Monday, September 27, 2010

Modern Chess Analysis by Robin Smith

I first came across a review for this book a couple years ago and found the review interesting, but never actually saw it for sale anywhere. I suspect the reason is because it’s not a book about improving your play using normal methods. It’s a book that focuses on using computers to improve your chess and attempts to assist serious players in improving their analysis.


Robin Smith is a correspondence chess grandmaster title and has twice been USA correspondence champion. He also won a world correspondence chess championship semi-final. SO Smith plays chess in a world where use of engines to assist in researching games and finding plans is a major factor.

I suspect one reason it turned out not to be a hot seller is because Smith starts with what to look for in purchasing a computer for use in playing chess. He discusses things like processor speed, memory, hash tables, etc. and even recommends specific models of computers. The problem is the book was published in 2004 so all that information was probably outdated even before the book was published! All top level CC players have the latest software and hardware, something most of us are unlikely be concerned with. Many of them run linked computers, let them analyze overnight, etc. That’s a lot of work, but that’s what it takes if you’re a titled CC player!

Despite the technical aspect which, unless you’re a computer geek as opposed to a chess geek, his discussion about the strengths of computers versus human players is quite interesting. Of course we all know engines are superior in doing tactical calculations and humans are better at long range planning. Another thing that probably outdated the book fairly quickly was his comparison of the strengths and weaknesses of various chess playing programs. All of that has probably changed in the last 10 years, making it all irrelevant today. Again, if you are technically oriented he discusses how different engines come up with their numerical assessment of a position. He also shows types of positions where computer evaluation is inaccurate.

He covers exceptions to the chess “rules” that engines use in deciding the evaluations of moves. What makes this interesting is his belief that unusual material imbalances are difficult engines to evaluate. This is encouraging because if you played through a couple games I posted on this Blog, you’ll remember there were some positions with material imbalances where I said I wasn’t so sure of the computers optimistic evaluation. Maybe I was right! Smith’s explanation as to when computer-assisted analysis is useful, and when it should be treated with caution, is instructive.

He then examines various methods for using the computer for analysis. Surprisingly he says the best use of the computer is as a sparring partner, where you try to come up with plans and then use the computer to check them. One disadvantage to this I see is that most of us aren’t strong enough to do that! It also validates my long standing claim that to rise to the top of the correspondence chess world you have to be a pretty good player to begin with. There’s more to playing at that level than just buying an engine. To rise to the top, you are going to have to be able to use your engine and your brain to defeat the other guys who are just using an engine to generate their moves.

Strong CC players recommend using more than one engine when analyzing because as Smith points out, there can be substantially differing results based on the program. He advises caution about reading too much into computer evaluations and he’s not the first strong CC player who has said this. Smith claims engine evaluations can’t always be trusted, even in tactical situations so he recommends conducting an engine tournament to test the evaluations. He also examines practical computer use and covers areas where computers may not be particularly useful.

So the short version is that those of us who thought engine evaluations were infallible will have our bubble burst. In short I don’t think this book will be of much use for the average player. It’s value might be in helping some aspiring correspondence players break through the gaggle of engine users lurking above a certain level in serious international play. Of course the real fly in the ointment here is that most of us don’t know when an engine’s output is useless and when it isn’t.

Still, it is an interesting read if you are just curious and that is the reason for this review. You can download a Windjview version of the book free. You will need the free Windjview program which is available HERE.

And then you can download the book itself HERE.

2 comments:

  1. Robin passed away a couple of years ago. Surprised to see this on public domain but maybe that is the reason

    ReplyDelete
  2. So is this file sharing site legal? It seems a little shady to me. I would prefer the author (or his family) get paid for his hard work.

    ReplyDelete