Crazyhouse Computer Championships 2016

[Ferdy]

You may download all games from,

https://sites.google.com/site/zhassociation/download

Look for 1st CCVA Comp Ch 2016 directory and find the file 1st_ccva_ch_2016.zip, it contains, per pairing pgn file and all games pgn file, ordo rating summary and head to head files are also included.

[MikeB]

Congratz to the Stockfish zh team winning the 1st Crazyhouse Chess Variant Computer Championships.
It builds an enormous lead with a score of 30.5 - 5.5 vs Imortal in the final match. The first player to reach 32.5 or more will win the match and the championships.

Games will be available once the match is finished.

Tables/brackets:
http://ccva.challonge.com/1st_ccva_comp_champ_2016

Site:
https://sites.google.com/site/zhassocia ... mpionships

Very impressive , will grab the latest release and make a special Mac Version with "Play by Elo" - Congrats to Dan and the SF zh team!

This is all Dan and team - just doing a service for my fellow macOS comrades.

https://github.com/ddugovic

macOS exe's on the most current variant source with play by ELO ( UCI_LImit_Strength & UCI_ELO & Move Delay)
https://github.com/MichaelB7/Stockfish/releases

[Toadofsky]

Thanks to Ferdinand Mosca for running this championship and to Fabian Fichter & others for an excellent collaborative effort!

[Carlos777]

Congratulations to the Stockfish Variant team for this outstanding result and to Ferdinand for running this nice tournament.

[Vinvin]

Thanks to Ferdinand Mosca for running this championship and to Fabian Fichter & others for an excellent collaborative effort!

+1 !

[Ferdy]

Thanks to Ferdinand Mosca for running this championship and to Fabian Fichter & others for an excellent collaborative effort!

We will do it again next year around the same time, I hope to have more entries by then.

It seems zh is one of the popular variants in Lichess, I saw chess.com is also into this variant.

I will write chessbase asking for zh support for their Fritz GUI.

[Vinvin]

Thanks to Ferdinand Mosca for running this championship and to Fabian Fichter & others for an excellent collaborative effort!

We will do it again next year around the same time, I hope to have more entries by then.

It seems zh is one of the popular variants in Lichess, I saw chess.com is also into this variant.

I will write chessbase asking for zh support for their Fritz GUI.

Good idea !

But the more I analyze the opening, the more I think 1.e4 is winning ...

e5 seems losing :
Some eval with 3 best moves after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3 after 8h30m, best move is h6 but -2.11 score.

Code: Select all

score cp -258 pv c5e7 c3d5 d7d6 d5e7 d8e7 B@h4 c8g4 c1g5 N@d7
score cp -326 pv d7d6 c1g5 g8h8 c3d5 f6d5 g5d8 N@e3 d1c1 c8g4
score cp -211 pv h7h6 c3d5

e6 :

I just started a long analyze with 3 best moves after 1.e4 e6 2. d4 d5 3.Nc3
Very good score for white too (>1.5) after 10 hours (6*4GHz) :

Code: Select all

depth 36 score cp -153 pv b8c6 e4e5 g8e7 g1f3 e7f5 f1b5 f8e7 e1g1 e8g8 b5d3 a8b8 d3f5 e6f5 N@h5 B@h8
depth 35 score cp -185 pv d5e4 c3e4 g8f6 e4f6 g7f6 g1f3 b8d7 f1d3 h8g8 e1g1 f6f5 d3b5 a7a6 b5d7 c8d7 N@e5
depth 35 score cp -221 pv g8f6 e4e5 f6d7 g1f3 f8e7 f1d3 b8c6 e1g1 c6b4 d3e2 e8g8 a2a3 b4c6 e2b5 a7a5 d1e2 

[Ferdy]

Thanks to Ferdinand Mosca for running this championship and to Fabian Fichter & others for an excellent collaborative effort!

We will do it again next year around the same time, I hope to have more entries by then.

It seems zh is one of the popular variants in Lichess, I saw chess.com is also into this variant.

I will write chessbase asking for zh support for their Fritz GUI.

Good idea !

But the more I analyze the opening, the more I think 1.e4 is winning ...

e5 seems losing :
Some eval with 3 best moves after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3 after 8h30m, best move is h6 but -2.11 score.

Code: Select all

score cp -258 pv c5e7 c3d5 d7d6 d5e7 d8e7 B@h4 c8g4 c1g5 N@d7
score cp -326 pv d7d6 c1g5 g8h8 c3d5 f6d5 g5d8 N@e3 d1c1 c8g4
score cp -211 pv h7h6 c3d5

e6 :

I just started a long analyze with 3 best moves after 1.e4 e6 2. d4 d5 3.Nc3
Very good score for white too (>1.5) after 10 hours (6*4GHz) :

Code: Select all

depth 36 score cp -153 pv b8c6 e4e5 g8e7 g1f3 e7f5 f1b5 f8e7 e1g1 e8g8 b5d3 a8b8 d3f5 e6f5 N@h5 B@h8
depth 35 score cp -185 pv d5e4 c3e4 g8f6 e4f6 g7f6 g1f3 b8d7 f1d3 h8g8 e1g1 f6f5 d3b5 a7a6 b5d7 c8d7 N@e5
depth 35 score cp -221 pv g8f6 e4e5 f6d7 g1f3 f8e7 f1d3 b8c6 e1g1 c6b4 d3e2 e8g8 a2a3 b4c6 e2b5 a7a5 d1e2 

I think Sf zh is not yet there where you can believe on the stability of its score.

Sample run (from your first example position) at 10 minutes/node for the 3 nodes (blue square) and 10 minutes x 2 for root node (red circle) each node is run at MultiPV=3.
It likes c5e7 after 20 minutes of analysis, but in the end the worst move d7d6 is actually the move that produces the best score (after mini-maxing) for the root position after searching further the children of the square nodes.
I use Stockfish zh 22Dec2016.



If you like to experiment further, from your given position,
1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3
You already have this record,

Code: Select all

score cp -258 pv c5e7 c3d5 d7d6 d5e7 d8e7 B@h4 c8g4 c1g5 N@d7 
score cp -326 pv d7d6 c1g5 g8h8 c3d5 f6d5 g5d8 N@e3 d1c1 c8g4 
score cp -211 pv h7h6 c3d5

Do this:
a) 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3 c5e7
Run again in multipv 3 and record the same

Do similar to the 2 positions below
b) 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3 d7d6
c) 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.d3 Nf6 5.o-o o-o 6.Nc3 h7h6

Then we can compare the forward score and the min-max score. This is an interesting experiment to see the search score stability and reliability of the engine.

[Toadofsky]

I think Sf zh is not yet there where you can believe on the stability of its score.

Empirically (based on experience implementing other variants) I agree, usually scores are only stable after all other issues are resolved (and then the parameters tuned).

[George Tsavdaris]

Congratz to the Stockfish zh team winning the 1st Crazyhouse Chess Variant Computer Championships.
It builds an enormous lead with a score of 30.5 - 5.5 vs Imortal in the final match. The first player to reach 32.5 or more will win the match and the championships.

Games will be available once the match is finished.

Tables/brackets:
http://ccva.challonge.com/1st_ccva_comp_champ_2016

Site:
https://sites.google.com/site/zhassocia ... mpionships

Thanks, it was interesting.
Will Imortal 3 will be released?