Hi Marco,

Thanks for sharing this information. Regarding SugaR can you provide more information about the difference between SugaR and Stockfish. I believe you have implemented Monte Carlo search amongst other functionalities. How do you see it as playing differently to Stockfish and for searches or Correspondence chess where do you see SugaR when compared to Stockfish.

Perhaps you can explain how the Monte Carlo approach hopes to achieve and equalize the artificial intelligence approach of LC0?

Also I am curious since there is a lot of hype about LC0 and Leila what do you see as improvements with these compared to Stockfish or SugaR? Do you see the possibility that they will at some point overtake chess as the only method?

Me I am sceptical because I do not personally see a load of tables that make up an Encyclopedia as intelligence myself. I see it more as just a database of evaluated positions, but then again I am not involved in programming myself so perhaps I am just being dumb.

I tend to in my simple mind think of LC0 approach as "anyone can act smart" but are they really?

I mean by that, the simplest of person can make themselves sound intelligent in a chat room when all he has to do is search in Google to find a clever response and then copy and paste it

Creativity and intelligence to me is to look at for example Chess Genius 1 written in 1993 I think, which is a 61 KB program which even in our tournament has the ability to beat a 2600 ELO program which has maybe 20 - 40 times more code and no end game table bases and a simple opening book of 438 KB. This of course makes me wonder what a Genius 1 could achieve if it could play at the same speed as engines playing on an I7 or I9 multi-core processor.

Maybe someday soon by creating a DOSBox Windows 98 it might be possible to really compare assuming Arena were possible inside Windows 98 to see actual imrovements between say a Stockfish and Genius 1 for example at the same speed?

So I guess what I view as intelligence is a simple program that works something out by itself without having to pull information from other sources.. ie encyclopedias.

ps. since you have worked for years with the development of Stockfish teams and SugaR, what are you seeing as the future development and progression of computer chess? Do you think that there might even be a trend towards going backwards and create programs that the human or Grandmaster can enjoy and play against? I mean head to head and no messing around with putting the engine down to for example 0.01% as a setting

Best regards
Nick

SugaR uses the MCTS option (MonteCarloTreeSearch), it is activated by default, the motor behavior is similar to the AlphaZero concepts.
creates three files for machine learning purposes:
-experience.bin when no more than 40 moves are played, there are non more than 6 pieces on the chessboard and at a not low depth in analysis
-pawngame.bin
when there are no more than 2 pieces and the game's phase is not the ending
-openings.bin, in the form .bin (>=1) at the initial stage of game with memorized the move played, the depth and the score. In this mode, the engine is not less strong than Stockfish in a match play without learning, but a lot better in analysis mode and to solve hard positions. With learning, the engine became stronger and stronger. The default mode, conversely, is stronger than Stockfish in a match play, but not as good as MCTS for the rest and can't improve its play because of no learning enabled.
Example:


Dynamic Strategy
To be used as additional support in the analysis of particularly complex positions. With the increase of the score, that is how much the motor is in advantage or fundamentally closer to the checkmate. Or In all favorable pressing situations; The advanced Pawns are penalized and the King gains more importance because we must pay attention to the compactness and the other way around.

Regards
Marco

Hi Marco
Thanks for your explanations. I know you have seen SugaR entered in some International Correspondence tournaments. What feedback are you receiving from the people who are using it?

Best regards
Nick

My friend GM Pirš, Matjaž (captain of the German team) and others ... but I say no more, I will reveal everything in my little article. "YearBook 2019"

Regards
Marco

My friend GM Pirš, Matjaž (captain of the German team) and others ... but I say no more, I will reveal everything in my little article. "YearBook 2019"

Regards
Marco

Hi Marco

welcome in this forum. By the way: I use Sugar for about 1 year now in correspondence chess together with ASMFish, Stockfish and Komodo. As soon as I have a good GPU Leela will join too.

Are yu also a correspondence player (because of your article in the ASIGC). Maybe we have met at ICCF.

Greetings

Hartmut