62.Kc2 Bd8 63.Nc1 Nxg3 64.Nd3+ Kf5 65.Bc6 Ne3+ 66.Kd2 Bb6 67.Be8 Nd5 68.Ke1 Nf6 69.Bf7 Nfe4 70.Be8 Nd6 71.Bc6 h5 72.Bd7+ Ke4 73.Nf2+ Bxf2+ 74.Kxf2 Kf4 75.Kg2
-+ (-3.24) Depth: 42/66 00:01:32 1171MN, tb=3243793
50.Nf6+
-+ (-2.94 ++) Depth: 43/67 00:02:18 1719MN, tb=5063652
50.Nf6+
-+ (-2.62 ++) Depth: 43/67 00:02:18 1719MN, tb=5064262
50.Nf6+
+- (2.70 ++) Depth: 43/67 00:02:19 1723MN, tb=5086954
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Kc6 Ba3 55.Qd1+ Kxg3 56.Qg1+ Kf3 57.Bd1+ Ke4 58.Qh1+ Ke5 59.Qd5+ Kf6 60.Qd8+ Ke5 61.Qd6+ Ke4 62.Qd5+ Kf4 63.Qc4+ Ne4 64.Qxa6 c2 65.Bxc2 Bb2 66.Bd1 Nf2 67.Qe2 Nxd1 68.Qxd1 Kg5 69.Kd5 Bd4 70.Ke4 Kg6 71.Qg4+ Kf7
+- (3.16) Depth: 43/67 00:02:20 1737MN, tb=5144281
50.Nf6+
+- (3.02 ++) Depth: 44/49 00:02:21 1757MN, tb=5242176
50.Nf6+
+- (3.23 ++) Depth: 44/49 00:02:25 1798MN, tb=5436598
50.Nf6+
+- (3.50 ++) Depth: 44/49 00:02:27 1831MN, tb=5625249
50.Nf6+
+- (3.84 ++) Depth: 44/49 00:02:32 1896MN, tb=5990277
50.Nf6+
+- (4.27 ++) Depth: 44/49 00:02:55 2213MN, tb=7956931
50.Nf6+
+- (4.82 ++) Depth: 44/67 00:03:50 2959MN, tb=12675594


A little faster than
Analysis by CorChess 3 dev-20230423-7a447ace :

50.Nxe3 Ba5 51.Kc6 Bd8 52.Kb5 Nb8 53.Bg4 Kg6 54.Kxc5 Ne4+ 55.Kd4 Nf2 56.Bf5+ Kg5 57.Be6 Nd1 58.Nc2 Kf6 59.Bh3 Ba5 60.Kc4 Nf2 61.Kb5 Bd8 62.Bg2 Nxd7 63.Kb4 Bb6 64.Kxc3 Kf5 65.Bb7 Ne4+ 66.Kb2 Nd6 67.Ba8 Nf6 68.g4+ Nxg4 69.Ne1 Ne5 70.Kb1 Ne4 71.Ng2 h5 72.Kc2
-+ (-3.55) Depth: 36/47 00:02:25 1802MN, tb=5750119
50.Nf6+
-+ (-3.49 ++) Depth: 37/58 00:03:07 2291MN, tb=6941978
50.Nf6+
-+ (-3.46 ++) Depth: 37/58 00:03:08 2291MN, tb=6942780

50.Nf6+
+- (3.99 ++) Depth: 39/49 00:03:41 2667MN, tb=8408596
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Qf6 Nb8 55.Qf4+ Kh3 56.Qxe3 Kg2 57.Qe2+ Kxg3 58.Qe3+ Kg2 59.Kc7 Na6+ 60.Kb7 Ba3 61.Kxa6 Bb2 62.Qf4 Ba3 63.Be4+ Kh3 64.Qe3+ Kg4 65.Bc2 Bb2 66.Bd1+ Kf5 67.Kb5 Kf6 68.Qf4+ Ke7 69.Kc6 c2 70.Bxc2 Bd4
+- (4.31) Depth: 39/52 00:03:54 2821MN, tb=9401252
50.Nf6+
+- (4.39 ++) Depth: 40/25 00:04:08 3014MN, tb=10860532

But more tactical focused engines fair better ...

Analysis by Deep Rybka 4.1 SSE42 x64:

50.d8Q Nf7+[] 51.Kd7 Nxd8[] 52.Kxd8 e2 53.Bxe2 c4 54.Bxc4 Ba5+ 55.Ke7 Nc5 56.Ne3[] Kg6 57.Kd6 Ne4+ 58.Kc6 Kg5 59.Bb3 Nxg3 60.Kb5 Bc7 61.Nd5 Be5 62.Kc4 Ne2 63.Ba4 h5 64.Kd3
-+ (-2.13) Depth: 21 00:00:15 9154kN, tb=171
50.Nf6+ Kg7[] 51.Nh5+
-+ (-2.01 ++) Depth: 21 00:00:16 9645kN, tb=179
50.Nf6+ Kg7[] 51.Nh5+
-+ (-1.86 ++) Depth: 21 00:00:16 9767kN, tb=180
50.Nf6+ Kg7[] 51.Nh5+[]
-/+ (-1.46 ++) Depth: 21 00:00:17 10171kN, tb=180
50.Nf6+ Kg7[] 51.Nh5+[] Kg6 52.Bc2+[] Kxh5 53.d8Q Kg4 54.Qc8+ Kf3 55.Qf5+ Kxg3 56.Qe5+ Kf2 57.Qh2+ Ke1 58.Qh1+ Ke2 59.Qh2+ Ke1 60.Qh1+ Ke2 61.Qh2+ Ke1 62.Qh1+ Ke2 63.Qh2+ Ke1 64.Qh1+ Ke2 65.Qh2+
=/+ (-0.66 ++) Depth: 21 00:00:18 10950kN, tb=184
50.Nf6+ Kg7[] 51.Nh5+[]
+/- (0.94 ++) Depth: 21 00:00:19 12576kN, tb=188
50.Nf6+ Kg7 51.Nh5+[] Kg6 52.Bc2+[] Kxh5 53.d8Q c4+ 54.Kd5 Kg4 55.Kxc4 Kxg3 56.Qd3 Bc5 57.Qxc3 Kg2 58.Qd3 Ba7 59.Qe2+ Kg3 60.Qe1+ Kg2 61.Bd3 Nc5 62.Bf1+ Kf3 63.Be2+ Kg2 64.Qa5 Kf2 65.Qxa7
+- (4.14 ++) Depth: 21 00:01:53 61361kN, tb=2936

Analysis by Solista Attack v2-bmi2:
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kc4 Bf6 56.Nd5 Kf7 57.Nxc3 Ne5+ 58.Kd5 Ng4 59.Nd1 Bd8 60.Kd4 Bc7 61.Kd3 Bxg3 62.Ne3 Ne5+ 63.Ke2 h5 64.Kf1 Kf6 65.Nd5+ Ke6 66.Ne3 Nef3 67.Nf5 Be5 68.Kf2 Nh2 69.Ne3 Ng4+ 70.Ke2 Kf6 71.Nd5+ Kg7 72.Bf5 Nf7 73.Kf1
-+ (-4.71) Depth: 32/50 00:00:13 208MN, tb=625100
50.Nxe3 Ba5 51.Kc6 Nb8+ 52.Kb5 Bd8 53.Bc2+ Kg8 54.Kxc5 Nxd7+ 55.Kc4 Bf6 56.Nd5 Kf7 57.Nxc3 Ne5+ 58.Kd5 Ng4 59.Nd1 Bd8 60.Kc4 Bc7 61.Kd3 Bxg3 62.Ne3 Ne5+ 63.Ke2 h5 64.Kf1 Nef3 65.Bd1 Nd2+ 66.Ke2 Nde4 67.Kf1 h4 68.Nf5 Kf6 69.Nxh4 Bxh4 70.Bg4 Ke5 71.Bd7 Nc3 72.Bc6 Kf5 73.Be8 Ke5
-+ (-4.86) Depth: 33/53 00:00:19 285MN, tb=884374
50.Nf6+
-+ (-4.43 ++) Depth: 34/59 00:00:23 347MN, tb=1075540
50.Nf6+
-+ (-4.06 ++) Depth: 34/59 00:00:23 347MN, tb=1076512
50.Nf6+
-+ (-3.58 ++) Depth: 34/59 00:00:23 348MN, tb=1079114
50.Nf6+
-+ (-2.99 ++) Depth: 34/59 00:00:23 349MN, tb=1080440
50.Nf6+
-+ (-2.23 ++) Depth: 34/59 00:00:23 349MN, tb=1080669
50.Nf6+
-/+ (-1.28 ++) Depth: 34/59 00:00:23 349MN, tb=1082388
50.Nf6+
= (-0.08 ++) Depth: 34/59 00:00:23 349MN, tb=1082560

50.Nf6+
+- (6.14 ++) Depth: 35/65 00:00:31 464MN, tb=1621417
50.Nf6+ Kg7 51.Nh5+ Kg6 52.Bc2+ Kxh5 53.d8Q Kg4 54.Qe8 c4+ 55.Kc6 Nc5 56.Qxe3 Nce6 57.Kd5 Bc5 58.Qxc3 Bf2 59.Kxc4 Nd4 60.Bd1+ Kh3 61.Qd3 Kg2 62.Qg6 Nh3 63.Bg4 Ng1 64.Qxh6 Nc2 65.Be6 Ne3+ 66.Kd3 Kxg3 67.Qg5+ Kf3 68.Qg6 Kf4 69.Qf6+ Kg3 70.Qe5+ Kf3 71.Qh5+ Kf4 72.Qf7+ Kg3
+- (6.12) Depth: 35/65 00:00:33 489MN, tb=1774542
50.Nf6+
+- (6.58 ++) Depth: 36/43 00:00:39 587MN, tb=2347309
50.Nf6+
+- (7.14 ++) Depth: 36/53 00:00:46 695MN, tb=3011624

Peter

[M.Z]

before 30 seconds finds 50. Cf6+

[Nemesis]

before 30 seconds finds 50. Cf6+

► Show Spoiler

Gimme Gimme ..