Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Thu Oct 15, 2020 11:43 pm
On my mobile phone.
TalkChess.com
Hosted by Your Move Chess & Games - chessusa.com
http://talkchess.com/forum3/
INTERESTING Engines with Unique Styles (Unlike SF)
Page 6 of 6
Where is MChess Pro?
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Fri Oct 16, 2020 2:41 am
Just a short addition to my original "Interesting engines list"
33. Deep Shredder 10..Wow..had so much fun with Deep Shredder 10...my first copy didn't even have a opening book
and played what I considered at that time "almost flawless chess with great battles vs Chessmaster.. Junior 7 and for some reason extra tough in the Lokasoft GUI...
34. Toga 1.2.1 The Standard for tactics in its time..and still not too bad today...Still plays at GM level Tactically..
35. Tao...Tao 5.4 and v5.6 etc.. A cool name with a cool game a excellent tactician...great engine
36. Pro Deo 1.2 and Pro Deo 1.1 (Alekhine) not an easy set up but has loads of options and settings..books etc..lots of fun!
39. Arasan...had a ton of fun with Arasan lots of engine releases and is both UCI or winboard....Still relevant today!
40.Colossus...like this one nice solid style ..not too strong..not too weak..lets of different versions..
41.Greko..I like this engine...tons of versions and has a great style.. not overpowering in strength.
42.Alfil.. Nice engine...lot of different versions...Nice style...
43. Smar-Think...had a lot of fun with this engine..lots of different versions...old classics and newer versions..
44.Chess Genius..had quite a bit of fun with Chess Genius 3..Chess Genius 6 was extremely difficult to set up engines and the CG 5-7 engines were kind of generic in style...and a bit dated (w16!) Mchess Pro..was just OK..
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Fri Oct 16, 2020 7:09 am
Hmmm it seems the Sapeli repo either vanished or was set to private?hafni wrote: ↑Wed Oct 14, 2020 12:42 pm-----------------------------------------------------------
My engine Sapeli 1.93.
https://github.com/SamuraiDangyo/Sapeli :(
------------------------------------------------------------
Sorry but your link is wrong or there is a problem with github
Best regards
I hope Toni will inform us about what happened.
A lot older binaries are still there in a general bin repo.
https://github.com/SamuraiDangyo/bins/t ... i-binaries
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Fri Oct 16, 2020 11:55 am
I closed the source for a while to focus better. I rewrote lots of things. Ie much faster insertion sort algorithm vs old selection sort. Sapeli now sorts checks much higher than it used to.Guenther wrote: ↑Fri Oct 16, 2020 7:09 amHmmm it seems the Sapeli repo either vanished or was set to private?hafni wrote: ↑Wed Oct 14, 2020 12:42 pm-----------------------------------------------------------
My engine Sapeli 1.93.
https://github.com/SamuraiDangyo/Sapeli
------------------------------------------------------------
Sorry but your link is wrong or there is a problem with github
Best regards
I hope Toni will inform us about what happened.
A lot older binaries are still there in a general bin repo.
https://github.com/SamuraiDangyo/bins/t ... i-binaries
Managed to trim down 100 lines.
Code: Select all
Language files blank comment code
-------------------------------------------------------------------------------
C 1 217 80 1766
-------------------------------------------------------------------------------This game is a good example of better sorting algortithm/ tons of speedups. Horrible opening but managed to find a few threats.: https://lichess.org/yoeXD1t3/black
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Sun Oct 18, 2020 3:29 pm
I see tons of engines but not many interesting.
That is really strange.
That is really strange.
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Sun Oct 18, 2020 3:37 pm
It plays around 2600 elo and can't weaken it but open tal 1.0 is very unique in style
Re: INTERESTING Engines with Unique Styles (Unlike SF)
Posted: Sat Oct 24, 2020 11:50 pm
It turns out style is all about mistakes and inaccuracies. Nothing else.
Something quite revealing from the NNUE wave that happened is that Stockfish had a very distinct playing style, and it's completely different from Stockfish Pure NNUE. Stockfish's style seemed to be about avoiding complications, material imbalances, and exaggerating or undervaluing King Safety.
Fixing all of that was worth 100 elo.
The games between Stockfish with Classic eval and NNUE can be fantastic! When the complication is better, when the material imbalance is better, when your king safety doesn't matter and you can push all your pawns, or when you can sacrifice material to launch a king attack, NNUE would do it and kill Classic on the spot!
This made NNUE's style pop out, which was awesome because not only did we have a stronger engine (stronger than anything else!) but its style was amazing!
...but this was only by comparison!
The NNUE v NNUE games were nothing worth calling home about. And then it went hybrid.
Stockfish Classic, Stockfish Pure NNUE and Stockfish Hybrid have all very different playing styles, but why doesn't anybody talk about them? Because when you look at the games there's no move played that talks to you. Nothing memorable.
The memorable things are the inaccuracies, blunders make up the style, and what NNUE showed is that the accurate chess lines are thin. That's why NNUE would kill off any engine that tried deviating from the accurate chess lines (even the strongest Stockfish with Classic eval!)
If you showed attractive moves played by the engines mentioned in this thread, they'd all probably be inaccurate moves that would be punished by Hybrid. If style was more than blunders it'd be easy for some programmer to make a module that played Hybrid's moves most of the time, but had the stylistic engine take over if it suggested an attractive move. This is impossible because attractive moves just can't perform well.
Accurate chess is boring, that's why the games we tend to remember are the losses and victories.