Bug in Crystal ?

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

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Vinvin
Posts: 4949
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Bug in Crystal ?

Post by Vinvin » Thu Jun 04, 2020 10:03 am

Here Crystal give an eval around +0.40 for 73.Rd6+, but when the move is played the eval goes to +96. Even with backward analyze.

[d]8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 73

Code: Select all

FEN: 8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 73

Crystal_20200526:
Found 511 tablebases
...
 23/81-	01:04	 451 939 393	6 988 501	+0,42	73.Rd6+ Ke7
 23/81-	01:05	 454 299 793	6 987 292	+0,32	73.Rd6+ Ke7
 23/81	01:07	 473 682 564	6 976 281	+0,35	73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rh5 Kf8 76.Rd6 Rg7+ 77.Kh4 Ne1 78.Rf6+ Rf7 79.Kg5 Ke7 80.Rxf7+ Kxf7 81.Kf4 Nd3+ 82.Ke4 Nf2+ 83.Kf5 Ke7 84.Rh2 Nd1 85.Re2+ Kf7 86.Re5 Nb2 87.Rb5 Nc4 88.Rb7+ Ke8 89.Rc7 Nb6 90.Ke6 Kd8 91.Rc2 Nc8 92.Rd2+ Kc7 93.Rd1 Nb6 94.Rb1
 24/61-	01:09	 484 130 251	6 965 702	+0,25	73.Rd6+ Ke7
 24/84+	01:16	 527 352 849	6 938 761	+0,35	73.Rd6+
 24/112+	01:19	 551 762 868	6 914 754	+0,50	73.Rd6+
 24/114	02:04	 837 561 957	6 715 807	+0,45	73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rh5 Kf8 76.Rf6+ Ke8 77.Rh8+ Ke7 78.Ra6 Ne5+ 79.Kf5 Nd7 80.Rh7+ Kd8 81.Ra8+ Kc7 82.Ra3 Rb5+ 83.Kg6 Rb7 84.Rh8 Kd6 85.Rd3+ Kc7 86.Rc3+ Kd6 87.Ra3 Kc7 88.Ra1 Rb6+ 89.Kf5 Rb5+ 90.Kf4 Nf6 91.Ra7+ Rb7 92.Ra2 Nd5+ 93.Ke5 Ne7 94.Kd4 Nf5+ 95.Kc5 Nd6 96.Rh7+ Kb8 97.Rxb7+ Nxb7+ 98.Kd5 Nd8 99.Rc2 Kb7 100.Rc3 Kb6
 25/42-	02:06	 847 506 968	6 706 764	+0,34	73.Rd6+ Ke7
 25/83+	02:14	 895 303 527	6 659 056	+0,45	73.Rd6+
 25/104-	02:27	 966 784 443	6 574 841	+0,24	73.Rd6+ Ke7
 25/106+	02:36	1 019 475 461	6 533 088	+0,42	73.Rd6+
 25/107+	02:53	1 127 871 649	6 498 865	+0,70	73.Rd6+
 25/107	03:03	1 195 659 729	6 505 787	+0,33	73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Ra8 Ke6 76.Kg3 Ne5+ 77.Kg2 Kd6 78.Rd8+ Ke7 79.Raa8 Nc6 80.Rd1 Rc5 81.Kg3 Ne5 82.Rb1 Kd7 83.Ra7+ Ke6 84.Rh7 Kf6 85.Rbb7 Kf5 86.Rh5+ Ke6 87.Rbh7 Rc4 88.Rg7
 26/98+	03:13	1 258 683 941	6 488 228	+0,43	73.Rd6+
 26/98+	03:14	1 261 737 501	6 487 716	+0,53	73.Rd6+
 26/98-	03:22	1 313 883 364	6 486 551	+0,23	73.Rd6+ Ke7
 26/114+	03:37	1 407 510 423	6 467 149	+0,46	73.Rd6+
 26/114	04:05	1 585 371 409	6 458 224	+0,37	73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Ra3 Ne5+ 77.Kf5 Nf7 78.Re6+ Kf8 79.Kf6 Nd8 80.Rd6 Rf7+ 81.Kg6 Rg7+ 82.Kf5 Ke7 83.Rd2 Ne6 84.Re3 Rg5+ 85.Ke4 Kf6 86.Rf3+ Ke7 87.Rh3 Nc5+ 88.Kd4 Ne6+ 89.Kc4 Re5 90.Rh7+ Kf6 91.Rdh2 Re4+ 92.Kb3 Nc5+ 93.Ka3 Ke6 94.Rh8 Ra4+ 95.Kb2 Rb4+ 96.Kc3 Re4 97.R8h5 Kd6 98.Rf2 Rg4 99.Rd2+ Kc6 100.Rh6+ Kc7
 27/61+	04:06	1 594 781 191	6 460 396	+0,47	73.Rd6+
 27/61-	04:07	1 602 352 446	6 462 088	+0,27	73.Rd6+ Ke7
 27/108+	04:18	1 675 454 642	6 472 735	+0,42	73.Rd6+

Code: Select all

FEN: 8/8/p2R1k2/R7/6K1/1r1n4/8/8 b - - 1 73

Crystal_20200526:
...
 20/57	00:03	 24 022 461	7 242 225	+96,00	73. ... Ke7 74.Rdxa6 Rc3 75.Rg6 Rc7 76.Rh5 Kf8 77.Rd5 Nb4 78.Rdd6 Nc2 79.Kf4 Na3 80.Rd5 Nc2 81.Rf6+ Kg7 82.Rdd6 Na3 83.Rf5 Nc4 84.Re6 Rf7 85.Rxf7+ Kxf7 86.Rc6 Na3 87.Ke5 Nb1 88.Rc7+ Kg6 89.Ke4 Kf6 90.Kd5
 21/40	00:03	 24 556 726	7 233 203	+96,00	73. ... Ke7
 21/52	00:03	 25 474 970	7 239 263	+96,00	73. ... Ke7 74.Rdxa6 Rc3 75.Rg6 Rc7 76.Rh5 Kf8 77.Rd5 Nb4 78.Rdd6 Nc2 79.Kf4 Na3 80.Rd5 Nc2 81.Re6 Kf7 82.Rdd6 Rd7 83.Ke5 Rxd6 84.Rxd6 Ne3 85.Rf6+ Ke7 86.Rf4 Kd7 87.Rd4+ Kc6 88.Re4 Nd1 89.Rc4+ Kb5 90.Rc1
 22/63	00:04	 30 264 572	7 243 794	+96,00	73. ... Ke7
 22/63	00:04	 30 474 494	7 243 758	+96,00	73. ... Ke7

jhellis3
Posts: 488
Joined: Fri Aug 16, 2013 10:36 pm

Re: Bug in Crystal ?

Post by jhellis3 » Thu Jun 04, 2020 3:34 pm

I am not sure if this is a bug or not, but I would need some more information to draw any conclusions.

One point first though, Rd6 is winning and +96.00 is the correct score for TB win at root with 4 pieces on the board for the losing side. Keeping that score thus should not be a problem. And indeed +96.00 should be the default score for any such position assuming you have the correct TBs installed - in this case 7-man.

If for some reason you don't want Crystal to hold onto scores in backward analysis, you can disable this by enabling clean search. This will clear the hash every time the position changes.

Now for some questions:

Which size TBs are you using?
What is SF eval in both positions?

For my part, I checked the position with both Crystal and Stockfish with 5-man TBs, and got similarly low scores even after capturing the pawn. The problem there is the DTZ (with best resistance) is 75 after the pawn capture, so finding the 5-man TB hit is going to take quite some search I imagine.

In conclusion, I can only see 2 potential issues.
1) If you are using 7m TBs, why the +96.00 score is not shown in the first instance...
2) If you are using 6m TBs, why the score is so low when the DTZ is only 2...

For 1, I could only guess corrupted TB file, maybe some glitch in the GUI, or maybe some underlying issue with SF.
2 is more concerning, but it would help to know how SF behaves in this regard in order to know if it is something Crystal specific.

From what you have reported, I would guess you are using at least partial 7m TBs. Since both positions should be contained in the same file, you should get a score of +96.00 in both instances. And in any TB position in root you should have either 0, -9X.xx+, or +9X.xx+. Any non-zero score with absolute value less than 95 should be impossible.

jhellis3
Posts: 488
Joined: Fri Aug 16, 2013 10:36 pm

Re: Bug in Crystal ?

Post by jhellis3 » Thu Jun 04, 2020 5:44 pm

After a little more research, it appears you have 6m TBs. The file count for 3-4-5 + 6m I found online is 510, but your posted analysis shows 511? Not sure which is correct.

Also, the scores for the positions found via search should be 96.97 for the first fen, 96.98 for the second, and 96.99 with white to move and capture the pawn. After that, you should have a base score of 97.00 since it is a 6 piece position in root with 3 weak side pieces.

peter
Posts: 2270
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: Bug in Crystal ?

Post by peter » Thu Jun 04, 2020 8:16 pm

jhellis3 wrote:
Thu Jun 04, 2020 5:44 pm
After a little more research, it appears you have 6m TBs. The file count for 3-4-5 + 6m I found online is 510, but your posted analysis shows 511? Not sure which is correct.

Also, the scores for the positions found via search should be 96.97 for the first fen, 96.98 for the second, and 96.99 with white to move and capture the pawn. After that, you should have a base score of 97.00 since it is a 6 piece position in root with 3 weak side pieces.
Could it be that Crystal misevaluates the mate in 57- position as for 50 moves rule? After 73.Rd6+ Ke7 74.Rxa6 next move reducing a piece on board is not until 113th at optimal play according to Losomosov 7men:


Here's SF with 6men Syzygys:

8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 1

Analysis by Brainfish 280520 64 POPCNT:

73.Td6+ Kg7 74.Tdxa6 Tb7
+- (154.12) Tiefe: 60/17 00:00:15 423MN, tb=6992771

This output doesn't change anymore for minutes.

As for Crystal at my trials with 6men Syzygys it keeps (almost) drawing eval until 74.Rxa6 is played, then it changes at once:



Analysis by Crystal 270520 64 POPCNT:

74...Tc3 75.Th6 Tc7 76.Tb5 Ke8 77.Thh5 Tc4+ 78.Kf3 Ke7 79.Th6 Sc5 80.Tc6 Se4 81.Tb7+ Kf8 82.Th6 Sg5+ 83.Kg3 Se4+ 84.Kg2 Kg8 85.Tg6+ Kf8 86.Ta6 Tc8 87.Th6 Kg8 88.Te6 Sc5 89.Tg6+ Kf8 90.Tf6+ Kg8 91.Te7 Td8 92.Tf5 Td2+ 93.Kf3 Sd7 94.Te1 Td3+ 95.Ke4 Tg3 96.Kd5 Sb6+ 97.Kd4 Tg4+ 98.Kc5 Sa4+ 99.Kd6 Sb2 100.Te7 Sd3 101.Te3 Tg6+ 102.Kd7 Sb4 103.Te7 Tg1 104.Kd6 Sd3 105.Tf6 Tg4 106.Tc7 Te4 107.Tg6+ Kf8 108.Tgg7 Sf4 109.Tcf7+ Ke8 110.Tf6 Td4+ 111.Ke5
+- (97.00) Tiefe: 32/107 00:01:54 1695MN, tb=83694846

One ply backward without clearing hash:



Analysis by Crystal 270520 64 POPCNT:

74.Tdxa6 Tc3 75.Th6 Tc7 76.Tb5 Ke8 77.Thh5 Tf7 78.Kh4 Kf8 79.Tb8+ Ke7 80.Tb6 Kf8 81.Td6 Sb2 82.Tc5 Ke8 83.Tc2 Tb7 84.Kg5 Ke7 85.Td4 Ke6 86.Kf4 Tb6 87.Th2 Tb5 88.Th6+ Ke7 89.Ke4 Tg5 90.Tdd6 Sa4 91.Tc6 Kd7 92.Kd4 Th5 93.Thd6+ Ke7 94.Tg6 Kd7 95.Kc4 Sb2+ 96.Kb3 Tb5+ 97.Ka3 Ke8
+/- (1.28) Tiefe: 31/106 00:00:56 890MN, tb=33859275

And after full backward to starting position:

8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 1

Analysis by Crystal 270520 64 POPCNT:

73.Td6+ Ke7 74.Tdxa6 Tc3 75.Th6 Tc7 76.Tb5 Tc4+ 77.Kf3 Sc5 78.Tc6 Se4 79.Tb7+ Kf8 80.Ta6 Sg5+ 81.Kg3 Se4+ 82.Kg2 Tc8 83.Th6 Kg8 84.Te6 Sc5 85.Tg6+ Kf8 86.Tf6+ Kg8 87.Te7 Td8 88.Tf5 Td2+ 89.Kf3 Sd7 90.Te1 Td3+ 91.Ke4 Tg3 92.Kd5 Sb6+ 93.Kd4 Tg4+ 94.Kc5 Sa4+ 95.Kd6 Sc3 96.Ke5 Kg7 97.Te3 Sa4 98.Tff3 Tg5+ 99.Kd6 Sc5 100.Tf4 Kh7 101.Kc6 Sa6 102.Te7+ Kh6 103.Te6+ Kh5 104.Tc4 Tg7 105.Tee4 Tg2 106.Kb6
+/- (1.28) Tiefe: 32/95 00:01:04 1047MN, tb=38288164

Sorry for german abbrevations in output- lines.
Peter.

Vinvin
Posts: 4949
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Bug in Crystal ?

Post by Vinvin » Thu Jun 04, 2020 8:49 pm

OK, as I wasn't clear enough :

I found 2 bugs

1) the main bug
(EGTB : I have all the 6-pieces installed + one 7-piece : KRppKRp)
Here, Crystal gives an eval around +0.40 for 73.Rd6+ but it's clear the position after 73.Rd6+ is completely lost.

[d]8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 73

Code: Select all

FEN: 8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 73

Crystal_20200526:
Found 511 tablebases
...
 25/106+	02:36	1 019 475 461	6 533 088	+0,42	73.Rd6+
 25/107+	02:53	1 127 871 649	6 498 865	+0,70	73.Rd6+
 25/107	03:03	1 195 659 729	6 505 787	+0,33	73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Ra8 Ke6 76.Kg3 Ne5+ 77.Kg2 Kd6 78.Rd8+ Ke7 79.Raa8 Nc6 80.Rd1 Rc5 81.Kg3 Ne5 82.Rb1 Kd7 83.Ra7+ Ke6 84.Rh7 Kf6 85.Rbb7 Kf5 86.Rh5+ Ke6 87.Rbh7 Rc4 88.Rg7
 26/98+	03:13	1 258 683 941	6 488 228	+0,43	73.Rd6+
 26/98+	03:14	1 261 737 501	6 487 716	+0,53	73.Rd6+
 26/98-	03:22	1 313 883 364	6 486 551	+0,23	73.Rd6+ Ke7
 26/114+	03:37	1 407 510 423	6 467 149	+0,46	73.Rd6+
 26/114	04:05	1 585 371 409	6 458 224	+0,37	73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Ra3 Ne5+ 77.Kf5 Nf7 78.Re6+ Kf8 79.Kf6 Nd8 80.Rd6 Rf7+ 81.Kg6 Rg7+ 82.Kf5 Ke7 83.Rd2 Ne6 84.Re3 Rg5+ 85.Ke4 Kf6 86.Rf3+ Ke7 87.Rh3 Nc5+ 88.Kd4 Ne6+ 89.Kc4 Re5 90.Rh7+ Kf6 91.Rdh2 Re4+ 92.Kb3 Nc5+ 93.Ka3 Ke6 94.Rh8 Ra4+ 95.Kb2 Rb4+ 96.Kc3 Re4 97.R8h5 Kd6 98.Rf2 Rg4 99.Rd2+ Kc6 100.Rh6+ Kc7
 27/61+	04:06	1 594 781 191	6 460 396	+0,47	73.Rd6+
 27/61-	04:07	1 602 352 446	6 462 088	+0,27	73.Rd6+ Ke7
 27/108+	04:18	1 675 454 642	6 472 735	+0,42	73.Rd6+

2) learning (via hashtable) bug :
When I play 73.Rd6+ on the board, the score rise to +96,00
and then I go back the previous move and the score go back to +0,40 (instead of staying at +96)

jhellis3
Posts: 488
Joined: Fri Aug 16, 2013 10:36 pm

Re: Bug in Crystal ?

Post by jhellis3 » Thu Jun 04, 2020 9:17 pm

I don't think the issue has anything to do with the hash since you have a 7 man file. It should show the correct result regardless in that case. Though I wonder if not having both DTZ and WDL are causing an issue there.

Regardless, the search to a 6 piece position from there should be trivial... the capture is only 3 ply away, so the fact that the score is not jumping to 96.97 ... 96.98... 96.99 is very suspicious.

Can you try this version of Crystal and report results? https://www.dropbox.com/s/d4yqe4g7x7ric ... l.exe?dl=0

Could it be that Crystal misevaluates the mate in 57- position as for 50 moves rule? After 73.Rd6+ Ke7 74.Rxa6 next move reducing a piece on board is not until 113th at optimal play according to Losomosov 7men
Not really, syzygy reports DTZ of 75, the line posted is less than 80... Crystal uses the rule50 count for the cutoff, so it can't be a hard cutoff (doesn't happen until 100+). The correct moves would be played and the base score of +97.00 would remain intact regardless as root is TB.

peter
Posts: 2270
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: Bug in Crystal ?

Post by peter » Thu Jun 04, 2020 10:17 pm

jhellis3 wrote:
Thu Jun 04, 2020 9:17 pm
Can you try this version of Crystal and report results? https://www.dropbox.com/s/d4yqe4g7x7ric ... l.exe?dl=0

8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 1

Analysis by Crystal 040620 64 POPCNT:

Code: Select all

73.Td6+ Ke7 74.Th6 Tb5 75.Taxa6 Td5 76.Kh4 Td4+ 77.Kg3 
  +/-  (0.85)   Tiefe: 7/14   00:00:00  212kN, tb=4980
73.Td6+ Ke7 74.Tdxa6 Tb4+ 75.Kf3 Tb7 76.Th5 Kf8 77.Ke4 Sb4 78.Tf6+ Kg7 79.Thh6 
  +/-  (0.84)   Tiefe: 8/38   00:00:00  4308kN, tb=146714
73.Td6+ Ke7 74.Tdxa6 Tb4+ 75.Kf3 Tb7 76.Th5 Kf8 77.Ke4 Sb4 78.Tf6+ Kg7 79.Thh6 Te7+ 80.Kf5 Sd5 81.Thg6+ Kh7 82.Td6 
  +/-  (0.94)   Tiefe: 9/38   00:00:00  4558kN, tb=155906
...
73.Td6+ 
  +-  (97.97 ++)   Tiefe: 14/4   00:00:04  57335kN, tb=2209601
73.Td6+ Ke7 
  +/=  (0.40 --)   Tiefe: 14/62   00:00:04  65707kN, tb=2491544
73.Td6+ 
  +-  (97.97 ++)   Tiefe: 14/62   00:00:04  65708kN, tb=2491576
73.Td6+ Ke7 74.Tdxa6 Tc3 75.Th6 Tc7 76.Tg5 Tc4+ 77.Kf3 Kf8 78.Tf5+ Ke7 79.Ta5 Tc7 80.Ke3 Sb4 81.Tb5 Sc6 82.Tc5 Kd7 83.Tc2 Se7 84.Td2+ Ke8 85.Tf2 Tc6 86.Th8+ Kd7 87.Th7 Tg6 88.Kd4 Kd6 89.Kd3 Sd5 
  +/=  (0.46)   Tiefe: 14/102   00:00:05  73587kN, tb=2754533
73.Td6+ 
  +-  (97.97 ++)   Tiefe: 15/4   00:00:05  73588kN, tb=2754567
73.Td6+ Ke7 
  +/=  (0.36 --)   Tiefe: 15/80   00:00:06  98453kN, tb=3591258
...
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 27/42   00:01:52  1545MN, tb=76073618
73.Rd6+ 
  =  (0.10 ++)   Depth: 28/4   00:01:54  1583MN, tb=78212705
73.Rd6+ 
  =  (0.20 ++)   Depth: 28/54   00:01:55  1596MN, tb=78835809
73.Rd6+ 
  +/=  (0.35 ++)   Depth: 28/54   00:01:55  1596MN, tb=78836019
73.Rd6+ 
  +/=  (0.56 ++)   Depth: 28/62   00:01:58  1636MN, tb=81051982
73.Rd6+ 
  +/-  (0.84 ++)   Depth: 28/62   00:02:01  1678MN, tb=83262808
73.Rd6+ 
  +/-  (1.21 ++)   Depth: 28/62   00:02:10  1793MN, tb=89394176
73.Rd6+ 
  +-  (1.70 ++)   Depth: 28/62   00:02:12  1816MN, tb=90622683
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 28/114   00:02:45  2260MN, tb=114061166
...
73.Rd6+ 
  =  (0.10 ++)   Depth: 30/4   00:02:45  2261MN, tb=114120945
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 30/70   00:02:46  2263MN, tb=114191728
...
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Ke8 77.Rhh5 Rc4+ 78.Kf3 Ke7 79.Rh6 Nc5 80.Rc6 Ne4 81.Rb7+ Kf8 82.Rh6 Ng5+ 83.Kg3 Ne4+ 
  =  (0.00)   Depth: 43/42   00:10:19  8397MN, tb=402987809
Going to position after 74.Rdxa6 brings up 97.00 again, but just one ply backwards with full hash it's back to 0.00.

Gave it another try with empty hash again, also because at the upper one I had (as I'm used to) Dynamic Contempt Both On, even if Contempt itself was 0, so I did set everything back to default this time:

Code: Select all

8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 1

Analysis by Crystal 040620 64 POPCNT:

73.Rd6+ Kg7 74.Rf5 Rc3 75.Kg5 Nb4 76.Rg6+ Kh8 77.Kg4 Kh7 
  +/-  (1.15)   Depth: 7/21   00:00:12  360kN, tb=10993
73.Rd6+ Ke7 74.Rdxa6 Nf2+ 75.Kf4 Nd3+ 76.Ke4 Nf2+ 77.Ke5 Rb7 78.Re6+ Kd7 79.Kf5 Rc7 80.Rd5+ Kc8 81.Re8+ Kb7 82.Rb5+ Kc6 83.Ra5 Kd7 
  +/-  (1.03)   Depth: 8/32   00:00:12  2211kN, tb=71943
73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rh6 Kf8 76.Rd5 Kg7 77.Rc6 Nf2+ 78.Kg3 Rf7 79.Rd4 
  +/-  (0.78)   Depth: 9/42   00:00:12  7221kN, tb=224452
73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rg6 Ke8 76.Kg5 Rd7 77.Rf6 Ke7 78.Rf3 Ke6 79.Raf5 Ne5 80.Re3 Rg7+ 81.Kf4 
  +/-  (0.83)   Depth: 10/46   00:00:13  11748kN, tb=369271
73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rh6 Kf8 76.Rd5 Nb4 77.Rd8+ Ke7 78.Rdh8 Rd7 79.Kf5 Rd5+ 80.Kf4 Rd4+ 81.Ke3 Rd7 82.Ke2 Nc2 83.R6h7+ Kd6 
  +/-  (0.80)   Depth: 11/46   00:00:13  15497kN, tb=494966
73.Rd6+ Ke7 74.Rdxa6 Rb7 75.Rh6 Kf8 76.Rd5 Nf2+ 77.Kf5 Kg7 78.Rg6+ Kf7 79.Rc6 Kg7 80.Ke6 Rf7 81.Rf5 Rxf5 82.Kxf5 Nh3 83.Kg4 Nf2+ 84.Kf4 Nd3+ 85.Kf5 Kh7 
  +/-  (0.71)   Depth: 12/46   00:00:13  18221kN, tb=598950
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Kg3 Kf8 77.Rh4 Nc5 78.Ra8+ Ke7 79.Rh6 Ne6 80.Ra3 Nc5 81.Rh1 Kf8 82.Rf1+ Rf7 83.Rf4 Rxf4 
  +/=  (0.56)   Depth: 13/58   00:00:13  27198kN, tb=942067
73.Rd6+ 
  +/-  (0.74 ++)   Depth: 14/4   00:00:15  51900kN, tb=1841309
73.Rd6+ Ke7 
  +/=  (0.46 --)   Depth: 14/29   00:00:15  51923kN, tb=1842159
73.Rd6+ 
  +/=  (0.65 ++)   Depth: 14/41   00:00:15  55700kN, tb=1984260
73.Rd6+ 
  +/-  (0.86 ++)   Depth: 14/41   00:00:15  55701kN, tb=1984327
...
73.Rd6+ 
  +-  (97.97 ++)   Depth: 24/53   00:01:27  1126MN, tb=43101814
73.Rd6+ Kf7 
  +-  (64.36 --)   Depth: 24/81   00:01:28  1139MN, tb=43501488
73.Rd6+ 
  +-  (97.24 ++)   Depth: 24/81   00:01:28  1144MN, tb=43656567
73.Rd6+ Kf7 74.Kh4 Rb4+ 75.Kg5 Ne1 76.Rf6+ Kg7 77.Raf5 Rb7 78.Re6 Ng2 79.Rg6+ Kh8 80.Rf3 a5 81.Rh3+ Rh7 82.Rxh7+ Kxh7 
  +-  (97.80)   Depth: 24/119   00:01:35  1244MN, tb=47062914
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1247MN, tb=47168950
73.Rd6+ Ke7 
  +-  (97.58 --)   Depth: 25/32   00:01:35  1247MN, tb=47184296
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1247MN, tb=47184319
73.Rd6+ Ke7 
  +-  (65.87 --)   Depth: 25/32   00:01:35  1248MN, tb=47187780
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47187799
73.Rd6+ Ke7 
  +-  (63.35 --)   Depth: 25/32   00:01:35  1248MN, tb=47190899
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47190916
73.Rd6+ Ke7 
  +-  (62.61 --)   Depth: 25/32   00:01:35  1248MN, tb=47192865
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47192893
73.Rd6+ Ke7 
  +-  (61.64 --)   Depth: 25/32   00:01:35  1248MN, tb=47195242
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47195270
73.Rd6+ Ke7 
  +-  (55.44 --)   Depth: 25/32   00:01:35  1248MN, tb=47198249
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47198265
73.Rd6+ Ke7 
  +-  (16.25 --)   Depth: 25/32   00:01:35  1248MN, tb=47199893
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47199924
73.Rd6+ Ke7 
  +-  (9.95 --)   Depth: 25/32   00:01:35  1248MN, tb=47205891
73.Rd6+ 
  +-  (97.97 ++)   Depth: 25/32   00:01:35  1248MN, tb=47206231
73.Rd6+ Ke7 
  +-  (6.63 --)   Depth: 25/32   00:01:35  1248MN, tb=47211032
...
73.Rd6+ 
  +-  (49.33 ++)   Depth: 26/116   00:02:00  1613MN, tb=60839523
73.Rd6+ Ke7 
  =  (0.24 --)   Depth: 26/116   00:02:01  1629MN, tb=61411084
73.Rd6+ 
  +-  (49.33 ++)   Depth: 26/116   00:02:01  1629MN, tb=61411103
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rg5 Rc1 77.Rg7+ Kf8 78.Ra7 Ne5+ 79.Kf5 Nf7 80.Ra8+ Kg7 81.Rg6+ Kh7 82.Rg3 Nh6+ 83.Kg5 Rc7 84.Kf6 Rd7 85.Ra5 Rf7+ 86.Ke6 Rg7 87.Rxg7+ 
  =  (0.00)   Depth: 26/116   00:02:09  1736MN, tb=65476872
73.Rd6+ 
  +-  (97.65 ++)   Depth: 27/4   00:02:09  1736MN, tb=65477037
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rg5 Rc1 77.Rg7+ Kf8 78.Ra7 Ne5+ 79.Kf5 Nf7 80.Ra8+ Kg7 81.Rg6+ Kh7 82.Rg3 Nh6+ 83.Kg5 Rc7 84.Kf6 Rd7 85.Ra5 Rf7+ 86.Ke6 Rg7 87.Rxg7+ 
  =  (0.00)   Depth: 27/84   00:02:10  1756MN, tb=66186078
73.Rd6+ 
  +-  (97.64 ++)   Depth: 28/4   00:02:10  1756MN, tb=66186176
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rg6 Rd4 81.Ra6 Ng5+ 82.Kg2 Rg4+ 83.Kf1 Rf4+ 84.Ke1 Re4+ 85.Kd2 Re6 86.Rxe6 
  =  (0.00)   Depth: 28/28   00:02:11  1760MN, tb=66359966
73.Rd6+ 
  +-  (97.63 ++)   Depth: 29/4   00:02:11  1760MN, tb=66360061
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rg6 Rd4 81.Ra6 Ng5+ 82.Kg2 Rg4+ 83.Kf1 Rf4+ 84.Ke1 Re4+ 85.Kd2 Re6 86.Rxe6 
  =  (0.00)   Depth: 29/28   00:02:13  1794MN, tb=67740293
73.Rd6+ 
  +-  (97.62 ++)   Depth: 30/4   00:02:13  1797MN, tb=67900065
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 30/87   00:02:21  1903MN, tb=72170844
73.Rd6+ 
  =  (0.10 ++)   Depth: 31/61   00:02:33  2070MN, tb=80723116
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 31/70   00:02:39  2144MN, tb=84788300
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rh6 Rc7 76.Rb5 Rc4+ 77.Kf3 Nc5 78.Rc6 Ne4 79.Rb7+ Kf8 80.Rh6 Ng5+ 81.Kg3 Ne4+ 
  =  (0.00)   Depth: 32/53   00:02:44  2210MN, tb=88704708
73.Rd6+ 
  =  (0.10 ++)   Depth: 33/78   00:03:02  2459MN, tb=101840117
73.Rd6+ 
  =  (0.20 ++)   Depth: 33/78   00:03:09  2547MN, tb=106713892
73.Rd6+ 
  +/=  (0.35 ++)   Depth: 33/78   00:03:12  2592MN, tb=109188158
73.Rd6+ 
  +/=  (0.56 ++)   Depth: 33/101   00:05:18  4240MN, tb=195927196
73.Rd6+ 
  +/-  (0.84 ++)   Depth: 33/101   00:05:18  4240MN, tb=195934125
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rb6 Rc4+ 76.Kg3 Nc5 77.Ra7+ Nd7 78.Rb3 Rd4 79.Re3+ Kd6 80.Rxd7+ 
  =  (0.00)   Depth: 33/101   00:05:21  4270MN, tb=197573876
73.Rd6+ 
  +/-  (0.84 ++)   Depth: 34/4   00:05:21  4270MN, tb=197574058
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rb6 Rc4+ 76.Kg3 Nc5 77.Ra7+ Nd7 78.Rb3 Rd4 79.Re3+ Kd6 80.Rxd7+ 
  =  (0.00)   Depth: 34/16   00:07:55  6280MN, tb=311067695
73.Rd6+ 
  =  (0.10 ++)   Depth: 35/4   00:07:55  6280MN, tb=311067830
73.Rd6+ Ke7 74.Rdxa6 Rc3 75.Rb6 Rc4+ 76.Kg3 Nc5 77.Ra7+ Nd7 78.Rb3 Rd4 79.Re3+ Kd6 80.Rxd7+ 
  =  (0.00)   Depth: 35/16   00:14:42  11512MN, tb=612117817
Little difference as for eval and output- lines over time, but with same result.
Peter.

jhellis3
Posts: 488
Joined: Fri Aug 16, 2013 10:36 pm

Re: Bug in Crystal ?

Post by jhellis3 » Thu Jun 04, 2020 11:09 pm

Well, this is certainly interesting...

It shows the correct & expected score of 97.97 at depths 14 and 15. Then the score gets overwritten. This can happen and is allowed, but when overwritten by a lower score, the TBs should activate once again and keep overwriting any non winning score.

This is quite odd.... the only thing I can think of is that the best move is the only move (which it is in this case) and also the PV move (which it is), and thus gets searched regardless, and the searched score gets reported.

Under previous code, this could happen resulting in an unusually low score from time to time. However, because the moves must be the PV in order for this to occur, there was no real harm as the engine would (and still does) follow the optimal path. If it somehow switched off the optimal path, it would be corrected immediately via hash.

However, I recently added code, which was supposed to correct the visually anomalous low scores. And I can not see any reason for that to not currently be the case.

TLDR: Optimal play should not be an issue, since the only time this can happen is if the best move is already the PV move. Irregularly low score should not appear unless PV path is *only* winning path. Irregularly low score shouldn't appear at all now, or at the most only momentarily. Will investigate further...

User avatar
MikeB
Posts: 4742
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Bug in Crystal ?

Post by MikeB » Fri Jun 05, 2020 6:34 am

jhellis3 wrote:
Thu Jun 04, 2020 11:09 pm
Well, this is certainly interesting...

It shows the correct & expected score of 97.97 at depths 14 and 15. Then the score gets overwritten. This can happen and is allowed, but when overwritten by a lower score, the TBs should activate once again and keep overwriting any non winning score.

This is quite odd.... the only thing I can think of is that the best move is the only move (which it is in this case) and also the PV move (which it is), and thus gets searched regardless, and the searched score gets reported.

Under previous code, this could happen resulting in an unusually low score from time to time. However, because the moves must be the PV in order for this to occur, there was no real harm as the engine would (and still does) follow the optimal path. If it somehow switched off the optimal path, it would be corrected immediately via hash.

However, I recently added code, which was supposed to correct the visually anomalous low scores. And I can not see any reason for that to not currently be the case.

TLDR: Optimal play should not be an issue, since the only time this can happen is if the best move is already the PV move. Irregularly low score should not appear unless PV path is *only* winning path. Irregularly low score shouldn't appear at all now, or at the most only momentarily. Will investigate further...

Joe , it does appear to be a bug of some sorts. Stockfish does not behave this way. Black-Diamond, based on Crystal does ( behaves same as Crystal).
I used 6 man EGTB - 510 files.

This how the functional equivalent SF behaves:

$ Stockfish-XI-r2
Stockfish XI-r2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
p f 8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 73
s t 60
info string Hash Table: Windows Large Pages, 256 Mb
Confirmation: Threads set to 60
s h 65536
info string Hash Table: Windows Large Pages, 65536 Mb
Confirmation: Hash set to 65536 Mb
s z c:/syzygy
info string Found 510 tablebases
Confirmation: SyzygyPath set to c:/syzygy
g i
info depth 1 seldepth 2 multipv 1 score cp 63 nodes 1666 nps 555333 tbhits 18 time 3 pv a5f5 f6e6
info depth 2 seldepth 4 multipv 1 score cp 63 nodes 19071 nps 3814200 tbhits 374 time 5 pv a5f5 f6e6 d4e4 e6d7
info depth 3 seldepth 6 multipv 1 score cp 9999 nodes 22144 nps 4428800 tbhits 474 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 4 seldepth 6 multipv 1 score cp 9999 nodes 25737 nps 5147400 tbhits 569 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 5 seldepth 7 multipv 1 score cp 9999 nodes 27265 nps 5453000 tbhits 607 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 6 seldepth 8 multipv 1 score cp 9999 nodes 28428 nps 5685600 tbhits 638 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 7 seldepth 9 multipv 1 score cp 9999 nodes 29409 nps 5881800 tbhits 667 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 8 seldepth 9 multipv 1 score cp 9999 nodes 30138 nps 6027600 tbhits 687 time 5 pv d4d6 f6e7 d6a6 b3b4
info depth 9 seldepth 13 multipv 1 score cp 9999 nodes 94414 nps 11801750 tbhits 2714 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 10 seldepth 6 multipv 1 score cp 9999 nodes 98836 nps 12354500 tbhits 2832 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 11 seldepth 6 multipv 1 score cp 9999 nodes 101123 nps 12640375 tbhits 2946 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 12 seldepth 6 multipv 1 score cp 9999 nodes 103412 nps 12926500 tbhits 3024 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 13 seldepth 6 multipv 1 score cp 9999 nodes 105853 nps 13231625 tbhits 3093 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 14 seldepth 6 multipv 1 score cp 9999 nodes 108689 nps 13586125 tbhits 3189 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 15 seldepth 6 multipv 1 score cp 9999 nodes 111901 nps 13987625 tbhits 3265 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 16 seldepth 6 multipv 1 score cp 9999 nodes 115911 nps 14488875 tbhits 3367 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 17 seldepth 6 multipv 1 score cp 9999 nodes 119950 nps 14993750 tbhits 3509 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 18 seldepth 6 multipv 1 score cp 9999 nodes 123981 nps 15497625 tbhits 3608 time 8 pv d4d6 f6e7 d6a6 b3b4
info depth 19 seldepth 6 multipv 1 score cp 9999 nodes 128853 nps 14317000 tbhits 3716 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 20 seldepth 6 multipv 1 score cp 9999 nodes 133284 nps 14809333 tbhits 3839 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 21 seldepth 6 multipv 1 score cp 9999 nodes 137281 nps 15253444 tbhits 3960 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 22 seldepth 6 multipv 1 score cp 9999 nodes 142861 nps 15873444 tbhits 4133 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 23 seldepth 6 multipv 1 score cp 9999 nodes 147199 nps 16355444 tbhits 4262 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 24 seldepth 6 multipv 1 score cp 9999 nodes 154245 nps 17138333 tbhits 4592 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 25 seldepth 6 multipv 1 score cp 9999 nodes 157199 nps 17466555 tbhits 4698 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 26 seldepth 6 multipv 1 score cp 9999 nodes 162801 nps 18089000 tbhits 4881 time 9 pv d4d6 f6e7 d6a6 b3b4
info depth 27 seldepth 28 multipv 1 score cp 9999 nodes 512966 nps 42747166 tbhits 22544 time 12 pv d4d6 f6e7 d6a6 b3b4
info depth 28 seldepth 8 multipv 1 score cp 9999 nodes 537540 nps 41349230 tbhits 23833 time 13 pv d4d6 f6e7 d6a6 b3b4
info depth 29 seldepth 8 multipv 1 score cp 9999 nodes 643960 nps 45997142 tbhits 29801 time 14 pv d4d6 f6e7 d6a6 b3b4
info depth 30 seldepth 39 multipv 1 score cp 9999 nodes 1959499 nps 75365346 tbhits 102372 time 26 pv d4d6 f6e7 d6a6 b3b4
info depth 31 seldepth 6 multipv 1 score cp 9999 nodes 1969794 nps 75761307 tbhits 102844 time 26 pv d4d6 f6e7 d6a6 b3b4
info depth 32 seldepth 6 multipv 1 score cp 9999 nodes 1986236 nps 73564296 tbhits 103718 time 27 pv d4d6 f6e7 d6a6 b3b4
info depth 33 seldepth 6 multipv 1 score cp 9999 nodes 2021694 nps 74877555 tbhits 105320 time 27 pv d4d6 f6e7 d6a6 b3b4
info depth 34 seldepth 10 multipv 1 score cp 9999 nodes 2670234 nps 80916181 tbhits 140024 time 33 pv d4d6 f6e7 d6a6 b3b4
info depth 35 seldepth 37 multipv 1 score cp 9999 nodes 3301334 nps 84649589 tbhits 170903 time 39 pv d4d6 f6e7 d6a6 b3b4
info depth 36 seldepth 9 multipv 1 score cp 9999 nodes 3398475 nps 84961875 tbhits 176923 time 40 pv d4d6 f6e7 d6a6 b3b4
info depth 37 seldepth 9 multipv 1 score cp 9999 nodes 4316069 nps 88083040 tbhits 230154 time 49 pv d4d6 f6e7 d6a6 b3b4
info depth 38 seldepth 9 multipv 1 score cp 9999 nodes 4593914 nps 88344500 tbhits 245915 time 52 pv d4d6 f6e7 d6a6 b3b4
info depth 39 seldepth 9 multipv 1 score cp 9999 nodes 6090151 nps 90897776 tbhits 325118 time 67 pv d4d6 f6e7 d6a6 b3b4
info depth 40 seldepth 9 multipv 1 score cp 9999 nodes 6130307 nps 91497119 tbhits 327229 time 67 pv d4d6 f6e7 d6a6 b3b4
info depth 41 seldepth 8 multipv 1 score cp 9999 nodes 6495325 nps 91483450 tbhits 348022 time 71 pv d4d6 f6e7 d6a6 b3b4
info depth 42 seldepth 19 multipv 1 score cp 9999 nodes 8623731 nps 93736206 tbhits 466336 time 92 pv d4d6 f6e7 d6a6 b3b4
info depth 43 seldepth 41 multipv 1 score cp 9999 nodes 11558766 nps 96323050 tbhits 614938 time 120 pv d4d6 f6e7 d6a6 b3b4
info depth 44 seldepth 36 multipv 1 score cp 9999 nodes 12570554 nps 96696569 tbhits 663789 time 130 pv d4d6 f6e7 d6a6 b3b4
info depth 45 seldepth 25 multipv 1 score cp 9999 nodes 13281712 nps 96946802 tbhits 699618 time 137 pv d4d6 f6e7 d6a6 b3b4
info depth 46 seldepth 48 multipv 1 score cp 9999 nodes 16366933 nps 97422220 tbhits 859036 time 168 pv d4d6 f6e7 d6a6 b3b4
info depth 47 seldepth 24 multipv 1 score cp 9999 nodes 17414491 nps 97834219 tbhits 913633 time 178 pv d4d6 f6e7 d6a6 b3b4
info depth 48 seldepth 38 multipv 1 score cp 9999 nodes 23928610 nps 98878553 tbhits 1245898 time 242 pv d4d6 f6e7 d6a6 b3b4
info depth 49 seldepth 9 multipv 1 score cp 9999 nodes 28438956 nps 99436909 tbhits 1472553 time 286 pv d4d6 f6e7 d6a6 b3b4
info depth 50 seldepth 25 multipv 1 score cp 9999 nodes 32581361 nps 99637189 tbhits 1667449 time 327 pv d4d6 f6e7 d6a6 b3b4
info depth 51 seldepth 48 multipv 1 score cp 9999 nodes 52960716 nps 101847530 tbhits 2658786 time 520 pv d4d6 f6e7 d6a6 b3b4
info depth 52 seldepth 14 multipv 1 score cp 9999 nodes 60602981 nps 102369900 tbhits 3036412 time 592 pv d4d6 f6e7 d6a6 b3b4
?
info depth 53 seldepth 14 multipv 1 score cp 9999 nodes 872961008 nps 111731858 hashfull 56 tbhits 33237315 time 7813 pv d4d6 f6e7 d6a6 b3b4
bestmove d4d6 ponder f6e7


For Black-Diamond, it goes back to at least 2-14-20 ( or prior)
$ Black-Diamond-021420.exe
Black Diamond 021420 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
info string Hash LargePages 16 Mb
p f 8/8/p4k2/R7/3R2K1/1r1n4/8/8 w - - 0 7
s syzygypath c:/syzygy
info string Found 510 tablebases
Confirmation: syzygypath set to c:/syzygy
s threads 60
Confirmation: threads set to 60
s hash 65536
info string Hash LargePages 65536 Mb
Confirmation: hash set to 65536
g i
info depth 1 seldepth 2 multipv 1 score cp 55 nodes 10893 nps 3631000 tbhits 74 time 3 pv a5f5 f6e6
info depth 2 seldepth 4 multipv 1 score cp 55 nodes 16500 nps 3300000 tbhits 114 time 5 pv a5f5 f6e6 d4e4 e6d7
info depth 3 seldepth 5 multipv 1 score cp 53 nodes 75536 nps 9442000 tbhits 1150 time 8 pv g4h5 d3e5 d4f4 f6e6 a5a6 e6d7
info depth 4 seldepth 5 multipv 1 score cp 53 nodes 80330 nps 10041250 tbhits 1308 time 8 pv d4d6 f6e7 d6a6 b3b1 a5d5
info depth 5 seldepth 8 multipv 1 score cp 50 nodes 85535 nps 10691875 tbhits 1405 time 8 pv g4h5 d3e5 d4f4 f6e6 f4e4 b3b5
info depth 6 seldepth 10 multipv 1 score cp 60 nodes 93394 nps 10377111 tbhits 1669 time 9 pv d4d6 f6e7 d6a6 b3b8 a5d5 b8g8 g4f5 g8f8 f5e4
info depth 7 seldepth 20 multipv 1 score cp 54 nodes 1124039 nps 62446611 tbhits 26374 time 18 pv d4d6 f6g7 a5g5 g7f7 g5f5 f7e7 d6d5 d3b4 d5e5 e7d6 e5e4
info depth 8 seldepth 20 multipv 1 score cp 67 nodes 1563944 nps 71088363 tbhits 37860 time 22 pv d4d6 f6e7 d6a6 d3f2 g4f4 e7d7 a6h6 d7c7 a5a7 c7b8 a7f7 b8a8
info depth 9 seldepth 31 multipv 1 score cp 60 nodes 3901275 nps 84810326 tbhits 111740 time 46 pv d4d6 f6f7 d6a6 d3f2 g4f4 b3b7 a5h5 f7f8 h5f5 f8g7 f4e3 f2d1 e3d4 b7f7
info depth 10 seldepth 50 multipv 1 score cp 58 nodes 18912615 nps 93626806 tbhits 696963 time 202 pv d4d6 f6g7 g4g5 d3f2 d6g6 g7h7 g6a6 b3b7 g5f6 f2g4 f6f5 g4e3 f5f4 e3c4 a5f5 b7b1 f5f7 h7g8 f7c7 b1f1 f4g5
info depth 11 seldepth 30 multipv 1 score cp 59 nodes 20119760 nps 93580279 tbhits 748096 time 215 pv d4d6 f6f7 d6a6 b3b7 a6d6 d3c1 a5c5 f7e7 d6h6 c1b3 c5c8 b7b4 g4g3 b3d4 c8c7 e7d8
info depth 12 seldepth 28 multipv 1 score cp 54 nodes 22204778 nps 93691046 tbhits 835019 time 237 pv d4d6 f6f7 d6a6 b3b7 a6d6 d3c1 a5c5 c1b3 c5f5 f7e7 d6d3 b7b4 g4g5 b3d4 f5d5 d4e6 g5f5 e6f8 d5e5 e7f7
info depth 13 seldepth 28 multipv 1 score cp 61 nodes 24216704 nps 93500787 tbhits 916991 time 259 pv d4d6 f6e7 d6a6 b3b4 g4f3 b4b7 a5h5 e7f8 f3e4 d3c1 e4d4 b7e7 d4c4 c1e2 h5f5 f8g7
info depth 14 seldepth 36 multipv 1 score cp 54 nodes 33313194 nps 92794412 tbhits 1297225 time 359 pv d4d6 f6e7 d6a6 d3f2 g4f4 f2d3 f4f5 b3b7 a6f6 d3b4 a5a8 b7c7 f6h6 c7c5 f5g4 e7f7 a8a7 f7g8 a7d7 b4c6 g4f4 c6e5
info depth 15 seldepth 42 multipv 1 score cp 50 nodes 51364493 nps 91233557 tbhits 2102064 time 563 pv d4d6 f6e7 d6a6 b3b7 a6h6 e7f8 h6d6 d3b4 g4f5 f8e7 d6e6 e7f7 e6h6 f7g7 h6g6 g7f7 f5g5 b7e7 g6f6 f7g7 f6d6 b4c2 d6g6 g7f7 g5h6 c2d4 g6g4 e7e6 h6h5 d4c6 a5a8
info depth 16 seldepth 68 multipv 1 score cp 49 nodes 90276965 nps 89118425 hashfull 9 tbhits 3820806 time 1013 pv d4d6 f6e7 d6a6 b3c3 a6h6 c3c7 a5g5 c7c4 g4f3 e7f8 g5f5 f8e8 f5a5 c4c7 f3e4 d3b4 a5b5 c7c4 e4e3 e8d7 b5b7 d7c8 b7g7 c4c7 g7g5 c7e7 e3d4 b4c2 d4d3
info depth 17 seldepth 52 multipv 1 score cp 50 nodes 117628247 nps 87979242 hashfull 11 tbhits 5077768 time 1337 pv d4d6 f6e7 d6a6 b3c3 a6h6 c3c7 a5g5 c7c4 g4f3 e7f8 g5f5 f8g7 h6h5 c4c6 f3e4 d3b4 h5g5 g7h6 e4d4 b4c2 d4d3 c2b4 d3e4 c6e6 f5e5 e6e5 g5e5 h6g7 e5e6 g7f8 e6b6 b4a2 e4d3 f8e8
info depth 18 seldepth 45 multipv 1 score cp 42 nodes 128035075 nps 87635232 hashfull 12 tbhits 5556538 time 1461 pv d4d6 f6e7 d6a6 b3c3 a6h6 c3c7 a5g5 c7c4 g4f3 e7f8 g5b5 c4c7 f3e3 d3c5 e3d4 c5d7 b5f5 f8g7 h6a6 d7b8 a6d6 c7c1 f5g5 g7f7 d6h6 b8c6 d4e3 c6e7 h6h7 f7e8 h7g7 c1e1 e3f2
info depth 19 seldepth 68 multipv 1 score cp 44 nodes 164667526 nps 86575986 hashfull 17 tbhits 7275075 time 1902 pv d4d6 f6e7 d6a6 b3b7 a5h5 e7f8 a6a8 f8g7 h5g5 g7f7 g5f5 f7g7 a8d8 d3c1 d8d6 b7c7 f5b5 c1a2 g4f3 a2c3 b5a5 c7e7 d6d4 e7c7 f3e3 g7f6 d4d6 f6e7 d6h6 e7f8 e3d3 c3d1 d3d4 c7d7 d4c5
info depth 20 seldepth 71 multipv 1 score cp 44 nodes 247540567 nps 85211899 hashfull 28 tbhits 11277718 time 2905 pv d4d6 f6f7 d6a6 b3b7 a6h6 f7f8 h6d6 d3b4 a5a4 b4c2 a4a8 f8f7 a8c8 f7e7 d6h6 c2e3 g4f3 b7b3 f3e4 e3g4 c8c7 e7d8 h6c6 b3b6 c6b6 d8c7 b6g6 g4f2 e4e3 f2d1 e3d4 d1f2 g6g2 f2h3 g2a2
info depth 21 currmove d4d3 currmovenumber 2
info depth 21 currmove a5a6 currmovenumber 3
info depth 21 currmove g4h5 currmovenumber 4
info depth 21 currmove a5e5 currmovenumber 5
info depth 21 currmove a5g5 currmovenumber 6
info depth 21 currmove g4f3 currmovenumber 7
info depth 21 currmove a5h5 currmovenumber 8
info depth 21 currmove a5d5 currmovenumber 9
info depth 21 currmove g4g3 currmovenumber 10
info depth 21 currmove d4c4 currmovenumber 11
info depth 21 currmove a5f5 currmovenumber 12
info depth 21 currmove a5c5 currmovenumber 13
info depth 21 currmove d4d5 currmovenumber 14
info depth 21 currmove d4d7 currmovenumber 15
info depth 21 currmove a5a3 currmovenumber 16
info depth 21 currmove d4f4 currmovenumber 17
info depth 21 currmove a5a4 currmovenumber 18
info depth 21 currmove d4d8 currmovenumber 19
info depth 21 currmove g4h3 currmovenumber 20
info depth 21 currmove g4h4 currmovenumber 21
info depth 21 currmove a5b5 currmovenumber 22
info depth 21 currmove d4b4 currmovenumber 23
info depth 21 currmove a5a2 currmovenumber 24
info depth 21 currmove a5a1 currmovenumber 25
info depth 21 currmove d4a4 currmovenumber 26
info depth 21 currmove d4e4 currmovenumber 27
info depth 21 seldepth 47 multipv 1 score cp 38 nodes 284217304 nps 84739804 hashfull 32 tbhits 13033141 time 3354 pv d4d6 f6f7 d6a6 b3b7 a6h6 f7f8 h6d6 d3b4 a5a4 f8g7 g4f5 b7f7 f5e4 b4c2 e4d3 c2e1 d3e2 e1f3 d6d5 f3g1 e2d3 f7f3 d3e4 f3f1 d5f5 f1f5 e4f5 g1e2 a4c4 e2g3 f5g4
info depth 22 currmove d4d6 currmovenumber 1
info depth 22 seldepth 30 multipv 1 score cp 44 lowerbound nodes 287905929 nps 84703127 hashfull 32 tbhits 13205410 time 3399 pv d4d6
info depth 22 currmove d4d6 currmovenumber 1
info depth 22 currmove a5a6 currmovenumber 2
info depth 22 currmove g4h5 currmovenumber 3
info depth 22 currmove a5h5 currmovenumber 4
info depth 22 currmove a5f5 currmovenumber 5
info depth 22 currmove g4f3 currmovenumber 6
info depth 22 currmove g4g3 currmovenumber 7
info depth 22 currmove a5g5 currmovenumber 8
info depth 22 currmove a5d5 currmovenumber 9
info depth 22 currmove d4d7 currmovenumber 10
info depth 22 currmove a5e5 currmovenumber 11
info depth 22 currmove d4f4 currmovenumber 12
info depth 22 currmove a5b5 currmovenumber 13
info depth 22 currmove a5a1 currmovenumber 14
info depth 22 currmove d4d5 currmovenumber 15
info depth 22 currmove d4d8 currmovenumber 16
info depth 22 currmove d4e4 currmovenumber 17
info depth 22 currmove a5a2 currmovenumber 18
info depth 22 currmove a5a4 currmovenumber 19
info depth 22 currmove d4c4 currmovenumber 20
info depth 22 currmove a5c5 currmovenumber 21
info depth 22 currmove d4b4 currmovenumber 22
info depth 22 currmove g4h3 currmovenumber 23
info depth 22 currmove a5a3 currmovenumber 24
info depth 22 currmove g4h4 currmovenumber 25
info depth 22 currmove d4a4 currmovenumber 26
info depth 22 currmove d4d3 currmovenumber 27
info depth 22 seldepth 55 multipv 1 score cp 41 nodes 296973520 nps 84559658 hashfull 34 tbhits 13632156 time 3512 pv d4d6 f6e7 d6a6 b3c3 a6h6 c3c7 a5g5 c7c1 g5g7 e7f8 g7a7 d3e5 g4f5 e5f7 a7a8 f8g7 h6g6 g7h7 g6g2 f7h6 f5e4 c1c7 g2h2 h7g7 a8a6 c7e7 e4d5 h6f7 h2f2 f7d8 f2g2 g7f7 a6d6 e7e1 d6h6 d8e6 h6h7 f7f6 g2f2 f6g6
info depth 23 currmove d4d6 currmovenumber 1
?info depth 23 seldepth 62 multipv 1 score cp 48 lowerbound nodes 344435843 nps 83967782 hashfull 37 tbhits 15884149 time 4102 pv d4d6
info depth 23 currmove d4d6 currmovenumber 1
info depth 23 currmove d4d3 currmovenumber 2
info depth 23 currmove a5a6 currmovenumber 3
info depth 23 currmove g4h5 currmovenumber 4
info depth 23 currmove g4f3 currmovenumber 5
info depth 23 currmove a5f5 currmovenumber 6
info depth 23 currmove g4g3 currmovenumber 7
info depth 23 currmove d4c4 currmovenumber 8
info depth 23 currmove a5h5 currmovenumber 9
info depth 23 currmove a5g5 currmovenumber 10
info depth 23 currmove g4h3 currmovenumber 11
info depth 23 currmove g4h4 currmovenumber 12
info depth 23 currmove a5a1 currmovenumber 13
info depth 23 currmove a5a3 currmovenumber 14
info depth 23 currmove a5c5 currmovenumber 15
info depth 23 currmove a5a2 currmovenumber 16
info depth 23 currmove d4d5 currmovenumber 17
info depth 23 currmove d4f4 currmovenumber 18
info depth 23 currmove a5b5 currmovenumber 19
info depth 23 currmove d4d7 currmovenumber 20
info depth 23 currmove a5d5 currmovenumber 21
info depth 23 currmove a5a4 currmovenumber 22
info depth 23 currmove a5e5 currmovenumber 23
info depth 23 currmove d4e4 currmovenumber 24
info depth 23 currmove d4d8 currmovenumber 25
info depth 23 currmove d4b4 currmovenumber 26
info depth 23 currmove d4a4 currmovenumber 27
info depth 23 seldepth 107 multipv 1 score cp 34 upperbound nodes 390244875 nps 83528440 hashfull 41 tbhits 17837086 time 4672 pv d4d6 f6e7
info depth 23 currmove d4d6 currmovenumber 1

info depth 23 seldepth 107 multipv 1 score cp 34 nodes 434083418 nps 83078166 hashfull 48 tbhits 19630707 time 5225 pv d4d6 f6e7
bestmove d4d6 ponder f6e7
Image

jhellis3
Posts: 488
Joined: Fri Aug 16, 2013 10:36 pm

Re: Bug in Crystal ?

Post by jhellis3 » Fri Jun 05, 2020 2:01 pm

Nah, it is not a bug, it can appear that way though. It is working as intended. There are a couple of things which can be tried to address the cosmetics though. I will probably upload a couple of exes later today for people to test.

Post Reply