SashChess is based on Stockfish. The difference is that this engine is an attempt to apply Alexander Shashin's theory based on his book Best Play: A New Method For Discovering The Strongest Move that was published in 2013.Read book review HERE
In the book physicist and master Alexander Shashin breaks down the position into mathematical ratios that compare the elements of material, mobility, safety and space which, supposedly, leads you to the proper plan and the right mental attitude in light of what’s happening on the board.
Based on the games of three world champions with distinctive playing styles (Tal, Capablanca, and Petrosian) and backed up by personal and computer-aided analysis, Shashin attempts to form guidelines to help make decisions in different situations, including those too wild and murky to provide clear-cut conclusions.
SashChess is based on Stockfish and the book. Depending on the position the engine has algorithms modeled after the play of Tal, Capablanca and Petrosian as well as "mixed," so it plays differently, based on the type of position it is analyzing.
Additionally, the engine has a GoldDigger mode in which it favors depth over the pruning of variations. In this mode it is supposed to be able to uncover hidden possibilities although it loses slightly in playing strength. It also means it will require more thinking time. I should mention that this mode does not seem suitable for auto-analysis in Fritz (or the tactical analysis as it's called in ChessBase 16) nor is it good for use in engine speed tournaments because it is slower.
It's this GoldDigger mode that I was interested in for correspondence play. As a test I entered an event on Lechenicher SchachServer where engine use is allowed. My opponents were all in the 2100 range and one had an ICCF title.
Of course, there was no way of knowing what engines they were using or how much time they were allowing engines to think before selecting their move, but I suspect that Stockfish was the norm and since most moved fairly quickly I assume their engine "thinking" time was not especially long.
For my part, all games were played in the GoldDigger mode and the minimum engine thinking time on three cores was five minutes, sometimes 15-30 minutes.
My main observation was that there was little difference between SashChess' moves and those of whatever engines my opponents were using because in none of the games did they make a move that was totally unexpected. All of the games were uneventful draws.
So, while SashChess is a most interesting concept, in my opinion it did not seem to offer any real advantage over Stockfish 15 which is still my preferred engine for analysis.
Speaking of engines, I also have Komodo 14 which I seldom use, but for those that are interested you can download Komodo 13 free from their site HERE.
The claim is that Komodo's advantage over brute force engines is its positional style of play and the fact that it relies on positional evaluation over depth. This means that "when most engines can't find a good plan, Komodo can seemingly create something out of nothing."
Komodo 14 also has different "personalities" the names of which are self-explanatory: aggressive, defensive, active, positional, endgame, beginner and human. The idea is to simulate different styles. The ChessBase site gives an example...say if you think the position offers enough attacking potential to warrant playing for a win then you can switch to the aggressive mode, etc. I think this sounds like sales hype or may it's something similar to SashChess' different modes. Testing it out on Lechenicher SchachServer is not something I want to bother with.
All that said, on the CCLR 40/15 rating list there is not all that much difference between Stockfish, Komodo Dragon and SashChess! If you are interested you can download SashChess HERE
No comments:
Post a Comment