SlowChess Blitz Classic 2.0

[Jim Logan]

There should be no problem. No doubt you have unzipped the file, and you see the engine. If you click the engine directly, it will open it's own interface. However, if you are in a Fritz interface, say Chessbase, simply "Create UCI Engine" and select that same file. No need to do anything else. The engine will be installed and you're good to go.

[Damir]

Sorry but it is not working. Not going to waste anymore of my time on this worthless engine..

[Modern Times]

There should be no problem. No doubt you have unzipped the file, and you see the engine. If you click the engine directly, it will open it's own interface. However, if you are in a Fritz interface, say Chessbase, simply "Create UCI Engine" and select that same file. No need to do anything else. The engine will be installed and you're good to go.

No, for some reason it will not install in the normal way in Fritz GUI (V17 anyway). You have to do it manually as AdminX sets out.

[Damir]

Ted if you have time can you log on my computer tomorrow via AnyDesk?
I will give you access so you can show me how to install the engine.
My nick on Playchess server is Pooper Scooper

[AdminX]

Ted if you have time can you log on my computer tomorrow via AnyDesk?
I will give you access so you can show me how to install the engine.
My nick on Playchess server is Pooper Scooper

Sure, just PM me when you are ready.

[Damir]

Will do that.

[jonkr]

Frank and Thomas,

Thanks for the encouragement. I've often ended up back at it developing, though this time I mostly stepped away from being active for months before deciding to release.

I think the engine is strong enough / tested enough that random ideas unlikely to work, at least not obviously. For instance when I first re-started development years ago, I added a tactical move test guided search & Q-search and got +30 Elo, I think was around 2900 computer chess Elo then. Now I'm thinking probably the idea could be simplified and tuned away, probably just any selectivity was helpful at that point, but a tweaked variation remains since easier than removing it. But was fun to test out and see results.

I tried experimenting with non playing-strength related tests, such as adding stats to my auto-training for moves to end of a won game + (<=30 moves, <=40, <=50, <=60). It's interesting but attempts to tune for it I don't think were very successful, where a small Elo difference remains unclear with many games, stats like quick-ended games vary to a significantly greater degree since they are only a smallish sub-set of games, and vary to an even greater degree depending on openings and opponents. Below is one of the better results from self-play, but then in other situations wasn't so noticeable :

Code: Select all

31. Slow2.9 - Slow 2.8 @ 30+0.3 	( 22.4 Elo ) 	+549 	-420 	=1031 	53.2% 	=51.5%	 +2, 5, 41, 157	 -0, 1, 16, 86

Also reaching the top divisions of online tournaments TCEC and CCC was fun to do a few times, but obviously not going to approach top 3 engines.

About Fritz Gui :
Unfortunately I don't have access to any recent Fritz Gui so I am not sure what is causing the installation issue or how to work around it. (I haven't purchased anything from ChessBase in quite a long time.) I don't think anything with Slow Chess protocol handling has changed. The engine itself supports UCI and Winboard, and recently I test in CuteChess on windows or linux, but previously I have also used Arena and Winboard without issues.

[Sylwy]

CMCanavessi :

If both bitbases and tablebases are loaded Slow will use both. The Syzygy tables are queried during the main search, and the bitbases are queried on any evaluation (but they are small and only certain 4-piece or less endgames.) Not printing the syzygy loaded info string to UCI is an oversight will be fixed next version. Since having a default console window is compile-time setting and I haven't found myself wanting to use a uci command line, I'm not sure if I'll get to it, but finding a way to handle it on startup it's on my list of things to look into.

Hi Jonathan

1.-Is your engine UCI or winboard engine?

1.-I have tried and create it as new engine in Fritz GUI. The author field is empty and engine does not want to be created?

I am using avx2 compile. My hardware is AMD 3990x 64 cores.

1.-SlowChess Blitz is both WB & UCI chess engine.......

2.-It's so simple to install........even in Fritz 18 GUI........Take a look to my installation in Arena 3.5.1 GI.......a simple command line parameter..... Success !!!

[Thomas Lagershausen]

Frank and Thomas,
I think the engine is strong enough /

V 2.9 is fantastic.

In todays game of worldchampion Carlsen it plays the complete preparation after 13.Be3 in seconds.

Absolutly great.

[pgn][Event "10th Norway Chess 2022"]
[Site "Stavanger NOR"]
[Date "2022.06.06"]
[Round "6"]
[White "Carlsen, M"]
[Black "Mamedyarov, S"]
[Result "*"]
[BlackElo "2759"]
[ECO "D78"]
[Opening "Grünfeld-Verteidigung mit 3.g3, 6.O-O c6 7.Nc3 dxc4"]
[WhiteElo "2864"]
[TimeControl "40/3600:40/3600:40/3600"]
[Termination "unterminated"]
[PlyCount "45"]
[WhiteType "human"]
[BlackType "human"]

1. Nf3 d5 2. g3 g6 3. Bg2 Bg7 4. d4 Nf6 5. O-O O-O 6. c4 c6 7. Nc3 dxc4 8.
e4 Bg4 9. h3 Bxf3 10. Bxf3 e5 11. dxe5 Nfd7 12. e6 fxe6 13. Be3 Qe7 14. Bg2
Na6 15. Qe2 Nb4 16. Qxc4 Nc2 17. Nd5 cxd5 18. Qxc2 d4 19. Bd2 Rac8 20. Qb3
Nc5 21. Qa3 Qd7 22. Rac1 b6 23. e5 *[/pgn]

[Frank Quisinsky]

Jonathan,

yes, same opinion Thomas have!

Is fact that your engine is fantastic!
No doubt about it.

- move average is great.
The master of desaster Frank made a mistake. The only engine I set a contempt. All others in my list are playing with standard settings. I don't know why I do that today. I think I was drunk, better I forgot the reason for it. Slow 2.83 have a move average from 100 moves in my list, but 81 moves only for won games. In my test games on other systems with 40 moves in 8 minutes the move average is not 100, the move average is very normally with around 88 moves (without contempt) !!



KI-Ratings:
https://www.amateurschach.de/t-ki/common/ki-ratings.jpg

If I start my list again ... thinking with FCP Tourney-2023 ... I set all engines to contempt = 0.
I have to reduce energy.

- aggressiveness is super!
- technical your engine is super, with the own GUI a lot of settings are possible. The opening book settings are great.
- with such small nn files to create such a strong engine is a wonder! A lot of originalty in your engine inside, should be clear for every one looking in stats.

John like the endgame from Slow. I like the earlier mid-game from Slow (not looking with many energy in all the endgames, more what engines do after the openings).

No idea what you can make better!
Maybe looking a bit in the very few lost games, I don't know?

Have the same opinion.
To play thousands of games for make an engine 10 Elo stronger is ... questionable ... if the program have a strong playing strength and much more important ... a great playing style. The day will come John say ... it is enough. A bad situation for me because 10 years later I have 500.000 games from the same version.

If Jan Lowman would be alive ...
Sure, Jan would like to say ...
Jonathan, you have created a good start point, no more and no less.
Jan like the provocation with the idea to drive to peak performances.
But all the time, with a very friendly way.
I am too stupid for this!

Best
Frank