ShashChess

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

Moderators: hgm, Rebel, chrisw

amchess
Posts: 345
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

Thanks for the many updates.
29.1 is the first one, that doesn't work correctly with MEA (Ferdinand Mosca). Given milliseconds for each position aren't interpreted correctly, everything <500 is ignored. No error message, run is completed and rated, just the TC is wrong.
Tried already to set Move Overhead and Slow Mover to minimum, doesn't help neither.
Best regards
Peter.
amchess
Posts: 345
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

peter wrote: Wed Mar 08, 2023 10:15 am
Thanks for the many updates.
29.1 is the first one, that doesn't work correctly with MEA (Ferdinand Mosca). Given milliseconds for each position aren't interpreted correctly, everything <500 is ignored. No error message, run is completed and rated, just the TC is wrong.
Tried already to set Move Overhead and Slow Mover to minimum, doesn't help neither.
Best regards
Hi.
ShashChess 29 uses the win probability to classify positions based on Shashin theory and uses for this the formula in Stockfish wdl model. This formula is rather computationally onerous.
Therefore, in ShashChess 29.1, we introduced a hashmap (to be precise, unordered_map), initialized at engine startup in a similar way to syzygy, so that we have immediate access to the win probability, given a score and depth of analysis.
I am trying to figure out how this might impact this tool that I am not familiar with.
peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

amchess wrote: Wed Mar 08, 2023 4:05 pm I am trying to figure out how this might impact this tool that I am not familiar with.
Thanks for explaining and looking into it, Andrea, just to notice again, with version 29 the TC was interpreted correctly still, 29.1 is the first one aberrant as for the tool and its fixed time per move syntax, regards
Peter.
peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

peter wrote: Wed Mar 08, 2023 5:53 pm
amchess wrote: Wed Mar 08, 2023 4:05 pm I am trying to figure out how this might impact this tool that I am not familiar with.
Thanks for explaining and looking into it, Andrea, just to notice again, with version 29 the TC was interpreted correctly still, 29.1 is the first one aberrant as for the tool and its fixed time per move syntax, regards
BTW, as for Banksia 0.58 positional test tool it's about the same, 200 msec/pos. given takes about 8.5 min for 512, that's well 5 times too much:

Code: Select all

File name          : 512.epd
Total test items   : 512
Test for           : best moves
Total engines      : 1
Timer              : movetime: 0.2
Expand ply         : 0
Elapsed            : 08:35
Laps               : 1
Total tests        : 512
Total corrects     : 420 (82%)
Ave correct elapse : 17 ms
Status             : completed


Correct/Total:
ShashChess 29.1: 420/512
Peter.
peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

peter wrote: Wed Mar 08, 2023 7:44 pm BTW, as for Banksia 0.58 positional test tool it's about the same, 200 msec/pos. given takes about 8.5 min for 512, that's well 5 times too much:

Code: Select all

File name          : 512.epd
Total test items   : 512
Test for           : best moves
Total engines      : 1
Timer              : movetime: 0.2
Expand ply         : 0
Elapsed            : 08:35
Laps               : 1
Total tests        : 512
Total corrects     : 420 (82%)
Ave correct elapse : 17 ms
Status             : completed


Correct/Total:
ShashChess 29.1: 420/512
>750 msec TC- interpretation starts coming nearer to the one of other engines, but yet even with 1"/move, there's still about 150% time consumption by ShashChess 29.1 than by other engines in GUIs that allow so (and in console like with MEA).

Shredder13- GUI with 1"/move ("exact time"- checked), like Dragon in this one example also other one SF- branches and e.g. Ethereal 14.00 take rather exactly 512", Shashchess 29.1 takes 850:

Code: Select all

Dragon 3.2:
Bisher gelöst: 405 von 512  ;  2:32m


         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   -   0   -   0   0   -   0   -   0   0   0   0   0   0   0   0   0   -   0   0
  20 |   0   -   0   0   0   -   0   0   0   -   0   0   -   -   0   0   0   0   0   0
  40 |   -   0   0   -   -   0   0   0   -   0   0   -   0   0   0   -   0   -   -   0
  60 |   -   0   -   0   0   0   -   0   0   -   0   0   0   0   -   0   0   0   0   0
  80 |   0   0   0   -   0   -   -   0   0   -   0   0   0   0   0   0   0   0   0   0
 100 |   0   0   0   0   0   -   -   -   0   -   0   -   0   0   -   0   0   0   0   0
 120 |   0   -   -   -   -   0   0   0   -   0   0   0   0   0   0   -   -   0   0   -
 140 |   0   -   0   -   0   0   0   0   -   -   -   -   0   -   0   0   0   0   -   0
 160 |   0   -   0   0   0   -   0   0   0   -   0   -   0   0   -   0   0   0   0   0
 180 |   0   0   0   0   0   0   0   0   0   0   0   -   0   0   0   -   0   0   -   -
 200 |   -   0   0   0   0   0   0   0   0   -   0   0   0   0   0   -   0   0   -   -
 220 |   0   0   0   0   0   0   0   0   -   -   -   0   0   0   -   0   -   0   0   0
 240 |   0   -   -   0   0   0   0   -   -   -   -   -   -   -   -   0   0   0   0   -
 260 |   -   0   0   -   -   -   0   0   -   -   0   0   0   0   0   0   0   0   0   0
 280 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 300 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 320 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 340 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 360 |   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   -   0   0
 380 |   0   0   0   0   0   0   0   0   0   0   0   -   0   0   -   0   0   0   0   -
 400 |   0   0   0   0   0   -   0   -   0   -   0   0   0   0   0   -   0   0   0   0
 420 |   0   0   0   0   0   0   0   0   0   0   -   -   0   0   0   -   0   0   0   0
 440 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   -   0   0   0   0
 460 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0
 480 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0
 500 |   0   -   -   -   -   0   0   0   0   -   0   -

   1 sec -> 405/512
  K/s: 1.973.838  
  TotTime: 8:32m    SolTime: 2:32m


ShashChess 29.1:
Bisher gelöst: 430 von 512  ;  2:59m

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   0   0   0   0   0   0   0   -   0   0   0   0   0   -   0   -   -   0   0   0
  20 |   0   -   -   0   0   -   0   0   0   0   -   0   -   -   0   0   0   0   0   0
  40 |   0   0   0   -   0   0   0   0   -   0   0   0   1   0   0   0   0   -   -   -
  60 |   -   0   -   0   0   0   0   0   0   -   0   0   0   0   0   0   0   0   -   0
  80 |   0   -   0   0   0   -   0   -   0   -   0   0   0   0   0   0   0   0   0   0
 100 |   -   0   0   0   0   0   0   -   -   -   0   -   0   0   0   0   0   0   -   0
 120 |   0   -   -   -   -   0   -   0   -   0   0   0   0   0   0   -   0   0   0   -
 140 |   0   -   -   0   0   0   0   0   0   -   0   -   0   0   0   0   -   0   0   0
 160 |   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 180 |   0   0   0   0   0   0   0   -   0   0   -   -   0   0   0   -   0   0   0   0
 200 |   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   -   0   0   -   -
 220 |   0   0   0   0   0   -   0   0   -   0   -   0   0   0   0   0   -   0   0   0
 240 |   0   0   -   0   0   0   0   0   -   0   -   -   0   0   0   0   0   0   -   0
 260 |   -   -   0   0   -   -   0   0   -   -   -   0   -   0   0   0   0   0   0   0
 280 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 300 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 320 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 340 |   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   0
 360 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 380 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 400 |   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   -   0   0   0   0
 420 |   0   0   0   0   0   0   0   0   0   0   -   0   0   0   0   -   0   0   0   0
 440 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   -   0   0   -   0
 460 |   0   0   0   0   0   0   0   -   -   0   0   0   0   -   0   0   0   0   0   0
 480 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0
 500 |   0   0   -   0   -   0   0   0   0   0   0   0

   1 sec -> 429/512
  K/s: 856.274  
  TotTime: 14:10m    SolTime: 2:59m
So far I saw only Fritz seeming to deny so much difference at 1"/move and more, regards
Peter.
Jouni
Posts: 3324
Joined: Wed Mar 08, 2006 8:15 pm

Re: ShashChess

Post by Jouni »

In my test 29.1 + Tal is very good in testsuites! But Crystal 5 is even better in all suites.
Jouni
peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

Jouni wrote: Thu Mar 09, 2023 5:03 pm In my test 29.1 + Tal is very good in testsuites! But Crystal 5 is even better in all suites.
5" is the shortest TC 29.1 at least in Fritz GUI interpretes correctly,
at 1" even in Fritz (18) 578" instead of 512 max given are used yet still:
Ergebnis: 443 aus 512 = 86.5%. Durchschnittszeit = 0.72s / 9.38
EloStatTS relativizes the time to solution additionally anyhow.. For 5"/postion and 30 threads of the 16x3.5 GHz- CPU I like e.g. these 256:
https://www.dropbox.com/s/s3jn6pxp16diqvt/256.epd?dl=0

Code: Select all


    Program                                    Elo   +/-  Matches  Score   Av.Op.   S.Pos.   MST1    MST2   RIndex

  1 Crystal5KWK                              : 3565    7   1603    60.8 %   3488   168/256    1.8s    2.9s   0.73
  2 BlueMarlin15.6                           : 3562    7   1534    60.4 %   3489   160/256    1.7s    3.0s   0.75
  3 Eman8.7064-bitBMI2                       : 3546    7   1452    57.7 %   3492   146/256    1.8s    3.1s   0.69
  4 ShashChess29.1                           : 3530    8   1454    55.2 %   3494   143/256    2.1s    3.4s   0.53
  5 ShashChess28                             : 3529    8   1442    55.0 %   3494   140/256    2.1s    3.4s   0.56
  6 Dragon3.2byKomodoChess                   : 3491    9   1337    48.4 %   3502   102/256    1.8s    3.7s   0.47
  7 Berserk11.1                              : 3460    9   1211    42.8 %   3511    85/256    2.1s    4.0s   0.38
  8 Ethereal14.00(NNUE)                      : 3424    9   1159    37.2 %   3515    67/256    2.3s    4.3s   0.26
  9 Rebel-16.2                               : 3416    9   1148    36.0 %   3516    60/256    2.1s    4.3s   0.27
 10 Koivisto9.2                              : 3406    9   1128    34.5 %   3517    52/256    1.7s    4.3s   0.28



MST1  : Mean solution time (solved positions only)
MST2  : Mean solution time (solved and unsolved positions)
RIndex: Score according to solution time ranking for each position
Peter.
peter
Posts: 3202
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: ShashChess

Post by peter »

peter wrote: Thu Mar 09, 2023 10:28 am >750 msec TC- interpretation starts coming nearer to the one of other engines, but yet even with 1"/move, there's still about 150% time consumption by ShashChess 29.1 than by other engines in GUIs that allow so (and in console like with MEA).

Shredder13- GUI with 1"/move ("exact time"- checked), like Dragon in this one example also other one SF- branches and e.g. Ethereal 14.00 take rather exactly 512", Shashchess 29.1 takes 850:

Code: Select all

Dragon 3.2:
Bisher gelöst: 405 von 512  ;  2:32m


         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   -   0   -   0   0   -   0   -   0   0   0   0   0   0   0   0   0   -   0   0
  20 |   0   -   0   0   0   -   0   0   0   -   0   0   -   -   0   0   0   0   0   0
  40 |   -   0   0   -   -   0   0   0   -   0   0   -   0   0   0   -   0   -   -   0
  60 |   -   0   -   0   0   0   -   0   0   -   0   0   0   0   -   0   0   0   0   0
  80 |   0   0   0   -   0   -   -   0   0   -   0   0   0   0   0   0   0   0   0   0
 100 |   0   0   0   0   0   -   -   -   0   -   0   -   0   0   -   0   0   0   0   0
 120 |   0   -   -   -   -   0   0   0   -   0   0   0   0   0   0   -   -   0   0   -
 140 |   0   -   0   -   0   0   0   0   -   -   -   -   0   -   0   0   0   0   -   0
 160 |   0   -   0   0   0   -   0   0   0   -   0   -   0   0   -   0   0   0   0   0
 180 |   0   0   0   0   0   0   0   0   0   0   0   -   0   0   0   -   0   0   -   -
 200 |   -   0   0   0   0   0   0   0   0   -   0   0   0   0   0   -   0   0   -   -
 220 |   0   0   0   0   0   0   0   0   -   -   -   0   0   0   -   0   -   0   0   0
 240 |   0   -   -   0   0   0   0   -   -   -   -   -   -   -   -   0   0   0   0   -
 260 |   -   0   0   -   -   -   0   0   -   -   0   0   0   0   0   0   0   0   0   0
 280 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 300 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 320 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 340 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 360 |   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   -   0   0
 380 |   0   0   0   0   0   0   0   0   0   0   0   -   0   0   -   0   0   0   0   -
 400 |   0   0   0   0   0   -   0   -   0   -   0   0   0   0   0   -   0   0   0   0
 420 |   0   0   0   0   0   0   0   0   0   0   -   -   0   0   0   -   0   0   0   0
 440 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   -   0   0   0   0
 460 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0
 480 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0
 500 |   0   -   -   -   -   0   0   0   0   -   0   -

   1 sec -> 405/512
  K/s: 1.973.838  
  TotTime: 8:32m    SolTime: 2:32m


ShashChess 29.1:
Bisher gelöst: 430 von 512  ;  2:59m

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   0   0   0   0   0   0   0   -   0   0   0   0   0   -   0   -   -   0   0   0
  20 |   0   -   -   0   0   -   0   0   0   0   -   0   -   -   0   0   0   0   0   0
  40 |   0   0   0   -   0   0   0   0   -   0   0   0   1   0   0   0   0   -   -   -
  60 |   -   0   -   0   0   0   0   0   0   -   0   0   0   0   0   0   0   0   -   0
  80 |   0   -   0   0   0   -   0   -   0   -   0   0   0   0   0   0   0   0   0   0
 100 |   -   0   0   0   0   0   0   -   -   -   0   -   0   0   0   0   0   0   -   0
 120 |   0   -   -   -   -   0   -   0   -   0   0   0   0   0   0   -   0   0   0   -
 140 |   0   -   -   0   0   0   0   0   0   -   0   -   0   0   0   0   -   0   0   0
 160 |   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 180 |   0   0   0   0   0   0   0   -   0   0   -   -   0   0   0   -   0   0   0   0
 200 |   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   -   0   0   -   -
 220 |   0   0   0   0   0   -   0   0   -   0   -   0   0   0   0   0   -   0   0   0
 240 |   0   0   -   0   0   0   0   0   -   0   -   -   0   0   0   0   0   0   -   0
 260 |   -   -   0   0   -   -   0   0   -   -   -   0   -   0   0   0   0   0   0   0
 280 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 300 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 320 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 340 |   0   0   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   0   0   0
 360 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 380 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
 400 |   0   0   0   0   0   0   0   -   0   0   0   0   0   0   0   -   0   0   0   0
 420 |   0   0   0   0   0   0   0   0   0   0   -   0   0   0   0   -   0   0   0   0
 440 |   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0   -   0   0   -   0
 460 |   0   0   0   0   0   0   0   -   -   0   0   0   0   -   0   0   0   0   0   0
 480 |   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   -   0
 500 |   0   0   -   0   -   0   0   0   0   0   0   0

   1 sec -> 429/512
  K/s: 856.274  
  TotTime: 14:10m    SolTime: 2:59m
As for MEA, TC given 100msec/pos. for ShashChess 29.1 results in about same time consumption (with Move Overhead and Slow Mover set to min.) like with other engines default setting and 750msec/move (like e.g with ShCh 29 too), all engines execpt LC0 (Nividia 3070ti GPU) ran with single thread of 3.5GHz CPU over these 2345 positions
https://www.dropbox.com/s/hqeqqn31s5ynqr2/2345.epd?dl=0

Code: Select all

Kovisto 9.2:
Time allocation  : GOOD!!
at <= et + mt and at >= et - mt
ExpectedTime     : 1758.8s
ActualTime       : 1766.7s
MarginTime/pos   : 0.2s
MarginTime       : 438.5s

ShashChess 29:
Time allocation  : GOOD!!
at <= et + mt and at >= et - mt
ExpectedTime     : 1758.8s
ActualTime       : 1765.3s
MarginTime/pos   : 0.2s
MarginTime       : 438.5s

ShashChess 29.1:
Time allocation  : BAD!! spending more time
at > et + mt
ExpectedTime     : 234.5s
ActualTime       : 1755.4s
MarginTime/pos   : 0.1s
MarginTime       : 117.2s

Code: Select all

   
    EPD  : epd\2345.epd
    Time : 750ms
                                                Max   Total   Time   Hash          
    Engine           Score   Found  Pos   ELO  Score   Rate    ms     Mb  Cpu      
 1  SF230305         29611   1908  2345  4108  32444  91.3%    750     8    1
 2  BlueMarlin15.6   29467   1893  2345  4086  32444  90.8%    750     8    1
 3  Eman8.70NN       29392   1897  2345  4077  32444  90.6%    750     8    1
 4  ShashChess29     29383   1893  2345  4077  32444  90.6%    750     8    1
 5  lc0230217-325M   29246   1896  2345  4054  32444  90.1%    800     8    2
 6  Dragon3.2        28739   1842  2345  3987  32444  88.6%    750     8    1
 7  Rebel16.2        28067   1792  2345  3892  32444  86.5%    750     8    1
 8  Ethereal14       27926   1785  2345  3874  32444  86.1%    750     8    1
 9  Koivisto9.2fix   27761   1777  2345  3852  32444  85.6%    750     8    1
10  ShashChess29.1   27666   1752  2345  3838  32444  85.3%    750     8    1

                                    Created with MEA
                                          by
                                       Ferdinand
                                         Mosca

Peter.
amchess
Posts: 345
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »