ShashChess

[Rowen]

https://github.com/amchess/Alexander/releases/tag/4.0

I don't think the reduced strength elo is accurate. Playing against Alexander 4.0 using CB Limit strength ticked and with the minimum Elo of 1320 selected I am not able to win or even draw. This was also true using past versions. My over the board FIDE is 1670, and testing the engine against other engines using elo of 1320 also suggests that it is stronger than 1320. Perhaps the 1320 is derived from engine games? It is presently too strong for beginners ,at least using 1320 elo, although reduced depth can be used, problem with that is the engine moves immediately.

If anyone is interested I have concluded that the LimitStrength_CB , ELO_CB parameters are not working in Fritz gui.This also seems to be the case with Alexander 4.1

[amchess]

https://github.com/amchess/Alexander/releases/tag/5.0

[Rowen]

https://github.com/amchess/Alexander/releases/tag/5.0

More feedback, Alexander 5 plays very slowly with simulate human blunders parameter enabled and often loses on time in blitz. Although this can be changed with adjusting the move overhead parameter.

[amchess]

Exciting News for Chess Enthusiasts!

For a limited time, experience the power of Alexander's Handicap Mode on Lichess with a series of demo bots designed to simulate authentic playing strength across various Elo ranges.

Introducing Alexander Bots:
♟ https://lichess.org/@/BotAlexnc (Elo 1320-1499) – Book Depth 4
♟ https://lichess.org/@/BotAlex3N (Elo 1500-1599) – Book Depth 6
♟ https://lichess.org/@/BotAlex2N (Elo 1600-1799) – Book Depth 8
♟ https://lichess.org/@/BotAlex1N (Elo 1800-1999) – Book Depth 10
♟ https://lichess.org/@/BotAlexCandidateM (Elo 2000-2199) – Book Depth 14
♟ https://lichess.org/@/BotAlexM (Elo 2200-2399) – Book Depth 18
♟ https://lichess.org/@/BotAlexMI (Elo 2400-2599) – Book Depth 22
♟ https://lichess.org/@/BotAlexGrandMaster(Elo 2600-2882) – Book Depth 28

Each bot accurately simulates the playing strength of its respective Elo range!

But that’s not all! Soon, we'll be introducing avatars of famous players, simulating not just their strength but also their unique playing style.

Stay tuned for more updates and don’t miss this unique opportunity!

[cpeters]

Hello.

Not working ("accurately") as intended - as you can see for yourself, sadly. The recommendation to test things like this beforehand and offline (various maja-builds are available) comes to late.

Why are you doing this?

[supernova]

Test Conditions
- OS: Windows 11 Home - HW: 13th Gen Intel(R) Core(TM) i7-13700HX, Lenovo 5 Legion
- Gui: Cute Chess Gui - Tournameent Type: Gaunlet - Time Setup: 40m/1m
- Games by Match: 100 as much as possible.
- Ponder: ON - Hash Table: 512Mb - Threads: 1 - Syzygy endgame tablebases: 3,4,5 complete + most important 6,7.
- Opening Book: Different Opening Test Sets consolidated in Polyglot Book
and progressively removing bad opening lines.
- PGN games are available upon request.

Code: Select all

ShashChess 38 PL.On-SS.AllOn :
	BrainLearn 4.2.1    							:  120  :  72+  :  48=  :   0-  :   80.0%
	BrainLearn Santiago 							:  120  :   1+  : 117=  :   2-  :   49.58%
	Brainlearn 30 PL.Off 							:  120  :   6+  : 110=  :   4-  :   50.83%
	Brainlearn 30 PL.On 							:  100  :   1+  :  95=  :   4-  :   48.5%
	CFish 20230115_FNNUE_GNv3_nn-76a8a7ffb820		 	        :  150  :   6+  : 141=  :   3-  :   51.0%
	Cfish 12.7          						        :  120  :  32+  :  88=  :   0-  :   63.33%
	ShashChess 37 PL.On-SS.AllOn 					        :  120  :   2+  : 115=  :   3-  :   49.58%
	ShashChess 38 MCTS.Ex.1-PL.On-SS.AllOn 			                :  120  :   4+  : 114=  :   2-  :   50.83%
	ShashChess 38 MCTS.Ex.1-PL.On-SS.AllOn-ChessDB 	                        :  120  :   9+  : 109=  :   2-  :   52.92%
	ShashChess 38 MCTS.Ex.2-PL.On-SS.AllOn 			                :  120  :   4+  : 116=  :   0-  :   51.67%
	ShashChess 38 PL.On-SS.Capablanca 				        :  120  :   5+  : 111=  :   4-  :   50.42%
	ShashChess 38 PL.On-SS.Petrosian 				        :  120  :   3+  : 116=  :   1-  :   50.83%
	ShashChess 38 PL.On-SS.Tal 					        :  120  :   3+  : 116=  :   1-  :   50.83%
	Stockfish 15        							:  100  :  12+  :  82=  :   6-  :   53.0%
	Stockfish 15.1      							:  100  :   8+  :  90=  :   2-  :   53.0%
	Stockfish 16        							:  100  :   6+  :  90=  :   4-  :   51.0%
	Stockfish 16.1      							:  100  :   4+  :  89=  :   7-  :   48.5%
	Stockfish 17        							:  100  :   5+  :  84=  :  11-  :   47.0%
	Stockfish Dev.25011820		 				        :  120  :   1+  : 118=  :   1-  :   50.0%
	Stockfish Dev.25022808		 				        :  100  :   6+  :  85=  :   9-  :   48.5%
	Vanilla 13          							:  100  :   7+  :  91=  :   2-  :   52.5%
	Vanilla 14c         							:  120  :   5+  : 112=  :   3-  :   50.83%
	Vanilla 15          							:  120  :   6+  : 114=  :   0-  :   52.5%
	
Remark: ShashChess 38 Persistent Learning On at the same level of Stockfish 16.

I am experiencing persistent issues with all versions of ShashChess when two specific options are enabled simultaneously:

• MCTS (Monte Carlo Tree Search)

• ChessDB Options, particularly when set to query ChessDB Endgame Tablebases

Observed Problems:

• It consistently stops working or loses on time when both options are active.

• Increasing the Overhead Time to generous values (up to 1000 and 1500) does not resolve the issue.

• It malfunction is also replicated during Pondering phases when these options are enabled.

[criko]

Test Conditions
- OS: Windows 11 Home - HW: 13th Gen Intel(R) Core(TM) i7-13700HX, Lenovo 5 Legion
- Gui: Cute Chess Gui - Tournameent Type: Gaunlet - Time Setup: 40m/1m
- Games by Match: 100 as much as possible.
- Ponder: ON - Hash Table: 512Mb - Threads: 1 - Syzygy endgame tablebases: 3,4,5 complete + most important 6,7.
- Opening Book: Different Opening Test Sets consolidated in Polyglot Book
and progressively removing bad opening lines.
- PGN games are available upon request.

Code: Select all

ShashChess 38 PL.On-SS.AllOn :
	BrainLearn 4.2.1    							:  120  :  72+  :  48=  :   0-  :   80.0%
	BrainLearn Santiago 							:  120  :   1+  : 117=  :   2-  :   49.58%
	Brainlearn 30 PL.Off 							:  120  :   6+  : 110=  :   4-  :   50.83%
	Brainlearn 30 PL.On 							:  100  :   1+  :  95=  :   4-  :   48.5%
	CFish 20230115_FNNUE_GNv3_nn-76a8a7ffb820		 	        :  150  :   6+  : 141=  :   3-  :   51.0%
	Cfish 12.7          						        :  120  :  32+  :  88=  :   0-  :   63.33%
	ShashChess 37 PL.On-SS.AllOn 					        :  120  :   2+  : 115=  :   3-  :   49.58%
	ShashChess 38 MCTS.Ex.1-PL.On-SS.AllOn 			                :  120  :   4+  : 114=  :   2-  :   50.83%
	ShashChess 38 MCTS.Ex.1-PL.On-SS.AllOn-ChessDB 	                        :  120  :   9+  : 109=  :   2-  :   52.92%
	ShashChess 38 MCTS.Ex.2-PL.On-SS.AllOn 			                :  120  :   4+  : 116=  :   0-  :   51.67%
	ShashChess 38 PL.On-SS.Capablanca 				        :  120  :   5+  : 111=  :   4-  :   50.42%
	ShashChess 38 PL.On-SS.Petrosian 				        :  120  :   3+  : 116=  :   1-  :   50.83%
	ShashChess 38 PL.On-SS.Tal 					        :  120  :   3+  : 116=  :   1-  :   50.83%
	Stockfish 15        							:  100  :  12+  :  82=  :   6-  :   53.0%
	Stockfish 15.1      							:  100  :   8+  :  90=  :   2-  :   53.0%
	Stockfish 16        							:  100  :   6+  :  90=  :   4-  :   51.0%
	Stockfish 16.1      							:  100  :   4+  :  89=  :   7-  :   48.5%
	Stockfish 17        							:  100  :   5+  :  84=  :  11-  :   47.0%
	Stockfish Dev.25011820		 				        :  120  :   1+  : 118=  :   1-  :   50.0%
	Stockfish Dev.25022808		 				        :  100  :   6+  :  85=  :   9-  :   48.5%
	Vanilla 13          							:  100  :   7+  :  91=  :   2-  :   52.5%
	Vanilla 14c         							:  120  :   5+  : 112=  :   3-  :   50.83%
	Vanilla 15          							:  120  :   6+  : 114=  :   0-  :   52.5%
	
Remark: ShashChess 38 Persistent Learning On at the same level of Stockfish 16.

I am experiencing persistent issues with all versions of ShashChess when two specific options are enabled simultaneously:

• MCTS (Monte Carlo Tree Search)

• ChessDB Options, particularly when set to query ChessDB Endgame Tablebases

Observed Problems:

• It consistently stops working or loses on time when both options are active.

• Increasing the Overhead Time to generous values (up to 1000 and 1500) does not resolve the issue.

• It malfunction is also replicated during Pondering phases when these options are enabled.

Moderation: Removed content that is not in line with the charter (3. Do not contain personal and/or libelous attacks on others as well as 6. Are not bullying, or vicarious bullying). Please re-read the charter: viewtopic.php?t=84538

[amchess]

Thanks.
Andrea