Testing LazySMP

[LazySMP]

I am not your slave Daniel. Clown.

Please be nice to me. I don't ever tell someone I'm not gonna answer their question, and I never ignore their question. If you believe that my engine has a bug or crashes, please send me the PGN file or FEN string.

Uahhh, 40 games, result 40:0 for Fruit, sudden death tc, exactly one startposition (the startpos) - 20 (near) 'doubles' as expected, never seen an engine developer test in a CB GUI too A wonderful test.

Two engines only support single thread, so there is a possibility that the games will be duplicated. If you really want to help, send the PGN file of the games in which my engine is crashing.

[Viren]

Code: Select all

No. Name         Win Draw Loss Unf.  Score Games       %
--------------------------------------------------------
  1 fruit-2.1    +62   =2   -1  *11   63.0    65   96.9%
  2 lazysmp-3.0   +1   =2  -62  *11    2.0    65    3.1%
  

This claim that lazysmp can beat fruit is way off the mark anyway.

Clown.

Moderation.

Insults are not acceptable here.

Warning given.

You can't go on like this Ed.

[Rebel]

I am not your slave Daniel. Clown.

Please be nice to me. I don't ever tell someone I'm not gonna answer their question, and I never ignore their question. If you believe that my engine has a bug or crashes, please send me the PGN file or FEN string.

Meaning to say, I think I gave you enough attention. Someone who can write an engine doing 8.6 million NPS does not need any help.

Be well.

[chessica]

Code: Select all

   Motor          Punkte          Al         La    S-B
1: Alexander1.1   10,0/10 ·········· 1111111111    0,00
2: LazySMP_popcnt 0,0/10  0000000000 ··········    0,00

Game level: Tournament game in 5 minutes
Hardware: AMD Athlon(tm) II X4 630 Processor with 11.8 GB memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit

Did you notice any bugs or unexpected crashes in my engine during the tournament? Can you estimate an ELO for my chess engine?!

chrashes: no
ELO: no

[hgm]

Thank you for your advice on how best to test my chess engine. I've run a number of matches between Fruit 2.1 and LazySMP 3.0:

https://filebin.net/blkcw211fmlb0rtr/game.pgn

There were no unexpected disconnections in any game. The important point is that in all games LazySMP lost to Fruit by a huge margin.

If you play deterministic engines against each other from the same starting position, they will repeat the same two games over and over again. So you basically played only two test games, and if your engine is only moderately buggy, two games would not be enough to detect the effect of any bugs. To conduct meaningful tests you should play all games from a different start position, or by forcing a different opening line.

Also, the 'disconnect' error message is specific for a particular GUI (or other tournament agent), I beleve CuteChess-CLI. If you use another GUI to play the match, the symptoms could be different. E.g. a crashing engine might just lead to a forfeit on time.

[Viz]

Thank you for your advice on how best to test my chess engine. I've run a number of matches between Fruit 2.1 and LazySMP 3.0:

https://filebin.net/blkcw211fmlb0rtr/game.pgn

There were no unexpected disconnections in any game. The important point is that in all games LazySMP lost to Fruit by a huge margin.

If you play deterministic engines against each other from the same starting position, they will repeat the same two games over and over again.

This is not strictly true if you actually play the games with time management.
Because frequency isn't really stable and thus it actually needs slightly different times to reach the same nodes and it then changes TM etc.
It strictly happens if you play node/move, depth/move and stuff like this.
Ofc if you use multiple CPUs it's also not true for any engine with LazySMP.

[hgm]

It is strictly true for deterministic engines (by definition), but in timed games engines are not strictly deterministic. The move variation will be very small, end most games will be the same, or will only diverge in the final phase.

[LazySMP]

If you play deterministic engines against each other from the same starting position, they will repeat the same two games over and over again. So you basically played only two test games, and if your engine is only moderately buggy, two games would not be enough to detect the effect of any bugs. To conduct meaningful tests you should play all games from a different start position, or by forcing a different opening line.

Thank you for your guidance and wisdom. More than 32 matches were held with different positions.

Code: Select all

[FEN "bbqnnrkr/pppppppp/8/8/8/8/PPPPPPPP/BBQNNRKR w KQkq - 0 1"]
[SetUp "0"]

[FEN "qbbnnrkr/pppppppp/8/8/8/8/PPPPPPPP/QBBNNRKR w KQkq - 0 1"]
[SetUp "1"]

[FEN "qbnnbrkr/pppppppp/8/8/8/8/PPPPPPPP/QBNNBRKR w KQkq - 0 1"]
[SetUp "2"]

[FEN "qbnnrkbr/pppppppp/8/8/8/8/PPPPPPPP/QBNNRKBR w KQkq - 0 1"]
[SetUp "3"]

[FEN "bbnqnrkr/pppppppp/8/8/8/8/PPPPPPPP/BBNQNRKR w KQkq - 0 1"]
[SetUp "4"]

[FEN "nbbqnrkr/pppppppp/8/8/8/8/PPPPPPPP/NBBQNRKR w KQkq - 0 1"]
[SetUp "5"]

[FEN "nbqnbrkr/pppppppp/8/8/8/8/PPPPPPPP/NBQNBRKR w KQkq - 0 1"]
[SetUp "6"]

[FEN "nbqnrkbr/pppppppp/8/8/8/8/PPPPPPPP/NBQNRKBR w KQkq - 0 1"]
[SetUp "7"]

[FEN "bbnnqrkr/pppppppp/8/8/8/8/PPPPPPPP/BBNNQRKR w KQkq - 0 1"]
[SetUp "8"]

[FEN "nbbnqrkr/pppppppp/8/8/8/8/PPPPPPPP/NBBNQRKR w KQkq - 0 1"]
[SetUp "9"]

[FEN "nbnqbrkr/pppppppp/8/8/8/8/PPPPPPPP/NBNQBRKR w KQkq - 0 1"]
[SetUp "10"]

[FEN "nbnqrkbr/pppppppp/8/8/8/8/PPPPPPPP/NBNQRKBR w KQkq - 0 1"]
[SetUp "11"]

[FEN "bbnnrqkr/pppppppp/8/8/8/8/PPPPPPPP/BBNNRQKR w KQkq - 0 1"]
[SetUp "12"]

[FEN "nbbnrqkr/pppppppp/8/8/8/8/PPPPPPPP/NBBNRQKR w KQkq - 0 1"]
[SetUp "13"]

[FEN "nbnrbqkr/pppppppp/8/8/8/8/PPPPPPPP/NBNRBQKR w KQkq - 0 1"]
[SetUp "14"]

[FEN "nbnrqkbr/pppppppp/8/8/8/8/PPPPPPPP/NBNRQKBR w KQkq - 0 1"]
[SetUp "15"]

[FEN "bbnnrkqr/pppppppp/8/8/8/8/PPPPPPPP/BBNNRKQR w KQkq - 0 1"]
[SetUp "16"]

[FEN "nbbnrkqr/pppppppp/8/8/8/8/PPPPPPPP/NBBNRKQR w KQkq - 0 1"]
[SetUp "17"]

[FEN "nbnrbkqr/pppppppp/8/8/8/8/PPPPPPPP/NBNRBKQR w KQkq - 0 1"]
[SetUp "18"]

[FEN "nbnrkqbr/pppppppp/8/8/8/8/PPPPPPPP/NBNRKQBR w KQkq - 0 1"]
[SetUp "19"]

[FEN "bbnnrkrq/pppppppp/8/8/8/8/PPPPPPPP/BBNNRKRQ w KQkq - 0 1"]
[SetUp "20"]

[FEN "nbbnrkrq/pppppppp/8/8/8/8/PPPPPPPP/NBBNRKRQ w KQkq - 0 1"]
[SetUp "21"]

[FEN "nbnrbkrq/pppppppp/8/8/8/8/PPPPPPPP/NBNRBKRQ w KQkq - 0 1"]
[SetUp "22"]

[FEN "nbnrkrbq/pppppppp/8/8/8/8/PPPPPPPP/NBNRKRBQ w KQkq - 0 1"]
[SetUp "23"]

[FEN "bbqnrnkr/pppppppp/8/8/8/8/PPPPPPPP/BBQNRNKR w KQkq - 0 1"]
[SetUp "24"]

[FEN "qbbnrnkr/pppppppp/8/8/8/8/PPPPPPPP/QBBNRNKR w KQkq - 0 1"]
[SetUp "25"]

[FEN "qbnrbnkr/pppppppp/8/8/8/8/PPPPPPPP/QBNRBNKR w KQkq - 0 1"]
[SetUp "26"]

[FEN "qbnrnkbr/pppppppp/8/8/8/8/PPPPPPPP/QBNRNKBR w KQkq - 0 1"]
[SetUp "27"]

[FEN "bbnqrnkr/pppppppp/8/8/8/8/PPPPPPPP/BBNQRNKR w KQkq - 0 1"]
[SetUp "28"]

[FEN "nbbqrnkr/pppppppp/8/8/8/8/PPPPPPPP/NBBQRNKR w KQkq - 0 1"]
[SetUp "29"]

[FEN "nbqrbnkr/pppppppp/8/8/8/8/PPPPPPPP/NBQRBNKR w KQkq - 0 1"]
[SetUp "30"]

[FEN "nbqrnkbr/pppppppp/8/8/8/8/PPPPPPPP/NBQRNKBR w KQkq - 0 1"]
[SetUp "31"]

To view the games, click on the following link:
https://filebin.net/3ochv3ewy8esjdjl/game1.txt

The only strange thing about these games was that there was only one draw with three fold repetition

Also, the 'disconnect' error message is specific for a particular GUI (or other tournament agent), I beleve CuteChess-CLI. If you use another GUI to play the match, the symptoms could be different. E.g. a crashing engine might just lead to a forfeit on time.

I see no 'disconnect' during the games. However, no one has found a crash in the engine so far.

Someone who can write an engine doing 8.6 million NPS does not need any help. Be well.

NPS doesn't mean anything, if you want high nps go back to pure alpha beta but may lose accuracy of the evaluation. Unfortunately, from the beginning of this thread, some users whose engine is among the top 10 tried to send insulting content, so no one would test my engine. There is no melodrama or self-pity to this engine, nor is there an ultimate pessimism.

[LazySMP]

This image appears to prove my point. As shown in the screenshot above, About two weeks ago, Mr. Gabor wanted to test my engine, but one of the members of this forum (Guenther) sent an insulting message, which made it not happen. Later, other people (Lucametehau) sent abusive messages on my same post, which were deleted by the moderators. Looks like they got their point and my post is totally destroyed, no one even wants to test my engine.

At least I was hoping that the CCRL members would explain why they didn't test my engine. Today's judgment, and all of the effort that has gone into getting to this engine, will be completely in vain if I do not step out in a similar position and make my voices heard.

This being my last word in this space, I dare to imagine what will become of the new engines.

Good luck,
Daniel

[chesskobra]

Daniel, if you have a linux executable, I can test it at least superficially against a bunch of engines I have installed (some small ones, some big ones). I have access to a Dell XPS 8960 with i7 processor, 24 cores, 32 GB RAM. I am not affiliated with any rating lists and can' t promise any extensive testing. I am just a layman.