It's all just the number of patterns you learn

http://www.amazon.com/gp/cdp/member-reviews/A8AMJZ8V1H36O/ref=cm_cr_dp_auth_rev?ie=UTF8&sort_by=MostRecentReview


http://www.amazon.com/gp/cdp/member-reviews/A8AMJZ8V1H36O/ref=cm_cr_dp_auth_rev?ie=UTF8&sort_by=MostRecentReview

This is the book that has spawned a large number of other books on the subject of the Grandmaster's thinking processes. It is a classic and for that reason gets 3 stars from me. However, I do believe that the premise upon which this book is based is flawed and for that reason I do not give it a higher rating. Let me explain what I mean.

In this book, Kotov outlines his theory on why GMs are better than IMs, why IMs are better than FMs etc. It all has to do with analysis. They analyse better. Yes, yes, yes. He is right. They do. But why do GMs analyse better? This is the key question. I think Kotov got the answer wrong.

Kotov claims that he was a poor analyst, but that he improved by doing regular exercises in which he analysed complex positions, writing down all the variations. Each position was analysed only once to create a "Tree of Analysis". Candidate moves are chosen and then each move analysed one by one, branch by branch until the analysis is complete. The problem with this idea is that if flies in the face of contradictory evidence that this approach works. I DO agree with Kotov that improvement in analysis is the key to becoming a stronger player. I do not agree that his method will do more than produce a small change in your playing strength.

The contradictory evidence:
1) As so clearly pointed out by Richard Reti in his classic "New Ideas in Chess" even if there is a choice of only 3 moves at each branch point in the tree of analysis, the number of branches becomes so thick that it is impossible to analyse each branch. What distinguishes a titled players analysis from the analysis of a weaker player is the ability to EXCLUDE irrelevant moves, not include ALL moves. Humans will never be like computers in this regard.
2) Brain imaging studies show very clearly that GMs different from IMs and so on down the food chain by their use of memory patterns. They do not think more deeply. This is clearly the conclusion of a number of serious studies on the psychological basis of chess talent (de Groot for example).
3) Strong players have the capacity to hold positions in their heads more easily that weaker players. But it is clear that this is due to pattern recognition. Give a strong player a random position and they are no more able to remember the position that a weak player. This then is the basis of the stronger players greater ability to visualise the board. It is memory for patterns which creates vision. For example, it doesn't take much effort to remember the position Pf2, Pg3, Ph2, Bg2, Rf1 and Kg1. You can visualise it and remember it in seconds.
4) It is clear that strong players perform extremely well at rapid chess. Visit ICC (www.chessclub.com) and watch a lightning match between two GMs. There is no time for analysis when the clock is set for 1 min. There is only time for pattern recognition. Yet lightning ratings correlate very well with FIDE ratings (they are not the same, but they correlate positively).

Club players often ask why it is so hard to improve. They often spend their money buying books which offer advice. In many cases, these books are written by GMs or very strong players. But just because a player is strong, does not mean that they understand the source of their chess talent.

The reason why improvement is hard is clear. First, playing strength is related to the number of patterns learnt. When a player learns the game, they first learn how to move the pieces and on which squares the pieces move. The number of pattern is much less than 100. They can achieve some success with this number of patterns. But as the player advances the number of patterns needed increases - exponentially. It is sometimes estimated (how, I do not know) that GMs need to know 100,000 patterns. FMs may need 10,000 patterns. Strong club players may get away with knowing only 1000 patterns. Do you see why it is so difficult to improve. To increase your rating by 400-500 points, you probably have to learn at least 10 times as many patterns as you know now.

A second point is that as a player matures, i.e. into the 20s and 30s, the capacity to learn new patterns decreases slightly. The brain is less plastic at later ages. This is a reason why mature players find it difficult to improve. This can only be overcome by increased practice.

A third reason why improvement comes early on, but less later is that there is substantial scope for improvement in the teens and early twenties at the period when the brain is myelinating the frontal cortex. At this stage, the player becomes more cautious and is able to concentrate better.

OK, so I have prattled on about why Kotov is wrong about his famous Tree of Analysis. I do want to add that analysis exercises do help with concentration and visualisation. However, they are not the only answer to improvement. Learning patterns is the key. There are few better ways to do this than by playing through Master games - particularly the games of players like Capablanca, Rubinstein, Botwinnik, Tal... Play them, study them, copy them. At later stages, the student can develop by studying games that are more intimately related to the opening repertoire that he or she adopts.

So is there anything good about the Kotov book? Of course there is. I just wanted to emphasise that this book is NOT the answer to How do you think like a GM? Many GMs would agree with me on this. You can learn from this book, but do not take it as the only or major route of study.

I particularly liked Kotovs anecdote in the book about Capablanca's treatment of an ending. There is much to be learned about endgame play in this chapter. There is also a great deal of good advice interspersed through the book, so it does earn its 3 stars.