Thursday, September 10, 2009

Beatiful Position posted by Susan Polgar today


Susan has the following position on her website from a game between Kraai and Kavutskiy this weekend in the So Cal Open in San Diego.



I spent time this afternoon analyzing it from a printout of the position, and when I got home I set it up on a board and analyzed it some more. Very fun position to work on.

Some analysis below:



1.Ne7+ Kh8 2.Q:f6

A) 2..Qg4 3.Qc3 {my earlier idea of 3.Bf3 pretty much gives up the advantage gained since the N is guarding g4 and can prevent f6 from being taken g:f6 B:g4 N:g4 f3 Ne5 f4 Nd7} Nd7 4.Bf3 {White can play it now though} Qg5 5.h4 Qf6 6.Q:f6 g:f6 7.f:e3 {and white is up 2 pieces}

B) 2..e5 3.B:e5 Qg4 4.Qd6 Re8 5.Rc8 R:c8 6.B:c8 Qd1+ 7.Q:d1 N:d1 8.Bg4 {this line was found by Rybka after I did my analysis. I didn't even consider e5 as a response.}

C) 2...Q:b2 3.Q:b2 {and white has a nice material advantage.}

D) 2..Nf5 3.N:f5 Qg4

D1) 4.Q:f7 {another line found by Rybka after I completed my analysis and a nice mating net to boot} h6 5.B:g7+ Kh7 6.B:f8+ Kh8 7.Bg7+ Kh7 8.Bf6+ Qg7 9.Q:g7#

D2) 4.Rc4 Nd7 5.Q:g7+ Q:g7 6.B:g7+ Kg8 7.Rg4 {this is the line I analyzed}


Saturday, September 5, 2009

Thoughts on training from Botvinnik

Probably one of the greatest players ever at training and preparing for competition was Botvinnik. Supposedly he discussed his methods in both the Flohr - Botvinnik match book, and in the book he authored with Ragozin about the 11th Soviet Championship. The discussion in the Flohr match book consists of about 4 pages of text, and I will probably translate it sometime in the near future. There is a bit about his methods in The Dover publication of Botvinnik's 100 Selected Games.


I am in the process of tracking down a translation of the tournament book for the 11th Soviet Championship to see if it has more insights to offer than the 2 pages in 100 Selected Games. It would be an interesting book to have anyways as this was the tournament where Kotov displayed the results of the training regimen put forth in Think Like a Grandmaster.

Here are some of the things Botvinnik had to say in 100 Selected Games:

  • Begins preparation with a review of the chess literature to acquaint himself with new and interesting games, making notes on questions which are of interest to him as he reads.
  • Studies the games of his rivals in the upcoming competition. He studies their peculiarites of play, and their opening repertoire.
  • Studies the lines he intends to employ in the competition. For a competition he usually employs 3-4 systems for each color, and is thoroughly prepared in those systems.
  • Next is to try out these systems in training games.
  • He champions the use of training games to work on all aspects of your chess, using the games to focus on a certain deficiency and to continue training games with that deficiency being the focus until it is eradicated. An example being that if you get into time trouble, you should play training games where the first consideration is the clock, not the quality of play or the result. This continues until you can play at peak strength without time trouble in the time control being trained.
  • He champions resting the final few days before competition to be fresh when it starts.
  • More than anything else to reach top potential you must master the art of analysis. To this end you should publish your analysis so that others can find the faults within it. I imagine you could also use a chess engine to critique your analysis. He makes an interesting point that you should not analyze games in the first hours immediately after it is played, because such analysis will most likely be purely negative and can become a bad habit.

Chess Hero, a chess training program

I just came across this program in a forum post on chess.com today. It is a program that presents you random positions selected from pgn files that you provide it. You choose your move for the position and then it scores it against a chess engine's analysis.



Somewhat of a variation on the guess-the-move feature at chessgames.com, which has you play as 1 side through a whole game, and scores your play based on engine analysis (using toga or fruit I believe). In this case you can use either uci or xboard engines and customize the init commands. You can create different training profiles that contain a selection of pgn files to be used in that training. You can set some as being books, so it can also train you on your knowledge of an opening, and not penalize you for not playing what the engine determines is the best move. Apparently it likes the modern variation of the Alekhine's best, not the exchange variation which leads to the voronezh, which is what I am studying. Once you add a pgn file with the lines that are your book it will no longer present those positions from non-book pgn files, and it won't score them when presented from book pgn files, just tell you whether you are correct or wrong.

It seems to be a fairly new program, so it will be interesting to see how he evolves it. I had considered trying to build something like this at one time, so I am glad someone else has taken on the task. I still have some other ideas for chess training programs that I might try putting together at some point. Right now I'd rather study and play chess in my free time than program :).