Crazyhouse tournaments and rating list

Ferdy · Post by **Ferdy** » Mon Oct 17, 2016 12:49 pm

GregNeto wrote:Hi! First of all thank you for your crazyhouse campaign. I started to do matches of the top three engines with long time controls, i will send you the pgn's.
Concerning chess960: Aren' t there only 480 positions without castling, because half of the positions are mirrored?

Regards Greg

It is better to include the mirrored positions because this will test the files or squares specific evaluation of the engines if there is.

hgm · Post by **hgm** » Mon Oct 17, 2016 8:46 pm

First win of my new engine in zh:

Code: Select all

&#91;Event "Computer Chess Game"&#93;
&#91;Site "ONTWIKKELLAPTOP"&#93;
&#91;Date "2016.10.17"&#93;
&#91;Round "-"&#93;
&#91;White "Pulsar2009-9b"&#93;
&#91;Black "CrazyWa 0.0"&#93;
&#91;Result "0-1"&#93;
&#91;TimeControl "40/60"&#93;
&#91;Variant "crazyhouse"&#93;

1. c4 &#123;+0.02/4&#125; e6 &#123;-1.20/10 4&#125; 2. Nf3 Qe7 &#123;-1.16/9 1.5&#125; 3. a3
&#123;+0.31/5 1.4&#125; Qc5 &#123;-0.39/8 0.8&#125; 4. d3 &#123;+0.55/5 1.4&#125; Be7 &#123;-0.38/7 0.9&#125; 5. b4
&#123;+0.99/5 1.4&#125; Qh5 &#123;+0.40/8 2.9&#125; 6. Bb2 &#123;+1.15/5 1.4&#125; Qh6 &#123;+0.43/8 1.0&#125; 7.
e4 &#123;+1.33/5 1.4&#125; f5 &#123;+0.31/8 3&#125; 8. exf5 &#123;+1.17/5 1.4&#125; exf5 &#123;+0.14/7 0.7&#125; 9.
Ne5 &#123;+1.85/4 1.4&#125; Nf6 &#123;+0.42/5 4&#125; 10. P@f7+ &#123;+2.19/4 1.4&#125; Kf8 &#123;+0.43/5 0.6&#125;
11. c5 &#123;+1.65/5 1.4&#125; d6 &#123;+2.53/6 5&#125; 12. cxd6 &#123;+2.10/5 1.4&#125; Bxd6 &#123;+2.54/4 4&#125;
13. Qd2 &#123;+0.02/3 1.5&#125; Bxe5 &#123;+3.70/5 1.0&#125; 14. Bxe5 &#123;-0.81/5 1.4&#125; Qxd2+
&#123;+3.82/5 1.5&#125; 15. Kxd2 &#123;-3.11/4 1.5&#125; N@b3+ &#123;+3.82/4 0.7&#125; 16. Kc3
&#123;-3.97/4 1.4&#125; Nxa1 &#123;+6.76/4 0.6&#125; 17. Q@d8+ &#123;-5.64/3 1.4&#125; Kxf7 &#123;+8.88/4 0.5&#125;
18. B@c4+ &#123;-9.76/3 1.4&#125; Kg6 &#123;+12.17/4 2.7&#125; 19. Qxh8 &#123;-12.55/4 1.4&#125; Q@c1+
&#123;+17.12/3 2.6&#125; 20. P@c2 &#123;-15.66/4 1.5&#125; P@d4+ &#123;+79.95/4 0.8&#125; 21. Kxd4
&#123;-10.30/4 1.4&#125; P@c5+ &#123;+79.96/4 0.6&#125; 22. Kxc5 &#123;-149.97/4 1.4&#125; b6+
&#123;+79.97/4 0.4&#125; 23. Kb5 &#123;-149.98/4 1.4&#125; Bd7+ &#123;+79.98/4 1.7&#125; 24. P@c6
&#123;-149.99/4 1.5&#125; Bxc6# &#123;+79.99/4 2.1&#125;
0-1

OK, it was only against Pulsar. But still a milestone!

Carlos777 · Post by **Carlos777** » Mon Oct 17, 2016 9:06 pm

hgm wrote:First win of my new engine in zh:

OK, it was only against Pulsar. But still a milestone!

Congratulations! A new zh engine.

Ferdy · Post by **Ferdy** » Tue Oct 18, 2016 4:00 am

hgm wrote:First win of my new engine in zh:

Code: Select all

&#91;Event "Computer Chess Game"&#93;
&#91;Site "ONTWIKKELLAPTOP"&#93;
&#91;Date "2016.10.17"&#93;
&#91;Round "-"&#93;
&#91;White "Pulsar2009-9b"&#93;
&#91;Black "CrazyWa 0.0"&#93;
&#91;Result "0-1"&#93;
&#91;TimeControl "40/60"&#93;
&#91;Variant "crazyhouse"&#93;

1. c4 &#123;+0.02/4&#125; e6 &#123;-1.20/10 4&#125; 2. Nf3 Qe7 &#123;-1.16/9 1.5&#125; 3. a3
&#123;+0.31/5 1.4&#125; Qc5 &#123;-0.39/8 0.8&#125; 4. d3 &#123;+0.55/5 1.4&#125; Be7 &#123;-0.38/7 0.9&#125; 5. b4
&#123;+0.99/5 1.4&#125; Qh5 &#123;+0.40/8 2.9&#125; 6. Bb2 &#123;+1.15/5 1.4&#125; Qh6 &#123;+0.43/8 1.0&#125; 7.
e4 &#123;+1.33/5 1.4&#125; f5 &#123;+0.31/8 3&#125; 8. exf5 &#123;+1.17/5 1.4&#125; exf5 &#123;+0.14/7 0.7&#125; 9.
Ne5 &#123;+1.85/4 1.4&#125; Nf6 &#123;+0.42/5 4&#125; 10. P@f7+ &#123;+2.19/4 1.4&#125; Kf8 &#123;+0.43/5 0.6&#125;
11. c5 &#123;+1.65/5 1.4&#125; d6 &#123;+2.53/6 5&#125; 12. cxd6 &#123;+2.10/5 1.4&#125; Bxd6 &#123;+2.54/4 4&#125;
13. Qd2 &#123;+0.02/3 1.5&#125; Bxe5 &#123;+3.70/5 1.0&#125; 14. Bxe5 &#123;-0.81/5 1.4&#125; Qxd2+
&#123;+3.82/5 1.5&#125; 15. Kxd2 &#123;-3.11/4 1.5&#125; N@b3+ &#123;+3.82/4 0.7&#125; 16. Kc3
&#123;-3.97/4 1.4&#125; Nxa1 &#123;+6.76/4 0.6&#125; 17. Q@d8+ &#123;-5.64/3 1.4&#125; Kxf7 &#123;+8.88/4 0.5&#125;
18. B@c4+ &#123;-9.76/3 1.4&#125; Kg6 &#123;+12.17/4 2.7&#125; 19. Qxh8 &#123;-12.55/4 1.4&#125; Q@c1+
&#123;+17.12/3 2.6&#125; 20. P@c2 &#123;-15.66/4 1.5&#125; P@d4+ &#123;+79.95/4 0.8&#125; 21. Kxd4
&#123;-10.30/4 1.4&#125; P@c5+ &#123;+79.96/4 0.6&#125; 22. Kxc5 &#123;-149.97/4 1.4&#125; b6+
&#123;+79.97/4 0.4&#125; 23. Kb5 &#123;-149.98/4 1.4&#125; Bd7+ &#123;+79.98/4 1.7&#125; 24. P@c6
&#123;-149.99/4 1.5&#125; Bxc6# &#123;+79.99/4 2.1&#125;
0-1

OK, it was only against Pulsar. But still a milestone!

Congratulations

Vinvin · Post by **Vinvin** » Tue Oct 18, 2016 7:54 am

Hello Ferdinand, just a question from me : is there multi-threaded engines or all are single-threaded ?

Vincent

Ferdy · Post by **Ferdy** » Tue Oct 18, 2016 8:07 am

Vinvin wrote:Hello Ferdinand, just a question from me : is there multi-threaded engines or all are single-threaded ?

Vincent

I have not known any so far that supports multi-threads, except NebiyuAlien but Daniel has not compiled and released yet a binary that supports multi-threads. I am actually waiting for NebiyuAlien before the champ's tour.

hgm · Post by **hgm** » Tue Oct 18, 2016 1:47 pm

Apparently QS can blow up enormously in Crazyhouse. In a test game my engine suddenly 'asserted out' because the maximum recursion depth was exceeded. Normally that points to an error (e.g. both sides in check, and extending). But in this case there seemed to be a line that just went on with captures (or promotions) that also happened to be checks, giving the other side te opportunity to drop pieces back as evasion. Eventually dropping something that inadvertantly also checked, so that he could take over the chase.

Code: Select all

&#91;Event "Edited game"&#93;
&#91;Site "ONTWIKKELLAPTOP"&#93;
&#91;Date "2016.10.18"&#93;
&#91;Round "-"&#93;
&#91;White "-"&#93;
&#91;Black "-"&#93;
&#91;Result "1-0"&#93;
&#91;Variant "crazyhouse"&#93;

1. e4 &#123;+0.02/4&#125; Nf6 &#123;+0.74/10 1.7&#125; 2. Nc3 Nc6 &#123;+0.68/10 2.8&#125; 3. d4
&#123;+0.43/5 1.5&#125; e6 &#123;+0.63/9 2.1&#125; 4. e5 &#123;+0.50/5 1.4&#125; Ng8 &#123;+0.81/9 2.8&#125; 5. Nf3
&#123;+1.16/5 1.4&#125; Bb4 &#123;+1.74/9 4&#125; 6. a3 &#123;+0.88/5 1.4&#125; Bxc3+ &#123;+4.36/8 1.1&#125; 7.
bxc3 &#123;+0.22/6 1.4&#125; Qe7 &#123;+4.27/8 4&#125; 8. Bg5 &#123;+2.56/5 1.4&#125; f6 &#123;+4.39/8 2.1&#125; 9.
B@h5+ &#123;+0.55/4 1.4&#125; Kf8 &#123;+5.43/7 0.8&#125; 10. Bd2 &#123;+0.60/5 1.4&#125; fxe5
&#123;+5.75/8 1.5&#125; 11. Nxe5 &#123;+1.36/5 1.4&#125; Nxe5 &#123;+5.83/7 1.5&#125; 12. dxe5
&#123;+0.96/5 1.5&#125; g6 &#123;+5.96/6 1.4&#125; 13. P@f6 &#123;+2.88/4 1.4&#125; Qc5 &#123;+6.87/7 1.4&#125; 14.
N@d3 &#123;+2.96/4 1.4&#125; Qa5 &#123;+3.73/6 2.7&#125; 15. Bxg6 &#123;+7.81/5 1.4&#125; P@g7
&#123;+1.33/5 1.9&#125; 16. P@e7+ &#123;+8.19/4 1.4&#125; Nxe7 &#123;-0.03/7 0.8&#125; 17. fxg7+
&#123;+7.85/5 1.4&#125; Kxg7 &#123;-0.92/7 1.9&#125; 18. P@f6+ &#123;+9.66/4 1.4&#125; Kg8 &#123;-0.93/6 1.6&#125;
19. fxe7 &#123;+9.66/4 1.4&#125; hxg6 &#123;-0.93/4 1.1&#125; 20. e8=Q+ &#123;+11.58/4 1.4&#125; Kg7
&#123;-4.18/5 1.9&#125; 21. Qxh8+ &#123;+13.05/4 1.4&#125; Kxh8 &#123;-2.03/5 0.7&#125; 22. R@d8+
&#123;+12.45/3 1.4&#125; Kg7 &#123;+0.59/6 1.4&#125; 23. N@e8+ &#123;+12.22/4 1.4&#125; Kf7 &#123;+1.80/6 0.9&#125;
24. c4 &#123;+10.95/4 1.4&#125; B@c3 &#123;+2.59/5 1.4&#125; 25. Qf3+ &#123;+10.29/4 1.4&#125; P@f5
&#123;+7.04/5 0.8&#125; 26. Kd1 &#123;+3.08/4 1.4&#125; Bxd2 &#123;+8.16/5 1.0&#125; 27. Nf6
&#123;+2.30/4 1.4&#125; Bh6 &#123;+10.80/4 2.4&#125; 28. Ne4 &#123;-0.38/4 1.4&#125; N@c3+ &#123;+13.32/5 1.1&#125;
29. Nxc3 &#123;-1.96/5 1.4&#125; Qxc3 &#123;+22.14/5 0.9&#125; 30. N@b3 &#123;-9.93/3 1.4&#125; Qxa1+
&#123;+79.95/5 1.2&#125; 31. Ndc1 &#123;-15.54/4 1.4&#125;

[d] r1bR4/pppp1k2/4p1pb/4Pp2/2P5/PN3Q2/2P2PPP/q1NK1B1R b
Hand: [ppnnbr]

The line on which it crashed in this position was:

Code: Select all

&#123;646404&#125; R@d4 &#123;45640401&#125; d1e1 &#123;30003&#125; h6d2 &#123;19030301&#125; b3d2 &#123;20002&#125; P@b2 &#123;10100&#125; -- &#123;1&#125; b2c1q &#123;2010101&#125; e1e2 
&#123;1&#125; d4d2 &#123;19010101&#125; e2e3 &#123;1&#125; a1a3 &#123;2c010101&#125; c2c3 &#123;1&#125; a3c3 &#123;2e010101&#125; f1d3 &#123;1&#125; c3d3 &#123;2f010101&#125; e3f4 &#123;1&#125; d3c4 
&#123;44010101&#125; f3e4 &#123;1&#125; c4e4 &#123;46010101&#125; f4g5 &#123;1&#125; e4g2 &#123;1c010101&#125; g5h6 &#123;1&#125; g2h2 &#123;1d010101&#125; h1h2 &#123;1&#125; d2f2 &#123;2010101&#125; h6h7 &#123;1&#125; f2h2 
&#123;1d010101&#125; B@h3 &#123;1&#125; h2h3 &#123;33010101&#125; Q@h5 &#123;1&#125; g6h5 &#123;1&#125; d8d7 &#123;87010101&#125; P@e7 &#123;1&#125; d7e7 &#123;88010101&#125; f7f8 &#123;1&#125; e7c7 &#123;1&#125; c1c7 
&#123;86010101&#125; P@e7 &#123;88010101&#125; f8f7 &#123;1&#125; e7e8q &#123;9e010101&#125; f7e8 &#123;86010101&#125; P@d7 &#123;87010101&#125; e8e7 &#123;1&#125; d7c8n &#123;9c010101&#125; e7e8 
&#123;86010101&#125; P@d7 &#123;87010101&#125; e8f7 &#123;1&#125; c8a7 &#123;1&#125; a8a7 &#123;1&#125; d7d8n &#123;9d010101&#125; f7e8 &#123;86010101&#125; P@f7 &#123;89010101&#125; e8d7 &#123;1&#125; f7f8n 
&#123;9f010101&#125; d7e8 &#123;86010101&#125; B@f7 &#123;89010101&#125; e8e7 &#123;1&#125; f7h5 &#123;1&#125; e7f8 &#123;86010101&#125; P@g7 &#123;8a010101&#125; f8e7 &#123;1&#125; g7g8q &#123;1&#125; c7e5 &#123;1&#125; d8b7 
&#123;1&#125; a7b7 &#123;1&#125; g8e6 &#123;72010101&#125; e7d8 &#123;85010101&#125; P@c7 &#123;86010101&#125; d8c7 &#123;1&#125; e6e5 &#123;5c010101&#125; c7c8 &#123;85010101&#125; P@c7 &#123;1&#125; b7c7 
&#123;86010101&#125; P@d7 &#123;87010101&#125; c8d7 &#123;1&#125; e5f5 &#123;5d010101&#125; R@e6 &#123;1&#125; f5e6 &#123;72010101&#125; d7e6 &#123;86010101&#125; R@e7 &#123;88010101&#125; e6d6 &#123;1&#125; e7c7 
&#123;1&#125; h3h5 &#123;5f010101&#125; R@h6 &#123;75010101&#125; d6d5 &#123;1&#125; h6h5 &#123;5f010101&#125; N@g5 &#123;5e010101&#125; h5g5 &#123;5e010101&#125; Q@f5 &#123;5d010101&#125; g5f5 &#123;5d010101&#125; d5e5

(There is one nullmove in the line, indicated as '--'.) This happened in the d=4 iteration, after it found a forced mate there:

Code: Select all

4 7996 17 97521 a1c1 b3c1 N@b2 d1e1 B@d2 e1e2 R@e1

So one side tries to beat the mate score, and no amount of material will satisfy him in the stand pat. The other side cannot stand pat, because he is in check all the time (and gets a 1-ply extension for the evasion).

Nevertheless it is fishy that at some point the loosing side can take over the checking. One would expect him to stand pat, happy as he is not to be checkmated. So perhaps my mate score is not high enough, so that after stripping the opponent bare the scorerises above it. In Crazyhouse scores can of course be double from what they are in Chess, if the piece values were the same, as every capture counts double.

Of course mate-distance pruning would also help here.

hgm · Post by **hgm** » Tue Oct 18, 2016 5:16 pm

Interesting. Pulsar sometimes claims stalemate when it gets checkmated:

Code: Select all

&#91;Event "Computer Chess Game"&#93;
&#91;Site "ONTWIKKELLAPTOP"&#93;
&#91;Date "2016.10.18"&#93;
&#91;Round "29"&#93;
&#91;White "Pulsar2009-9b"&#93;
&#91;Black "CrazyWa 0.0"&#93;
&#91;Result "1/2-1/2"&#93;
&#91;TimeControl "40/60"&#93;
&#91;Variant "crazyhouse"&#93;

1. c4 &#123;+0.02/4&#125; Nc6 &#123;-0.98/10 1.7&#125; 2. g3 Rb8 &#123;-0.97/9 1.2&#125; 3. d4
&#123;+0.09/6 1.4&#125; e5 &#123;-0.87/8 1.0&#125; 4. dxe5 &#123;+0.77/5 1.4&#125; Nxe5 &#123;-1.10/7 1.1&#125; 5.
Qd5 &#123;+0.32/4 1.4&#125; f6 &#123;-0.76/6 1.7&#125; 6. Nc3 &#123;+0.13/4 1.4&#125; Bb4 &#123;+0.50/6 1.1&#125;
7. Qd4 &#123;-0.47/4 1.4&#125; c5 &#123;+1.50/6 3&#125; 8. Qd6 &#123;-2.61/4 1.4&#125; Bxc3+
&#123;+1.36/6 1.7&#125; 9. bxc3 &#123;-2.02/5 1.4&#125; N@c2+ &#123;+1.35/5 1.1&#125; 10. Kd1
&#123;-3.92/5 1.5&#125; Nxa1 &#123;+7.09/6 2.9&#125; 11. B@h5+ &#123;-5.20/3 1.4&#125; g6 &#123;+10.11/6 2.7&#125;
12. Bxg6+ &#123;-7.00/4 1.4&#125; Nxg6 &#123;+11.97/6 3&#125; 13. P@f7+ &#123;-12.61/4 1.4&#125; Kxf7
&#123;+14.07/6 1.1&#125; 14. Ke1 &#123;-16.98/4 1.4&#125; P@b2 &#123;+16.67/5 1.2&#125; 15. Bxb2
&#123;-18.15/4 1.4&#125; R@b1+ &#123;+17.16/5 0.9&#125; 16. Kd2 &#123;-18.15/5 1.4&#125; B@h6+
&#123;+16.70/5 5&#125; 17. P@f4 &#123;-15.10/5 1.4&#125; Rxb2+ &#123;+15.48/5 1.8&#125; 18. Kc1
&#123;-16.00/5 1.4&#125; Rxa2 &#123;+16.29/5 3&#125; 19. P@e6+ &#123;-14.67/4 1.4&#125; Ke8
&#123;+22.38/6 2.3&#125; 20. Qxb8 &#123;-14.96/4 1.4&#125; Rc2+ &#123;+79.96/4 2.0&#125; 21. Kb1
&#123;-149.97/4 1.4&#125; Qb6+ &#123;+79.97/9 0.8&#125; 22. R@b5 &#123;-149.98/4 1.4&#125; Qxb5+
&#123;+79.98/77 0.6&#125; 23. cxb5 &#123;-149.99/4 1.4&#125; R@c1#
&#123;+79.99/2 0.1&#125; 1/2-1/2

From its output it seems to do that by removing its King

:

Code: Select all

2792565 >first &#58; R@c1
2792573 <first &#58; #testing crazyhouse evaluate is 764
2792573 <first &#58; # proccessed time
2792574 <first &#58; #testing crazyhouse evaluate is 764
2792574 <first &#58; #testing crazyhouse evaluate is 764
2792574 <first &#58; # in plop move where its written a+4
2792575 <first &#58; # finished call to getmove and turn is 46 and wstat is 2
2792576 <first &#58; #board before search
2792577 <first &#58; # 8 6 10 0 0 3 2 4
2792577 <first &#58; # 0 0 10 0 1 1 0 1
2792577 <first &#58; # 0 0 1 0 0 0 1 0
2792577 <first &#58; # 0 0 0 0 0 1 0 0
2792577 <first &#58; # 0 1 7 0 0 0 0 0
2792578 <first &#58; # 0 0 0 0 1 7 8 9
2792578 <first &#58; # 7 7 0 7 0 0 0 7
2792578 <first &#58; # 0 5 9 0 12 0 8 10
2792578 <first &#58; 1 -15000 0 0 b1b2 
2792581 <first &#58; 
2792581 <first &#58; # maxdepth = 1 
2792581 <first &#58; 2 -15000 0 0 b1b2 p@a2 
2792582 <first &#58; 
2792583 <first &#58; # maxdepth = 2 
2792583 <first &#58; 3 -15000 0 0 b1b2 p@a2 b1a1 
2792584 <first &#58; 
2792584 <first &#58; # maxdepth = 3 
2792584 <first &#58; # board before move
2792585 <first &#58; # 8 6 10 0 0 3 2 4
2792585 <first &#58; # 0 0 10 0 1 1 0 1
2792585 <first &#58; # 0 0 1 0 0 0 1 0
2792585 <first &#58; # 0 0 0 0 0 1 0 0
2792586 <first &#58; # 0 1 7 0 0 0 0 0
2792586 <first &#58; # 0 0 0 0 1 7 8 9
2792586 <first &#58; # 7 7 0 7 0 0 0 7
2792586 <first &#58; # 0 5 9 0 12 0 8 10
2792586 <first &#58; # 8 0 10 0 0 3 2 4
2792586 <first &#58; # 0 0 10 0 1 1 0 1
2792587 <first &#58; # 0 0 1 0 0 0 1 0
2792587 <first &#58; # 0 0 0 0 0 1 0 0
2792587 <first &#58; # 0 1 7 0 0 0 0 0
2792587 <first &#58; # 0 0 0 0 1 7 8 9
2792587 <first &#58; # 7 7 0 7 0 0 0 7
2792587 <first &#58; # 0 5 9 0 12 0 8 10
2792587 <first &#58; 
2792588 <first &#58; 1/2-1/2 &#40;Stalemate&#41;

Ferdy · Post by **Ferdy** » Tue Oct 18, 2016 6:09 pm

hgm wrote:Interesting. Pulsar sometimes claims stalemate when it gets checkmated:

Code: Select all

&#91;Event "Computer Chess Game"&#93;
&#91;Site "ONTWIKKELLAPTOP"&#93;
&#91;Date "2016.10.18"&#93;
&#91;Round "29"&#93;
&#91;White "Pulsar2009-9b"&#93;
&#91;Black "CrazyWa 0.0"&#93;
&#91;Result "1/2-1/2"&#93;
&#91;TimeControl "40/60"&#93;
&#91;Variant "crazyhouse"&#93;

1. c4 &#123;+0.02/4&#125; Nc6 &#123;-0.98/10 1.7&#125; 2. g3 Rb8 &#123;-0.97/9 1.2&#125; 3. d4
&#123;+0.09/6 1.4&#125; e5 &#123;-0.87/8 1.0&#125; 4. dxe5 &#123;+0.77/5 1.4&#125; Nxe5 &#123;-1.10/7 1.1&#125; 5.
Qd5 &#123;+0.32/4 1.4&#125; f6 &#123;-0.76/6 1.7&#125; 6. Nc3 &#123;+0.13/4 1.4&#125; Bb4 &#123;+0.50/6 1.1&#125;
7. Qd4 &#123;-0.47/4 1.4&#125; c5 &#123;+1.50/6 3&#125; 8. Qd6 &#123;-2.61/4 1.4&#125; Bxc3+
&#123;+1.36/6 1.7&#125; 9. bxc3 &#123;-2.02/5 1.4&#125; N@c2+ &#123;+1.35/5 1.1&#125; 10. Kd1
&#123;-3.92/5 1.5&#125; Nxa1 &#123;+7.09/6 2.9&#125; 11. B@h5+ &#123;-5.20/3 1.4&#125; g6 &#123;+10.11/6 2.7&#125;
12. Bxg6+ &#123;-7.00/4 1.4&#125; Nxg6 &#123;+11.97/6 3&#125; 13. P@f7+ &#123;-12.61/4 1.4&#125; Kxf7
&#123;+14.07/6 1.1&#125; 14. Ke1 &#123;-16.98/4 1.4&#125; P@b2 &#123;+16.67/5 1.2&#125; 15. Bxb2
&#123;-18.15/4 1.4&#125; R@b1+ &#123;+17.16/5 0.9&#125; 16. Kd2 &#123;-18.15/5 1.4&#125; B@h6+
&#123;+16.70/5 5&#125; 17. P@f4 &#123;-15.10/5 1.4&#125; Rxb2+ &#123;+15.48/5 1.8&#125; 18. Kc1
&#123;-16.00/5 1.4&#125; Rxa2 &#123;+16.29/5 3&#125; 19. P@e6+ &#123;-14.67/4 1.4&#125; Ke8
&#123;+22.38/6 2.3&#125; 20. Qxb8 &#123;-14.96/4 1.4&#125; Rc2+ &#123;+79.96/4 2.0&#125; 21. Kb1
&#123;-149.97/4 1.4&#125; Qb6+ &#123;+79.97/9 0.8&#125; 22. R@b5 &#123;-149.98/4 1.4&#125; Qxb5+
&#123;+79.98/77 0.6&#125; 23. cxb5 &#123;-149.99/4 1.4&#125; R@c1#
&#123;+79.99/2 0.1&#125; 1/2-1/2

From its output it seems to do that by removing its King

:

Code: Select all

2792565 >first &#58; R@c1
2792573 <first &#58; #testing crazyhouse evaluate is 764
2792573 <first &#58; # proccessed time
2792574 <first &#58; #testing crazyhouse evaluate is 764
2792574 <first &#58; #testing crazyhouse evaluate is 764
2792574 <first &#58; # in plop move where its written a+4
2792575 <first &#58; # finished call to getmove and turn is 46 and wstat is 2
2792576 <first &#58; #board before search
2792577 <first &#58; # 8 6 10 0 0 3 2 4
2792577 <first &#58; # 0 0 10 0 1 1 0 1
2792577 <first &#58; # 0 0 1 0 0 0 1 0
2792577 <first &#58; # 0 0 0 0 0 1 0 0
2792577 <first &#58; # 0 1 7 0 0 0 0 0
2792578 <first &#58; # 0 0 0 0 1 7 8 9
2792578 <first &#58; # 7 7 0 7 0 0 0 7
2792578 <first &#58; # 0 5 9 0 12 0 8 10
2792578 <first &#58; 1 -15000 0 0 b1b2 
2792581 <first &#58; 
2792581 <first &#58; # maxdepth = 1 
2792581 <first &#58; 2 -15000 0 0 b1b2 p@a2 
2792582 <first &#58; 
2792583 <first &#58; # maxdepth = 2 
2792583 <first &#58; 3 -15000 0 0 b1b2 p@a2 b1a1 
2792584 <first &#58; 
2792584 <first &#58; # maxdepth = 3 
2792584 <first &#58; # board before move
2792585 <first &#58; # 8 6 10 0 0 3 2 4
2792585 <first &#58; # 0 0 10 0 1 1 0 1
2792585 <first &#58; # 0 0 1 0 0 0 1 0
2792585 <first &#58; # 0 0 0 0 0 1 0 0
2792586 <first &#58; # 0 1 7 0 0 0 0 0
2792586 <first &#58; # 0 0 0 0 1 7 8 9
2792586 <first &#58; # 7 7 0 7 0 0 0 7
2792586 <first &#58; # 0 5 9 0 12 0 8 10
2792586 <first &#58; # 8 0 10 0 0 3 2 4
2792586 <first &#58; # 0 0 10 0 1 1 0 1
2792587 <first &#58; # 0 0 1 0 0 0 1 0
2792587 <first &#58; # 0 0 0 0 0 1 0 0
2792587 <first &#58; # 0 1 7 0 0 0 0 0
2792587 <first &#58; # 0 0 0 0 1 7 8 9
2792587 <first &#58; # 7 7 0 7 0 0 0 7
2792587 <first &#58; # 0 5 9 0 12 0 8 10
2792587 <first &#58; 
2792588 <first &#58; 1/2-1/2 &#40;Stalemate&#41;

In that case Pulsar behaves like playing an atomic variant

.

Just checking Pulsar's games in blitz where its rating are based, it got 1 draw by repeat. These games are played in Cute Chess. I am not sure though if Pulsar claimed a stalemate in one of those games and was flagged by Cute Chess as a loss.

Code: Select all

11&#41; Pulsar 2009 9b 32bit      2174.1 &#58;    357 (+147,=1,-209&#41;,  41.3 %

    vs.                              &#58;  games (   +, =,   -),   (%) &#58;    Diff,    SD, CFS (%)
    Imortal 2.9001 32bit             &#58;     32 (   1, 0,  31&#41;,   3.1 &#58;  -677.8,  42.2,    0.0
    TJchess 1.3 64bit                &#58;     16 (   0, 0,  16&#41;,   0.0 &#58;  -617.5,  41.2,    0.0
    Sunsetter 9 32bit                &#58;     16 (   1, 0,  15&#41;,   6.3 &#58;  -541.7,  47.2,    0.0
    Imortal 2.0 32bit                &#58;      9 (   0, 0,   9&#41;,   0.0 &#58;  -509.9,  39.4,    0.0
    Sunsetter 7g 64bit               &#58;     10 (   1, 0,   9&#41;,  10.0 &#58;  -385.8,  35.4,    0.0
    Sjaak II 1.3.1a 64bit            &#58;     48 (  13, 0,  35&#41;,  27.1 &#58;  -152.4,  29.7,    0.0
    Sjeng 11.2 32bit                 &#58;     36 (  15, 0,  21&#41;,  41.7 &#58;  -125.9,  30.3,    0.0
    NebiyuAlien 1.45a 32bit          &#58;     30 (  10, 0,  20&#41;,  33.3 &#58;  -122.4,  33.7,    0.0
    NebiyuAlien 1.45 32bit           &#58;     49 (  27, 0,  22&#41;,  55.1 &#58;   +76.3,  34.0,   98.7
    TSCP 1.1 32bit                   &#58;     64 (  39, 1,  24&#41;,  61.7 &#58;  +139.4,  30.9,  100.0
    KKFChess 2.6.7 beta 32bit        &#58;     47 (  40, 0,   7&#41;,  85.1 &#58;  +181.1,  32.2,  100.0

hgm · Post by **hgm** » Tue Oct 18, 2016 9:58 pm

I'm not sure if it is worth it, (as I hope to have much better versions in a few days), but I have uploaded the first more-or-less usable version of my engine to my website. It beat Pulsar 22.5-17.5 at 40 moves/min, with only one illegal move (which I adjudicated as a win, because the problem was that a mate-in-one move was not always returned by the search, so that it repeated the previous move. But this is now fixed).

The name of the engine is CrazyWa, because eventually it will play all Shogi variants as well, in particular Wa Shogi with drops, a Shogi variant on an 11x11 board. None of the other variants is currently operational, (or at least not tried even once), though.

This version cannot castle (but accepts castling from the opponent), and it also seems that it doesn't realize e.p. capture can be a method of check evasion. It is completely untuned.

URL: http://hgm.nubati.net/dropper.exe

Crazyhouse tournaments and rating list

Re: CCVA RR Tour Nr 15

Re: CCVA RR Tour Nr 15

Re: CCVA RR Tour Nr 15

Re: CCVA RR Tour Nr 15

Re: Crazyhouse tournaments and rating list

Re: Crazyhouse tournaments and rating list

Re: Crazyhouse tournaments and rating list

Re: Crazyhouse tournaments and rating list

Re: Crazyhouse tournaments and rating list

Re: Crazyhouse tournaments and rating list