PDA

View Full Version : Improving calculation



paulb
21-01-2008, 04:51 PM
Forgive me if this has been raised before, but I'd dearly like to know if anyone knows of any good books/techniques/training programs for improving one's ability to calculate/visualise. I'm also interested to know of anyone who has used such techniques and whether they have noticed any improvement.

In particular, what does one do about the following: after calculating more or less accurately for a certain number of moves, one gets to a point where the resulting position is vague in one's head, and one loses the confidence or ability to proceed. How do you clarify the mental picture (assuming, contra Wittgenstein, that it is ultimately a matter of visual imagination) and extend the range of calculation?

Thanks in advance for any help

Cheers - PaulB

Adamski
21-01-2008, 05:06 PM
See Jermey Silman - How To Reassess Your Chess.

More detailed level (and need to be strong player to visualise as much as he does ) A Kotov - Think like a Grand master.

Both mentioned in other threads - see the search function.

Rincewind
21-01-2008, 05:32 PM
Hi Paul,

I think doing tactical puzzles, lots of tactical puzzles is a good way to improve. The more used your mind gets to such calculations, the better it is a coping with longer combinations and more complicated positions with less mistakes. Books with lots of such positions like 1001 brilliant checkmates and the like by Rhinefeld are useful for this purpose. Perhaps I found a database version of this book and used that with chessbase/scid using their training mode and random position selection.

The Polgar book on chess I believe also has a lot of material for such a program though I haven't got this one. (I think Theo Rippas suggested this book to me a few years ago). Obviously I haven't tried this particular book first hand.

Another thing which may help is to try to play blindfold. This might be daunting at first but I think can only help with a little practice. I've never tried this one myself but I believe Alex MDC has. Well he has a blind account on FICS anyway. If you're interested I noticed someone was starting a blindfold club in another thread on the board.

If you run into Alex or Theo you might like to ask them about their ideas as I might not be relaying them accurately.

CameronD
21-01-2008, 07:13 PM
Forgive me if this has been raised before, but I'd dearly like to know if anyone knows of any good books/techniques/training programs for improving one's ability to calculate/visualise. I'm also interested to know of anyone who has used such techniques and whether they have noticed any improvement.

In particular, what does one do about the following: after calculating more or less accurately for a certain number of moves, one gets to a point where the resulting position is vague in one's head, and one loses the confidence or ability to proceed. How do you clarify the mental picture (assuming, contra Wittgenstein, that it is ultimately a matter of visual imagination) and extend the range of calculation?

Thanks in advance for any help

Cheers - PaulB

I'm currently working through silmans book, and it really helps a lot and changes the way you organize chess. He recommends creating a plan based on weaknesses before calculating a single move, you calculate a dream position and try to calculate it.

Basil
21-01-2008, 11:28 PM
I'd dearly like to know if anyone knows of any good books/techniques/training programs for improving one's ability to calculate/visualise.
Hi Paul

Andrew Soltis' book 'The Inner Game of Chess' (McKay Chess Library), subtitled 'How To Calculate & Win' is just what you're after. Weighty enough at 350 odd pages. Discusses in detail analysis trees, how to analyse, visualise, what to do in over-load etc..

As with all books, there will no doubt be detractors, but I for one heartily recommend this as an excellent place to start! You'll probably get more out of it than me :eek:

Cheers
Howard

Miguel
22-01-2008, 09:16 AM
Forgive me if this has been raised before, but I'd dearly like to know if anyone knows of any good books/techniques/training programs for improving one's ability to calculate/visualise. I'm also interested to know of anyone who has used such techniques and whether they have noticed any improvement.
This may seem rather banal, but I don't think there are any special techniques/training programs for improving calculating ability. I think Rincewind is spot on: Get a book on tactical puzzles and work through it. Have you tried Fritz's "Sparring" mode?

Aaron Guthrie
22-01-2008, 11:22 AM
Kotov's think like a grandmaster for calculation.

For visualization, sitting at a board with difficult puzzles (ones that you may or may not solve in 30 minutes) set up, and not giving up on them until it really is useless. Specifically something like the puzzles at the back of informator, or in the informator puzzle book. This suggestion requires that you are reasonably competent in simpler puzzles, however.

Capablanca-Fan
22-01-2008, 12:16 PM
Not sure that many people on ChessChat are in a position to advise someone of PaulB's level.


I think doing tactical puzzles, lots of tactical puzzles is a good way to improve. The more used your mind gets to such calculations, the better it is a coping with longer combinations and more complicated positions with less mistakes. Books with lots of such positions like 1001 brilliant checkmates and the like by Rhinefeld are useful for this purpose. Perhaps I found a database version of this book and used that with chessbase/scid using their training mode and random position selection.
I agree with lots of tactical studies. Reinfeld is often laughed at but Purdy called his books "sound and reliable".


I think doing tactical puzzles, lots of tactical puzzles is a Another thing which may help is to try to play blindfold.
Not sure. I play blindfold simuls much better than, for example, Larry Christiansen, who lost track of a lot of his positions (http://www.youthchess.net/christiansen.htm), but he can calculate much better than I can. Blindfold is not only about visualization but recall.

Not sure about Kotov's tree of analysis. Nunn's book Secrets of Practical Chess shows up some of the problems with "candidate moves'' if one doesn't retrace branches already analyses. E.g. a line in one candidate move may not work, but another line shows up a possible clue to make the first one work. some of his helpful hints are:


DAUT — don't analyse unnecessary tactics. Especially cogent from an author who's won fierce tactical slugfests with some of the world's best.
Learn about safety nets, i.e. a way to bail out with say perpetual check if a line turns out not to be as good as you thought. Even in the famous Kasparov's Immortal against Topalov (http://www.chessgames.com/perl/chessgame?gid=1011478), K knew that he was safe to enter the line because he had a perpetual a few moves down. Once he reached that point, he could look further and realise he had a win.

MichaelBaron
22-01-2008, 12:42 PM
For someone as strong as Paul, there are not many tactical books.

One book that comes to mind (and its quite advanced as i remember working on it after reaching candidate master level)

is "Together with Grandmasters" by Hort and Jansa. I may have the book title wrong as i read it in Russian and it may havea different title in English but the author's name should be enough for finding it.

Also, solving puzzles at the back of the "Chess Informators" is good..they are also quite advanced.

Another good way to practice is playing sharp tactical lines against Fritz, Rybka etc.

paulb
22-01-2008, 02:55 PM
Thanks for all the advice - I'll certainly follow it (some of it I've already been following!).

But a few more questions:

1. Presumably, everyone eventually will have trouble visualising a position after a long analysis. (For me it might after 5 moves, for Kasparov it might be 20 moves, but *eventually* it's an issue for everyone). So what do you do then?

A chess player I know said he read somewhere that masters playing blindfold will divide the board in four in their imagination. They are able to visualise 16 squares quite OK, but not 64 (that's certainly my situation!). They later put the four quarter-boards together in their minds, presumably by noting which pieces are about to take each other and so on - the relationships between the pieces.

First of all - does anyone know whether this is true of masters playing blindfold. And is it effective?

Either way, it's an example of a mental "technique" to improve calculating abilities.

Does anyone know of any other mental "techniques" that might extend the range of calculation (perhaps including mental techniques from non-chess fields). For example: hypnosis/meditation etc?

Similarly, some people use mnemonics to remember things (eg All Cows Eat Grass - ACEG - for the musical notes on the bass stave) - are there any comparable techniques for remembering the position of pieces in a chess game? Just a thought.


One other related observation: some very strong players are known to stare at the roof, or into the middle distance, rather than at the board when they're playing. I saw Greg Hjorth doing this in my youth, shortly before he uncorked some fantastic move. I've also heard that Ivanchuk does it, and some suggestion that autistic players do it. I'd be interested to know why they do this (is it perhaps because, once one's calculations go a few moves past the board position, the positions are so different that the board is no longer a useful reference point? Is it something that mere mortals like myself should try - is there a point at which seeing the physical board is no longer useful for calculation?

Gattaca
22-01-2008, 03:48 PM
Hi Paul.

I can play pretty reasonable blindfold chess, without forgetting the position at least.

I don't use any artificial technique like dividing the board into 4. Neither do I visualise the whole board perfectly like a picture in my mind. What I see is a whole bunch of relationships between the pieces, as you alluded to. So in any given blindfold game I will concentrate on a particular area of the board and then swing my attention back and forth, perhaps a bit like looking through a funnel. (Is that tunnel vision? :) )

I suspect that is also what we do when looking at a board in a normal game, but because our eyes are 'seeing' everything it appears different to a blindfold game. In actual fact though, we are not really thinking about everything our eyes are seeing, our brains are only thinking about a certain relevant part of it.

I doubt that when players stare into space they are doing much serious calculation. It's probably either resting their minds and relaxing for a moment, or just double checking that visual stimulus isn't interfering with what they are visualising. They usually only do it for a short time.

As to how to extend ones depth of accurate calculation, that is a tough one. Practice is certainly important. It's hard work calculating, so a lot of players are lazy and decide to mainly use their intuition, or concentrate on positional play.

If you simply decide to start calculating as far as you can for as long as it takes in your games, you'll certainly improve your ability in that area. It may not pay off straight away though, both because of clock time used and because human calculation is never going to be good enough on its own. Intuition about the positions you are 'seeing' and good assessments when you reach the end of your analysis are still very important.

For example, the key moment in a line might be 8 moves down the variation. It may not matter whether you can see 6 moves ahead or 7 moves in such a case... what is more important is whether you assess the probability of it being good correctly. Of course if you could calculate an extra move ahead then suddenly it would have made a big difference in such a case, but I'm just making the point that sometimes extra calculating depth won't immediately convert to extra points. If however, you add a whole few extra ply to your calculations, and more importantly become more accurate, that should obviously make a huge difference to results.

How to do it? I'm skeptical about any bizarre methods like hypnotism or zen. (I tried the latter unsuccessfully once) Practice and discipline at the board, definitely. Doing studies and problems without moving the pieces around, yes. Books like "Think like a Grandmaster", maybe. Playing blindfold a lot, probably. They're my initial thoughts on the matter for what it's worth.

Aaron Guthrie
22-01-2008, 07:08 PM
1. Presumably, everyone eventually will have trouble visualising a position after a long analysis. (For me it might after 5 moves, for Kasparov it might be 20 moves, but *eventually* it's an issue for everyone). So what do you do then?Then you give up and use the information you did get out if it to help you choose your move. Is there any other thing one can do?
One other related observation: some very strong players are known to stare at the roof, or into the middle distance, rather than at the board when they're playing.I do this to an extent (so some not very strong players do it too;)). I guess I can say two things about it 1)sometimes one just wants to think about the position in this manner (and my natural reaction is why does one need to look at the board?) 2)sometimes the pieces on the board just get in the way (especially if one has calculated a few moves ahead already).

paulb
22-01-2008, 10:04 PM
Then you give up and use the information you did get out if it to help you choose your move..

That's fine, unless you've reached a really sharp position where further analysis is really needed.

Also, in general there's a desire to extend the range of one's calculation, because one tends to get beaten by players who can calculate further ahead.

Aaron Guthrie
22-01-2008, 10:07 PM
That's fine, unless you've reached a really sharp position where further analysis is really needed.

Also, in general there's a desire to extend the range of one's calculation, because one tends to get beaten by players who can calculate further ahead.I thought you meant what do you do at the time.

paulb
23-01-2008, 02:36 AM
I thought you meant what do you do at the time.
I do, in this sense: is there anything I can do to clear the fog out of my brain. I'm interested in any techniques and tricks that can be applied at the board during a game, as well as any practice techniques for between games. Obviously I agree that at some point you have to make a move :)

paulb
23-01-2008, 04:24 PM
Thanks to everybody - and particularly Guy West and Jono - for their excellent advice.

Guy: It was very helpful to me to learn that you don't "visualise the whole board perfectly like a picture in my mind" and have to scan various parts of the board. The take-home message for me is not to abandon analysis once the position becomes foggy.

And your point that analysis simply is hard was also quite illuminating, suprisingly enough. In the past I tended to think "Gee, it's hard keeping track of this position as I analyse. There must be something wrong with me' ... whereas your remarks suggest otherwise. Hopefully this will spur me to keep going when the going gets tough.

Jono: thanks for the points about DAUT and safety nets, and the distinction between memory and analysis in blindfold (though, to some extent, failure to visualise correctly is often a case of failing to remember where a piece is, but cases of confusing board 4 for board 6 in a blindfold simul are clearly another matter :) )

xxx

But consider this: there is an army of intelligent professional and would-be professional players who spend many hours practising and doing precisely what you suggest. Yet some end up with vastly superior calculating abilities. (People like Pillsbury, with his incredible feats of blindfold chess, spring to mind.) Is this simply because their brains work better (how and why?), or are they thinking differently? How would we know? And if it is the latter, what exactly are they doing? These seem to me to be interesting and open questions. (With this select group of people, it does not seem to be a case of better calculation due to harder work - they're all working flat-chat).

Aaron Guthrie
23-01-2008, 04:40 PM
One thing I learnt about calculation that seems quite obvious now is that one need not calculate until the position has calmed down (i.e. tactics still abound). The issue is that I often get very complicated positions that are tactically rich, but dynamically equal, and one can just evaluate at a certain point that the position is dynamically equal, rather than calculate to the limit of how far one can. I only realized this through experience. So when you say

I do, in this sense: is there anything I can do to clear the fog out of my brain.clearing out the fog in my case is more a case of making a practical judgement about which move I want to play at some point.

I don't know if there is any way to make the position apear clear if it has gone foggy, aside from resting ones mind for a moment, and trying really hard again. It would seem to me you have to get the ability before the game, and that is unlikely any trick would help much (but if anyone knows one, share, share!).

MichaelBaron
24-01-2008, 02:29 AM
Getting Back to one of Paul's original concern: even if you can calculate 10 moves deeps it is impossible to analyze all lines till mate.

This is where positional chess and tactical/calculation abilities go hand in hand. Tactics is essential to analyse the sharp possibilities available but even if you can see 10 moves deep you still need a good positional ability to assess the position that is going to appear on the board once these 10 moves are going to be played out.

DarkHorse
28-01-2008, 01:18 PM
Forgive me if this has been raised before, but I'd dearly like to know if anyone knows of any good books/techniques/training programs for improving one's ability to calculate/visualise. I'm also interested to know of anyone who has used such techniques and whether they have noticed any improvement.

In particular, what does one do about the following: after calculating more or less accurately for a certain number of moves, one gets to a point where the resulting position is vague in one's head, and one loses the confidence or ability to proceed. How do you clarify the mental picture (assuming, contra Wittgenstein, that it is ultimately a matter of visual imagination) and extend the range of calculation?

Thanks in advance for any help

Cheers - PaulB

Very interesting post,I have been trying to improve my calculating ability recently too. One thing I did was get hold of Convekta's software CT-ART 3.0 (Basically a fancy name for "The art of chess tactics").
This is awesome software and allows you to improve calculation by constantly analysing complex tactical positions and finding solutions, you can do tests on it and be graded..and so compare progress over a period of time...
If one cant improve calculation greatly with this software I'd be very suprised.

For more information check this >>

http://www.chesscentral.com/software/ct-arts-chess-tactics.htm

Davidflude
28-01-2008, 05:26 PM
The toughest that I have seen are in Kaissiber.

Silman's books are magic but you also need tactical exercises and you need to practice playing through tactical problems.

The other thing that you need to develop is intuition. First to develop a feeling for when there is something special about the position. Second to know when to play a move without trying to calculate too far.

Denis_Jessop
28-01-2008, 08:13 PM
Getting Back to one of Paul's original concern: even if you can calculate 10 moves deeps it is impossible to analyze all lines till mate.

This is where positional chess and tactical/calculation abilities go hand in hand. Tactics is essential to analyse the sharp possibilities available but even if you can see 10 moves deep you still need a good positional ability to assess the position that is going to appear on the board once these 10 moves are going to be played out.

I would go even further by saying that even if you can calculate 10 moves deep it is generally impossible and impracticable in the time available in a tournament game to analyze all lines. There is a question that must be faced, namely, how much you need to analyze or whether you need to analyze at all. John Nunn deals with this very clearly in his "Secrets of Practical Chess". It is the best book I know of at doing what it says it does - giving practical advice on how to play a tournament game - especially the first third of the book called "At the Board" in relation to Paul's question.

DJ

bgriffenchop
06-02-2008, 04:08 PM
Don't do problems on a 2-D board or on a computer. Set-up all problems/positions on a table with a bright light.

Oh...and solve 200 rubix cubes a day.

That is all.

MichaelBaron
18-08-2008, 01:38 PM
Have a look at the game Werle - Wells! The last few moves are not terribly complicated to find but are really cute. Good game to analyze for tactical skill practice

http://www.howardstaunton.com/hsmt2008/rd7.shtml

EcoChess
21-10-2008, 03:44 AM
You must resolve puzzles all days (http://www.ecochess.com/tactical/tactical.htm
But, on the contrary, reasoning is fundamental in all positions (or the most positions).
TECHNIQUE:
Level 1:
1. Put on board initial position.
2. Move (with your hand) 1 move: For example: 1.e4-Nc6
3. Now, put again initial position.
4. Move (in your mind) 1.e4-Nc6
5. Repeat this about 20 times (IE: 1.d4-c5; 1.f4-e5, and so on).

Level 2: Similar to level 1, but you must calculate 2 moves, and so on (Level 3, 4,).

Just do it! Bye

Miranda
21-10-2008, 12:55 PM
TECHNIQUE:
Level 1:
1. Put on board initial position.
2. Move (with your hand) 1 move: For example: 1.e4-Nc6
3. Now, put again initial position.
4. Move (in your mind) 1.e4-Nc6
5. Repeat this about 20 times (IE: 1.d4-c5; 1.f4-e5, and so on).

Level 2: Similar to level 1, but you must calculate 2 moves, and so on (Level 3, 4,).


Ouch.
That sounds like it takes ages!!