Discussion of anything and everything relating to chess playing software and machines.
Moderators: hgm , Rebel , chrisw
Henk
Posts: 7216 Joined: Mon May 27, 2013 10:31 am
Post
by Henk » Mon May 22, 2017 4:28 pm
PV length from 16 back to 9. Progress of -7 ply. This is where everybody has been waiting for. Technology improvements go fast these days.
Code: Select all
Depth Value Time(seconds) Nodes
370 0.5800 23.29 21686168 g3f1 b3a5 f1e3 a5c6 e3d5 c6e5 d5e7 e5d3 b2a3
360 0.5800 60.82 48285036 g3f1 b3a5 f1e3 a5c6 e3d5 c6e5 d5e7 e5d3 b2a3
350 0.8400 18.52 15886315 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1f2
340 0.8700 16.30 15095423 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
330 0.8700 6.75 6531825 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
320 0.8700 6.03 4899636 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
310 0.7300 2.68 2668297 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
300 0.7300 2.95 2829587 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
290 0.7700 14.47 12191805 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
280 0.0000 3.95 3762018 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4
270 0.0000 2.96 2895635 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6
260 0.0000 1.23 1323095 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7
250 0.0000 1.09 1132453 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7
240 0.0000 0.62 615247 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4
230 0.0000 0.47 486695 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6
220 0.0000 0.30 270310 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7
210 0.0000 0.28 264342 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7
200 0.0700 1.51 1194084 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4
190 -0.1700 0.13 129474 g1f2 h7h6 e5b8 e7e6 b8a7 g7f8 b2a3 f8g8 f2f3 b3a1
180 -0.1700 0.07 78680 g1f2 h7h6 e5b8 e7e6 b8a7 g7f8 b2a3 f8e8 f2f1 b3a1
170 -0.1700 0.09 79314 g1f2 h7h6 e5b8 e7e6 b8a7 g7f8 b2a3 f8e8 f2f1
160 -0.3000 0.06 61654 g1f2 h7h6 f2e1 g7f8 g3f1 b3a5
150 -0.2300 0.04 39230 g1f2 h7h6 f2e1 g7f8 g3f1
140 -0.3300 0.10 90934 g1f2 h7h6 f2e1 g7f8 g3f1
130 0.1600 0.02 23962 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1f2 d2b1
120 0.1600 0.01 13489 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1f2 d2b1
110 0.2400 0.01 15554 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1g1
100 0.2100 0.00 6017 g3f1 b3a5 f1d2 a5b3
90 0.2000 0.00 4040 g1h1 b3d2 h1g1
80 0.2000 0.00 3323 g3f1 b3a5 f1d2
70 0.2100 0.00 2390 g1f2 b3d2 b2a1
60 0.2100 0.00 1424 g1f1 b3d2 f1e1 d2b1
50 0.1300 0.00 960 g3f1 b3a5
40 0.2100 0.00 478 g1h1 b3d2
30 0.2000 0.01 335 g3h1
20 0.2100 0.00 81 g1f1
10 0.2100 0.05 103 g3h1
Ras
Posts: 2487 Joined: Tue Aug 30, 2016 8:19 pm
Full name: Rasmus Althoff
Post
by Ras » Tue May 23, 2017 2:45 am
Probably because of hash tables, that can cut off a PV. Who cares, it's the move being played that counts.
That said, still wondering whether a micro-controller system can beat Skipper. Can I get an EXE file?
Henk
Posts: 7216 Joined: Mon May 27, 2013 10:31 am
Post
by Henk » Tue May 23, 2017 4:10 pm
No Skipper does no do hash cut offs. For it uses Zobrist keys so it can't trust the hash move anymore because of hash collisions. And I want to avoid spending time on errors which were caused by hash collisions.
Maybe I should go back to an old implementation that did not use Zobrist keys. Good old days.
Code: Select all
Depth Value Time(seconds) Nodes
390 1.2400 103.61 96109296 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 d5d7
380 1.2400 64.99 59604669 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 d5d7
370 1.2400 31.00 25529477 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 d5d7
360 1.1300 15.47 13858138 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 d5d7
350 1.1300 15.42 12184342 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 d5d7
340 0.9200 6.97 6219345 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
330 0.9200 9.74 8787204 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
320 0.9200 8.91 5605485 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3a5 g1h1
310 0.7800 2.59 2331090 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
300 0.7800 2.58 2360391 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
290 0.7800 1.77 1799205 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 d5a8
280 0.9600 1.01 834369 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 g1h1
270 0.9600 0.64 683186 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 g1h1
260 0.9600 0.63 555584 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 g1h1
250 0.6600 1.02 672401 g3f1 h7h6 f1e3 g7h7 e3d5 f6d5 e5d5 b3d2 g1h1
240 0.9200 1.73 1547120 g3f1 h7h6 f1e3 e7d8 e3g4 b3d2
230 0.0000 0.82 619640 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6
220 0.0000 0.47 475877 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7
210 -0.0100 0.48 459909 b2a3 e7a3 e5c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7 f4c7 g7h6 c7f4 h6g7
200 -0.0700 0.60 571915 e5e2 b3a5 g3f1 g7g8 g2g3 f6h5
190 0.1700 0.16 163154 g3e2 g7g8 e5b8 e7e8 b8e8 f6e8 g1f1 b3d2 f1e1 d2b1 h2h3
180 0.0900 0.25 250440 g3f1 h7h6 e5b8 e7e6 b8a7 g7f8 g1h1 b3a5
170 0.0900 0.06 60303 g3f1 h7h6 e5b8 e7e6 b8a7 g7f8 f1g3 b3d2
160 0.0100 0.08 68415 g3f1 g7g8 g2g3 b3a5 f1d2 a5b3
150 0.0700 0.24 131365 g3f1 g7g8 g2g3 b3a5 f1d2
140 0.2200 0.06 52385 g1h1 g7g8 h2h3 e7d8 g3f1
130 0.4400 0.02 14199 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1f2 g6g5
120 0.4400 0.01 10141 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1f2 g6g5
110 0.4000 0.01 14363 e5b8 e7e6 b8a7 g7h6 g1f1 b3d2 f1g1
100 0.4800 0.00 3903 g1h1 b3d2 b2a1 d2b1
90 0.4800 0.00 4093 e5e3 b3a5 g1f1
80 0.4900 0.00 2810 g1f2 b3d2 b2a1
70 0.4800 0.00 2063 g1f1 b3d2 f1e1 d2b1
60 0.4800 0.00 1281 g3f1 b3a5
50 0.4900 0.00 1150 e5e1 b3a5
40 0.4800 0.00 607 g3h1 b3d2
30 0.4900 0.01 278 g1h1
20 0.4800 0.00 83 e5e1
10 0.4900 0.05 75 e5e1
Henk
Posts: 7216 Joined: Mon May 27, 2013 10:31 am
Post
by Henk » Thu Jun 01, 2017 6:20 pm
Skipper not seeing a checkmate.
[d] 3qk3/8/8/4K3/8/8/8/8 w - - 0 1
Code: Select all
Depth Value Time(seconds) Nodes
210 -10.6400 66.25 74955998 e5f4 d8d2 f4g4 e8f7 g4h3 d2f2 h3g4 f7g6
200 -10.7300 7.55 9272677 e5e4 d8d2 e4f5 e8f7 f5e4 f7e6
190 -10.6400 15.22 17638139 e5f5 d8d3 f5e5 e8f7 e5f4 d3d4 f4g3 d4g1 g3f3 g1d1 f3e4 f7e6 e4f4
180 -10.6400 12.26 14485910 e5f4 d8d2 f4g4 e8f7 g4h3 d2d3 h3h2 d3e2 h2h1 e2f1 h1h2 f7g6
170 -10.6200 4.71 5388541 e5f4 d8d2 f4g4 e8f7 g4h3 d2d3 h3h2 d3e2 h2h1 f7g6
160 -10.6000 3.01 3854043 e5f5 d8d3 f5f4 e8e7 f4g4 d3e2 g4h3 e2h5 h3g2 e7f6
150 -10.6100 0.14 114086 e5e4 d8a5 e4e3 a5e1 e3d3 e1d1 d3c4 e8d7
140 -10.5300 0.65 734114 e5f4 e8f7 f4e3 d8g5 e3e2 f7e6
130 -10.6100 0.02 23453 e5e4 d8d2 e4f3 d2d1 f3g2 e8f7
120 -10.6000 0.15 186956 e5f5 d8d4 f5g5 e8f7
110 -10.5900 0.02 25269 e5f4 e8f7 f4e3 f7e6
100 -10.6600 0.05 46065 e5f5 d8d4 f5g5 e8f7
90 -10.5700 0.00 1227 e5e4 d8d2 e4f3
80 -10.6100 0.00 1033 e5e4 d8h4 e4d3 e8d7
70 -10.5400 0.00 1024 e5f4 d8d2 f4g3 e8f7
60 -10.5900 0.00 818 e5f4 e8f7
50 -10.6100 0.00 65 e5e4 e8e7
40 -10.6100 0.00 127 e5e4 e8e7
30 -10.5300 0.00 9 e5e4
20 -10.5300 0.00 9 e5e4
10 -10.4000 0.00 11 e5f4
Daniel Anulliero
Posts: 759 Joined: Fri Jan 04, 2013 4:55 pm
Location: Nice
Post
by Daniel Anulliero » Thu Jun 01, 2017 8:36 pm
What are the knowledges did you gave to poor skipper to solve this QK vs K ?
Isa download :
Henk
Posts: 7216 Joined: Mon May 27, 2013 10:31 am
Post
by Henk » Thu Jun 01, 2017 9:17 pm
Daniel Anulliero wrote: What are the knowledges did you gave to poor skipper to solve this QK vs K ?
Have to find out what he's doing for useless computations and memory allocations.
Kan Isa solve it ?
Daniel Anulliero
Posts: 759 Joined: Fri Jan 04, 2013 4:55 pm
Location: Nice
Post
by Daniel Anulliero » Thu Jun 01, 2017 10:13 pm
Henk wrote: Daniel Anulliero wrote: What are the knowledges did you gave to poor skipper to solve this QK vs K ?
Have to find out what he's doing for useless computations and memory allocations.
Kan Isa solve it ?
Yes
Isa download :
Daniel Anulliero
Posts: 759 Joined: Fri Jan 04, 2013 4:55 pm
Location: Nice
Post
by Daniel Anulliero » Thu Jun 01, 2017 10:31 pm
Daniel Anulliero wrote: Henk wrote: Daniel Anulliero wrote: What are the knowledges did you gave to poor skipper to solve this QK vs K ?
Have to find out what he's doing for useless computations and memory allocations.
Kan Isa solve it ?
Yes
Isa 2.0.45 find a mate in 7 at depth 14 but it don't show the good PV !
Code: Select all
1 -950 0 4 e5f5
2 -964 0 166 e5e4 e8e7
3 -972 0 288 e5e4 d8e7 e4d4 e8d7
4 -977 0 1337 e5e4 d8d6 e4e3 e8d7
5 -982 0 3705 e5e4 d8d6 e4e3 d6e5 e3d3 e8d7
6 -985 0 11115 e5e4 d8d6 e4e3 d6e5 e3d3 e8d7 d3c4
7 -990 0 28053 e5e4 e8d7 e4e5 d8g5 e5e4 d7e6 e4d4 g5d5 d4e3
8 -992 0 64357 e5e4 d8h4 e4e3 h4c4 e3f3 c4d4 f3g3 e8d7 g3f3
9 -996 0 79188 e5e4 d8h4 e4e3 e8d7 e3d3 d7e6 d3e3 e6d5 e3d3 h4d4 d3e2
10 -1003 0 357234 e5e4 e8d7 e4d5 d8h4 d5e5 d7c6 e5e6 h4f4 e6e7 f4e5 e7f7
11 -1010 0 316406 e5e4 e8d7 e4d5 d8h4 d5e5 h4g4 e5d5 g4f4
12 -1045 1 1207584 e5e4 e8f7 e4f3 d8d3 f3f2 d3e4 f2g3 e4e3 g3g2 e3f4 g2h1 f4f3 h1g1 f3e2 g1h1
13 -1057 2 1685162 e5e4 e8f7 e4f3 d8d3 f3f2 f7e6 f2g2 d3e3 g2f1 e3d2 f1g1 d2e2 g1h1 e6d5
14 -99986 3 2089316 e5e4 d8d2 e4f3 e8f7
Isa download :
zenpawn
Posts: 349 Joined: Sat Aug 06, 2016 8:31 pm
Location: United States
Post
by zenpawn » Thu Jun 01, 2017 11:18 pm
Daniel Anulliero wrote:
Isa 2.0.45 find a mate in 7 at depth 14 but it don't show the good PV !
Code: Select all
1 -950 0 4 e5f5
2 -964 0 166 e5e4 e8e7
3 -972 0 288 e5e4 d8e7 e4d4 e8d7
4 -977 0 1337 e5e4 d8d6 e4e3 e8d7
5 -982 0 3705 e5e4 d8d6 e4e3 d6e5 e3d3 e8d7
6 -985 0 11115 e5e4 d8d6 e4e3 d6e5 e3d3 e8d7 d3c4
7 -990 0 28053 e5e4 e8d7 e4e5 d8g5 e5e4 d7e6 e4d4 g5d5 d4e3
8 -992 0 64357 e5e4 d8h4 e4e3 h4c4 e3f3 c4d4 f3g3 e8d7 g3f3
9 -996 0 79188 e5e4 d8h4 e4e3 e8d7 e3d3 d7e6 d3e3 e6d5 e3d3 h4d4 d3e2
10 -1003 0 357234 e5e4 e8d7 e4d5 d8h4 d5e5 d7c6 e5e6 h4f4 e6e7 f4e5 e7f7
11 -1010 0 316406 e5e4 e8d7 e4d5 d8h4 d5e5 h4g4 e5d5 g4f4
12 -1045 1 1207584 e5e4 e8f7 e4f3 d8d3 f3f2 d3e4 f2g3 e4e3 g3g2 e3f4 g2h1 f4f3 h1g1 f3e2 g1h1
13 -1057 2 1685162 e5e4 e8f7 e4f3 d8d3 f3f2 f7e6 f2g2 d3e3 g2f1 e3d2 f1g1 d2e2 g1h1 e6d5
14 -99986 3 2089316 e5e4 d8d2 e4f3 e8f7
Is that 3 in the output really 3 centiseconds?! Impressive. What's your LMR and pruning like? RookieMonster finds it in 13 ply, but it takes 105 seconds and searches considerably more nodes than Isa (61x as many). I've got LMR and null move, but not futility (it always seems to hurt its play).
Anyway, to contribute to the Skipper discussion: Henk, even if it doesn't find mate right away, if you let Skipper play itself from this position, does it make progress in driving the lone king to the edge and certain doom?
zenpawn
Posts: 349 Joined: Sat Aug 06, 2016 8:31 pm
Location: United States
Post
by zenpawn » Thu Jun 01, 2017 11:43 pm
Incidentally, RookieMonster's mating line is e5e4 d8d2 e4f3 e8d7 f3e4 d7e6 e4f3 e6f5 f3g3 d2e2 g3h3 f5f4 h3h4 e2h2