Fire On Board

Shirov

     ...by Alexei Shirov is on of my favorite books. Shirov (July 4, 1972) is a Latvian and Spanish GM whose FIDE rating is 2636 which doesn't even put him in the top 100, but in 1994 he was ranked number two in the world. 
     He was the world under-16 champion in 1988, at the world under-20 championship in 1990 he tied for first with Ilya Gurevich, but finished second on tiebreaks. Awarded the GM title in 1990, Shirov was the winner of numerous strong international tournaments in the 1990s. 
     In his youth he studied under Mikhail Tal who strongly influenced his style and helped mold his attacking style. Shirov wrote that he always tried not to be just a tactician and worked had on studying strategy and endings. In fact, he believed that the endgame was his strongest area. In the forward to the book he wrote, "If you want to become involved in a world of unusual ideas, sacrifices, sharp moves and sometimes strange errors, then I believe that this is the book for you." 
     Two games in the book held a special fascination for me. The very first game was one he played in1983 against an IM named Valery Zhuralev in which Shirov sacrificed two pieces in a row.  The first R-sac couldn't be taken and the second, a B-sac, was risky. 
     Shirov wrote that the following game against Jeroen Piket played at Amsterdam in 1995 was one that gave him more pleasure annotating it than actually playing it. In his analysis of all the game he was assisted by Fritz 4 and Chess Genius. One thing I appreciated is that Shirov checked his analysis with engines, but the actual analysis was his own. And, he was very meticulous. Even going over the games with the aid of a modern day engine reveals few errors in analysis. 

Piket

     Piket (born in 1969) earned his GM title in 1989 and won the Dutch Championship in 1990, 1991, 1992 and 1994. He drew a match against Anatoly Karpov in 1999 that was held in Monaco in which all eight games were drawn. In 2000, he won an internet tournament organized by kasparovchess.com and beat Kasparov in the final. In 2001, he retired from chess to become the personal secretary of businessman Joop van Oosterom. A few years later, in 2005, Van Oosterom won the Correspondence World Championship, causing Tim Krabbe to write: "The Turk was operated by William Schlumberger, Mephisto was operated by Isidore Gunsberg, Ajeeb was operated by Harry Pillsbury and Joop van Oosterom is operated by Jeroen Piket." 
     Of course everybody is familiar with Fritz, but probably not Chess Genius. It was a program written by Richard Lang who has in the past written programs that have won the World Computer Chess Championship on 10 occasions. It was a continuation of a series of programs (which included various incarnations of the Mephisto program. Running on a Pentium PC, it was the first program to beat a world champion, Garry Kasparov, at a non-blitz time limit. One of the first master-strength programs, in an article comparing Chess Genius with Fritz John Nunn wrote that for raw playing strength one should go for Chess Genius.
     In the game, Piket's 23rd move allowed Shirov to win almost by force, but Shirov suggested that instead of 23...Re6 black should have played 23...Qc5 which he would then have answered with 24.Bg5! 
     What is so fascinating about that line is Shirov's analysis and his comment that such a deep analysis would never be made by an engine by itself. He added that his human analysis combined with engines was a good example of "combining two brains to bring the art of chess investigation to an entirely new level." That's interesting because, hate it or love it, that is exactly what modern day top level correspondence players are trying to do. GMs also do the same thing in their opening research. 
     What got me curious was how would Stockfish and Komodo manage the position after 23...Qc5? Would they find 24.Bg5 and how would they evaluate it compared to their top choice if they suggested a different move? 

Here are the recommendations of Stockfish 9 and Komodo 10 after 20 minutes:
Stockfish: 24. g4 (2.83) 
Komodo: 24. g4 (2.03) 

How about their evaluation of Shirov's suggestion of 24.Bg5? 
Stockfish: 24.Bg5 (0.47) 
Komodo: 24.Bg5 (0.55) 

     As a curiosity, I have a note penciled in the margin that Xtreme Chess recommended 24.Qe2 with an evaluation of 0.41 in white's favor. Xtreme was based on Fritz and I can't be sure of the exact date, but it came out sometime in the 1990s and it was advertised that it worked on Windows 3.1 and Windows 95. I purchased it at Radio Shack for around $20. 

     It was somewhat surprising that after 23...Qc5 Fritz 12 also evaluated 24.Qe2 as the best move and gave white the edge by 1.31. Engines have greatly advanced since the 1990s! 
     This game was played in the second Donner Memorial Tournament and was held in Amsterdam in August of 1995. Shirov at 2695 was top seeded, but first place was shared by Timman. who at 2590 was one of the lower rated players as was Julio Granda Zuniga, rated 2605. Timman's only loss was to Piket while Granda Zuniga lost to Pikek and Judit Polgar. Shirov suffered three defeats: Granda Zuniga, Polgar and John Nunn. 

1) Timman and Granda 7.5-3.5 
3) Polgar 7.0-4.0 
4-6) Seirawan, Huzman and Shirov 6.0-5.0 
7-10) Morozevich, Salov, Nunn and Khalifman 5.0-6.0 
11-12) Piket and Van Wely 3.0-8.0