Random Posts

Thursday, May 26, 2016

SmarThink Engine

     On the CCLR 40/40 rating list the SmarThink engine is rated way down at number 30 where it's tied with engines like Arasan 18.3, Naum 4.2, Gull 2, Rybka 2.2 and Komodo 1.3 at 2980. To be honest though, I don't think its rating should be that high because it's not played any of the top engines. But, that does not mean it may not be worth looking at.  Download it HERE.
     I came across the following game between Nisipeanu and Cornette in which white essayed a rarely played opening and I wanted to compare SmarThink's move selections against Komodo 8's because Komdo is supposed to be quite good in the opening phases. 
     Surprisingly, I found that SmarThink's selections did well in matching the GM's opening moves. For that reason I think it might be a good engine to use when looking for opening innovations.  Perhaps it has something to do with the way it incorporates  some ideas of Botvinnik's.  Of course, it's clearly weaker in the middle and end games. I ran a small 5-minute tournament using my expanded Fritz 12 Opening Book with the engines on a single core and set at 1024 Megabytes. The results, not unexpected, were: 

1) Komodo 8......x 1 1 1 = 3.0 
2) Stockfish.......0 x 1 1 = 2.0 
3) SmarThink...0 0 x 1 = 1.0 
4) Fritz 12.........0 0 0 x = 0.0 

     SmarThink, an UCI/WB compatible engine by Sergei S. Markoff with an aggressive attacking style and what makes it interesting is that it uses techniques in search and evaluation that includes, as mentioned, some ideas of Mikhail Botvinnik. SmarThink played some tournaments in Russia and was the Russian computer champion back in 2004. Only a single-processor version is available. 
     While outright blunders and forced wins are readily shown by engines, in many positions the best move may not be easy to determine. Some times it's impossible because there are so many plans and ideas that are equally good. As a result, engines are of little practical help in evaluating opening play due to the sheer number of possibilities, importance of long term strategy and lack of forcing tactics. 
     The best way to use an engine in the opening is to search for alternative moves that don't appear in the opening books. In situations without any tactics different engines may have different evaluations of the position and even today they often over-value material. Most opening innovations these days are often discovered with the help of an engine. SmarThink might be worth considering for this purpose. 

No comments:

Post a Comment