Tag Archives: Boris Gelfand

Gelfand’s Lofty Standard

This review has been printed in the February 2017 issue of Chess Life.  A penultimate (and unedited) version of the review is reproduced here. My thanks to the good folks at Chess Life for allowing me to do so.

—–

Gelfand, Boris (with Jacob Aagaard). Dynamic Decision Making in Chess. Glasgow: Quality Chess, 2016. ISBN 978-1784830137. HB 288pp. List $34.95.

Positional Decision Making in Chess, the first volume in Boris Gelfand’s ‘Decision Making’ series, was published by Quality Chess in 2015 to critical acclaim. (See the September 2015 issue of Chess Life for my rapturous review.) Now Gelfand’s second book, Dynamic Decision Making in Chess, is available. Does it live up to the lofty standard set by its predecessor?

The title of Gelfand’s new book accurately describes its contents. His main theme is decision making, with a particular focus on (a) how Gelfand makes practical decisions over the board and (b) his handling of dynamic positions. While Gelfand’s articulation of his thought processes is clear and mainly successful, the lens he uses (dynamical play) makes its complete exploration very difficult.

Gelfand describes the “core” of his books as follows:

I want to explain the thinking that has led to my reasonable success as a chess player, and not ‘cheat’ in the process. It is quite easy to analyze a variation with the engine and then explain why it works. And this certainly has its uses, but to me it is more interesting to talk about how we find the moves in the first place. This is the key to playing better chess. (260)

The goal of the books in this series thus far is to offer an honest accounting of how a super GM like Gelfand decides on his moves. The analysis tries to follow Gelfand’s in-game stream of consciousness, and because he cuts no corners, it can be incredibly complex. A recurring theme of the book is Gelfand’s warnings about overreliance on the computer.

It is a mistake to assume that Grandmasters think like engines. Because humans cannot begin to match the machine in terms of calculation, because we can’t see everything like the computer does, at some point we have to “guess.” (8, 86) Decision making on the basis of limited information (guessing) relies on intuition, evaluation, and judgment. (160, 218, 226).

Gelfand’s point seems to be this: humans cannot calculate their way to good decisions. We must rely on “general considerations” (15) while we play, and we must use our intuition to take decisions that we cannot fully calculate. How do we train intuition, and in this case, how do we train our sense of dynamics?

There’s the rub.

It’s important to be clear about what we’re talking about. Dynamics involves the ephemeral in chess. Some temporary feature of the position must be converted into an lasting advantage before it dissipates. (8) Dynamic chess involves intuition and calculation for Gelfand, but devolves to neither. (9). It is not strictly tactical or strategic in nature, the very distinction being somewhat artificial in his view. (61)

For all of the analysis in Dynamic Decision Making in Chess, and for all of the exquisitely careful explanation of decisions and thought processes, there is nothing to my eye that explains how Gelfand senses dynamism in a position. He just does, and more than that is hard to explain.

This is not a knock on Gelfand (or his co-author Jacob Aagaard). Dynamic Decision Making in Chess is a wonderful book, one of the best of its kind, but like every book on dynamics, there comes a point where analysis and explanation fail and we must simply bear witness to genius.

Again, let me be clear. I am not claiming that the great moves of the masters are somehow ineffable or beyond reason. Instead, what I am arguing follows from the block quote above.

It is easy, as Gelfand notes, to retroactively explain the logic of a brilliant move. What is more difficult is clearly articulating the move’s genesis without falling prey to what John Dewey called the ‘philosopher’s fallacy,’ where the results of analysis are taken to accurately represent what was experienced before analysis began.

Studies of dynamic play are, in my experience, particularly susceptible to this kind of fallacy. While Gelfand works diligently to break down the logic of his best moves – his 11…Ra6!! against Karjakin from the 2009 World Cup, for instance (227-239) – there is a level of analysis beyond which he cannot go. It took him 40 years of study and solving (54, 134) to find such moves, and their intuitive, unconscious origins are not easily excavated.

Does this mean that Dynamic Decision Making in Chess fails in its project? Absolutely not. It may lack the clarity and focus of Positional Decision Making in Chess, but this is due to Gelfand’s ambitious handling of a very difficult subject and his refusal to simplify his thought processes for the sake of expediency. The analysis is best suited for experts and above, but players of all strengths can’t help but learn from this book.

Advertisements

Learning from Gelfand

This review has been printed in the September 2015 issue of Chess Life.  A penultimate (and unedited) version of the review is reproduced here.  My thanks to the good folks at Chess Life for allowing me to do so.

—–

Gelfand, Boris, and Jacob Aagaard. Positional Decision Making in Chess. Glasgow: Quality Chess, 2015. ISBN 978-1-78483-006-9. 288pp. HB $34.95. [Note that Quality Chess has only released the hard cover version to specialized chess retailers, and a paperback should be on Amazon in the nearish future.]

Positional Decision Making in Chess is Boris Gelfand’s second book, the first being his 2005 My Most Memorable Games. Were it simply another batch of his annotated games, it would well be worth our attention. Very few of the world’s elite put pen to paper (fingers to keys?) while they are still active players.

Most of Kasparov’s many books emerged only after his retirement. Books by Anand and Kramnik predate their World Championship reigns, while the bulk of Shirov’s output now comes in DVD form. Recent works by Giri and Polgar are excellent, but Giri’s best years are ahead of him while Polgar has retired from tournament play.

So when a player like Boris Gelfand – a six-time Candidate, the 2012 Challenger for the World Championship and the 13th ranked active player in the world – writes [1] a book about his games, we chess bibliophiles tend to take notice. And all the more in this case, for Gelfand has given us a superlative book.

My Most Memorable Games is, on the whole, a traditional ‘best games’ collection. It is evident from even the first pages of Positional Decision Making in Chess that Gelfand has something else in mind with his new book. As he writes in the Preface,

…the intention of this book is not to focus on the accuracy of the moves I made at the board… but on the thought process that led me to finding them in the first place. … [T]hroughout we have focused on the reasons for the decisions and plans I made, and also the limitations of my thinking during the game. (8)

While (sometimes copious) analysis of Gelfand’s games is provided, the real focus of the book is how Gelfand takes decisions over the board, with positional decisions front and center. The games of Akiba Rubinstein – Gelfand’s favorite player – are enlisted in this effort, and special emphasis is placed on Rubinstein’s influence on Gelfand along with his relevance for contemporary chess theory.

There is much to like here. It’s good to see Rubinstein get his due as player and theoretician, especially as there are very few legitimate books about him in print. Gelfand’s annotations are clear, and his descriptions of his opponents are both respectful and revealing. The book’s surprisingly personal feel is amplified by the photographs strewn throughout its pages.

For me, however, the central theme of the book only appears between the lines of the text: Gelfand’s relationship to the computer. No one can dispute the changes wrought on chess and its play by our silicon friends. Nor, if we are honest, can we overlook the way in which most players trust engine evaluations blindly, almost outsourcing their thinking to the computer. (Look at Twitter or the ICC chat during the next big tournament if you doubt this.)

What is most interesting to me about Positional Decision Making in Chess is seeing how Gelfand, a member of the last generation to come of age before the rise of the machines, thinks about engines and their limitations. Gelfand trusts his intuition – this word appears repeatedly in the text – and prefers to view engines as tools for understanding rather than as infallible oracles. Rarely have I seen such honest and practical discussion of the topic. For instance:

…I am a strong believer in the value of a chess education built on thorough knowledge of the classics [like Rubinstein – JH]. Any attempt to emulate the engines and their 2,000,000 moves a second is doomed to fail. We need to supplement calculation with all other weapons available. And one of these is intuition, which is strongly rooted in pattern recognition. (58)

Extremely often the computer will suggest moves that no human would consider. And when we do not feel it delivers us a clear understanding of why this move is good, I cannot see that it makes sense to follow its recommendations. (199)

If only those kibitzers on ICC would heed Gelfand’s warning!

By providing us a window into his decision making, and by showing us – warts and all – both the limits and triumphs of his thought, Boris Gelfand does much more than merely offer us edifying games to study. The author of Positional Decision Making in Chess is an exemplar for all of us who struggle to learn from the computer without succumbing to its siren call. This might well be the book of the year, and serious students of modern chess practice should not miss out on its lessons.


[1] I would be remiss if I did not mention the role of Gelfand’s ‘helper,’ Jacob Aagaard, in the construction of this book. Aagaard, himself a very well regarded author and pedagogue, recorded extensive discussions with Gelfand and used them as the basis for the written text. It appears that most of the conceptual content should be attributed to Gelfand, while the style, structure, and some of the pedagogy are Aagaard’s.