Pattern Recognition of Key Positions

I have a long rambling question about pattern recognition of key positions if you are brave enough to read on I would appreciate your sensible thoughts.

I’ve read several times that strong players (strong as in players over 2000 or master strength players) use pattern recognition/intuition to a greater extent than weaker players, and it is perhaps this pattern recognition that helps set them so dramatically apart from such weaker players.

It has also been suggested that because of this such players actually calculate less than weaker players, presumably because a stronger player can instantly read a situation and can often intuitively make good moves. Such, apparently, is one of the reasons why a master player is able to perform so well during simuls against weaker players. Often the master is able to scan the board and quickly make a move without the need to calculate because of his vastly superior pattern recognition (he/she also possess vastly superior calculation skills too of course as well as every other aspect of play, but primarily uses pattern recognition or intuition in being able to quickly make a move, as they can read the board so well.)

If this is correct I also assume that stronger players need to calculate less because they know when calculation isn’t necessary. A strong player may be able to see the board and instantly see the mate in four for example because they recognise that particular motif, but I’m not just speaking in terms of forcing tactical positions, they might also be able to recognise an obvious move in a quiet, strategic position too and make that move without the need for much calculation.

With this in mind (and thanks for keeping up so far if you have!) is it not a really good idea to work on committing to memory key patterns and situations in a systematic way in order to improve this key area of chess? Yes this is what everyone does when they work on tactics problems to some degree, but I’m talking about systematically learning essential positions in an organised way. (The study of tactics for example on Tactics Trainer is completely random and many books on tactics are just mates in 3. It is not systematic and it is not necessarily working with key positions.)

For example I am looking at the book 300 Most Important Positions and Ideas by Lev Alburt. This is a good book and it has got me thinking whether it is so useful to be able to roll the idea of this book out to more positions, say 500 or even a 1000 and review them at optimal time to help commit the positions to long term memory. (I have only had the book a few days and have gone passed position 50 already.) Do you think such an undertaking would be a good idea? Why or why not?

For example I have reviewed some of the early positions in this book more than once, and as anybody would, can now instantly spot the solution, whereas upon first seeing such positions it might have taken me 5, 10 minutes to try to calculate the solution, if indeed I found it at all. 10 minutes calculation time reduced to a few seconds because I just learnt to recognise the pattern.

I can imagine the benefit if I was able to roll that out to encompass all of these key positions (key positions according to Lev) to me this would seem to be of great benefit, what do you think? The key is being able to commit them to long term memory. This can be achieved by studying by repetition at optimal time periods…I can go into that at a later stage.

It sounds like a lot of effort but I am simply talking about going over these positions (or positions like them from elsewhere) several times to help commit them to memory, just a minute or so on second reading and then it should just be a case of seconds after that. Lev in the into says that this can be done by simply flicking through the book at random from time to time, but I was thinking of something a little more structured personally maybe with the use of computer, not thought about how I would go about that completely yet, but anyway, overall do you think this is worthwhile? Do you think such an undertaking would be a good idea? Would it significantly improve playing strength do you think? I mean I'm not talking about anything new or original here just thinking about organising something a little better with what's already there.

Thanks.

Pattern recognition helps reduce the calculation load because it helps you to spot 'calculable positions' but also to trim variations faster when calculating.

Now your idea is a very good one, though the selection of positions may be a little tricky.

I have thought a good deal about it and have developed my own pattern training system, combining various ideas and tools.

"is it not a really good idea to work on committing to memory key patterns and situations in a systematic way in order to improve this key area of chess?"

Yes, but one difficulty is in choosing which positions are key.  IMO it depends a bit on the person and (again IMO) this is one reason why some players progress much more quickly than others.  That is to say what lessons someone is able to pull from a position and then to how many similar positions are they able to generalize this lesson/pattern/idea.

I don't have Lev's book.  I imagine it as being middlegame positions, but "key" or how about "instructive positions that are useful to generalize" can be found in openings and endgames as well.  IMO one of the best ways to discover and remember these key positions is to... play long games and analyse them as well as play over master games and analyse them :p  Not very groundbreaking I'm afraid.  Of course you could make a general one like Lev and it would probably be useful to many people willing to take the time to understand the positions.  I've often thought about putting together a collection on basic ideas behind different pawn moves (because they used to frustrate me as a beginner).  Neat to hear hicetnunc has dome something similar to Lev.

Anyway, a computer program to serve as a storage device / convenient flash cards sounds great.  I've usually just recorded my "key" positions in a note book or by ear marking a page in a book or marking the score sheet of a tournament game.

Actually, there's nothing new under the sun : Soviet players were storing cards with diagrams in binders, and one of my old chess friends cut diagrams in chess magazines with scissors and pasted them in a large notepad for regular review... 

Great, great post Ben (as usual)

my modest, humble view is that there are tooo many patterns, and so many combinations of them, and so every trainner recomendens the exposure to thousands of not only positions, but whole games, to add the requiered flexibility to the pattern recognition...i think just to focus in ,say, 800 positios maybe won't help you to spot a slighlty different position, say if on move 4 something goes awry; but if you have gone to  (again, say) 12000 positions, you will have seen your original 800 learnt-by-heart positions and many more to be able to coordinate alternate solutions.

On the other hand, as you know, there is no trainer or player who recomends to play JUST this 250 master games; quite the contrary people always recomend to play thousands of master games, no matter if you did not read this one or that others Capa's gem...so i think there is something similar with tactics,

finally is you DO 12000 puzzles, you WORK through 12000 positions; if you learnt the by heart that is difficult to prove useful in our game.

once again the paralelism with any practical knowleadge like talking english or riding a bycicle...some static knwoleadge might be useful but that is not the way anybody learns to speak even a foreing language, you learn mostly, by doing not by storing new concepts in your brain...with very little theory you are already ready to learn "just" by doing, and doing, and doing ...and of couse , failing which is part of the process.(someone said: next time , just fail better)

One last coda: if you are interested in this issue the W Hendriks, move first, think later etc book might apeal to you (though some thesis are a bit overstated)

Thanks for the comments. I'm just reading though and thinking about it.

For example generality immediately recognize the smothered mate pattern.

OK, just some thoughts that have come to mind after reading the above, thanks again.

Selection of positions – start with the ones in this book and then from there not sure, personal games, errors, significant master games? (The positions in the book are various, mostly middle games, but also some endgame motifs as well.)

In terms of the reviewing own games/playing over master games, I already do this (though I need to play slow games and do it, I will be joining a club after summer, too much going on in summer.) This is sound practical advice. However, I have looked at many master games and I am still doing that as well. The problem is though committing those lessons, any lessons be it going through a game or looking at positions, to memory. Rather than go through 1000s and remember a few things but forget the vast majority, is it not better to really grasp lessons from key positions? Go through a 1000 and remember 5 or go through 300 and remember them all? Sort of thing.

Yes some flashcard system or computer-based flash card system is what I’m thinking about here. (And if the Russian masters used to do that then it can only be a good thing.) As chess is a visual pursuit (and visual works well with memory) then theoretically you should be able to store positions effectively over other types of material.

Make a general one like Lev… Yes he actually talks about personalising a collection in the intro. He says that you should begin to archive positions that mean the most to you and that the positions in this book (his book) provide you with a foundation upon which to build and expand your own. So maybe that answers the above point about significant master games and your own games? He says: it’s extremely effective for you to archive positions from your own games. Record positions in which you faced problems or made errors.

Too many patterns and positions?

I don’t know you may be correct. However I think you can have a foundation of key positions that you can apply in various situations, as Clint says the threat of smothered mate is instantly recognisable for most people. Lev writes:

Here’s a promise: To be a strong player, you do not need to know hundreds of King and Pawn endgame positions – but only 12 key positions. Of course they have to be the right positions – and they’re in this book! To be a master you do not need to know thousands of King and Pawn endings. You need to know about 50 key positions.

The example he gives is this:

Which, because I’ve seen and followed the solution (as you probably have because you have the book) I’m able to solve instantly, where the first time I might have spent 5/10 minutes trying to work it out/remember the stuff about opposition which I always seem to forget! Stronger players will see that position and instantly know the correct move, weaker players with have to analyse/calculate, waste time and maybe still get it wrong.

Thanks, I’ll have a look at that book too.

For me it’s all about retaining the information. As a student and reading for a hobby I've read hundreds, if not thousands of books (English literature). For the majority of these books I can now barely recall the main characters/plot probably. In a few years I probably won't remember having read the book at all (like when you are watching a film and 20 minutes in you realise you have seen it before!) Well I know I will have read it but you know what I mean...

In terms of chess I’ve gone through 100s of master games but how much has that benefitted me if I haven’t stored/learned important concepts in my long term memory? Rather than attempt the impossible by going through thousands and thousands of games/positions and perhaps forget 99% of it anyway, maybe it is better to remember key positions and remember them all as I suggested earlier?

Roi, I really like that idea thanks.

@Anton, I don't think the 'too many patterns' is a real problem. You just try to memorize important patterns, not all patterns. If you know well 1000-2000 important patterns (maybe 70% would be tactical patterns), I think you already know enough to be a strong player.

As for the 'most important games', there is no absolute list of course, but there are good game collections, and if you work on those, you're probably doing fine. Now, the 12 000 games thing is for people who want to become masters (and have a good memory ! )

Also trainers should be able to provide references to thematic games, which may help improve awareness towards a specific point.

But beyond the 'usual patterns', Alburt makes a very good point : the most important new patterns are those that derive from your own games - some plan you didn't find, a new strategic idea, some trade that wasn't correctly evaluated, a practical decision, a piece which "belongs" on a given square, things like that. You should build your own data bank... 

For pattern-diggers, let me shamelessly advertise one of my recent blog posts, listing many sources with good quality material.

Thanks for the link and book suggestion. Great blog collection.

Just buy yourself some decent chess books and work through their many, many, examples.  Johan Hellsten, Mastering Chess Strategy (2010), comes to mind.  John Nunn's "intermediate level" books on the Middlegame and Endgame are also great training tools.

Or maybe even hire @Hicetnunc as your chess coach, even though he has blocked me from his notes.  Whatever. 

Either way, you will became a much stronger player, and (more importantly) save yourself a massive number of keystrokes in this somewhat silly thread.

Well the idea is to save myself time by learning key positions, no by mastering key positions, as opposed to coming across 1000s of random positions/games that I will never remember. Or to read through a book, enjoy it, then forget it. That's the point. It's about retaining key information and not collecting vast amounts of surplus material.

I have found 100 of the 300 positions in PGN file, I've added them into a database in Fritz, though I think I prefer the simple Winboard over Fritz as I've never like it. I could do with a better chess database/program though really.

https://secure.chess.com/forum/view/general/chess-advice-most-chess-players-dont-like-to-hear?page=2

read the post by dpruess in this thread...particularly the part about how to use tactics trainer for pattern recognition.

Thanks, yes that would seem to fit and is similar in principle to the advice that I follow from Dan Heisman in Tactics Trainer. He suggests setting it to a lower level in order to learn basic patterns/motifs over and over. I personally set it to 1200-1300, 300 or so points below my rating. The objective in doing this is not to 'solve' problems but to embed common patterns/tactics so that you can spot them quickly. I've been doing thise for several months and feel I am getting a lot from it. I do it daily from anywhere from 5 minutes to 45 minutes maybe.

This is a similar thing that I want to roll out with these key positions, only instead of getting random positions on the same theme, with TT, I hope to work first with these 300 positions and then to start adding ones of my own.

Can anyone suggest a simple flash card style database for this? One that could perhaps store both PGN files and the FEN strings. I want to be able to quickly click over the positions for review. I would ideally like to be able to organise them by theme too, check mate positions, endgames etc. I've not very good with some of the chess programs I find some of them (like Fritz) fiddly to use. I'm looking for something clear and easy to use, something like winboard but with maybe a bit more to it, like the ability to organise by theme etc.

Thanks again, comments greatly appreciated.

Both TT and Chess Mentor on here have your scores broken down by themes.  It will say what your percentage is for underpromotion or perpetual checks for example.  I've always wanted them to let you pick the theme you had problems with and let them just feed you those problems.  For all I know you can do that and I just don't how.  But if you can't, it would be a good option to have.

Yes I've thought about option it would be useful. However, as I set the TT to unrated, in order to lock in a lower rating, you don't get any feedback on your scores nor even the number of problems you have attempted or got right. If you look at my TT attempts over the past three months for example it looks as if I haven't hardly done any (only when I rarely switch on TT) but this has not been the case.

There is a body of knowledge that you must imbibe and learn cold, in order to reach a reasonable playing strength, say USCF 1800.

TT training and massive computer resources won't cut it.  Sorry.

The problem is "in your head" (you need more chess knowledge) not in the speed or efficiency of your chess hardware and software.

The more knowlege you acquire, the faster you will be able to play reasonably well.  Put a 5 or 10 second bonus into all your game speeds, and learn to play "your game" at roughly Game in 30/5 speeds or faster.

Focused study, (perhaps) coupled with a coach, and judicious but limited practice at quick chess speeds is the recipe for becoming a stronger player.  Couple those training tools with OTB play at regular time controls, plus other complimentary study and analysis.  Very Simple.

Extra gigabites and faster processors, and grand databases can be great.  But only if you know what's important to begin with.

If you really want to waste your time in these threads, just read all the blather written about various Openings.  Massive numbers of mindless keystokes, going round and round in circles.

Here's a great example of a mindless thread (going on for 5 years!) concerning a specific opening --

http://www.chess.com/opening/eco/B22_Sicilian_Defense_Alapin_Variation

Whatever. Knock yourself out.  You've been warned.

Best book to reach USCF B class, roughly 1600-1800 playing strength --

http://www.amazon.com/Chess-Tactics-Batsford-Book/dp/0713489340/ref=sr_1_1?s=books&ie=UTF8&qid=1369781002&sr=1-1&keywords=paul+littlewood+chess

Small book, only 150 pages.  Do it. 

Thanks for the book tip it does sound like a good tactics book.

I also agree with you about a number of things. Chess knowledge. Yes that is exactly what I am talking about trying to gain!! Chess knowledge. I'm not interested in powerful chess databases with go faster gigabytes or in spending a year learning the Alapin, however neither do any of my postings suggest that!!

What I'm simply talking about, which everyone else seem to get, was in learning key positions, actually remembering them learning, as opposed to blindly going through 1000s of tactics/master games/random positions in the hopes that some of them stick! Remembering key positions.

You are quite correct about standard OTB play and I have to admit that I am letting myself down here by not joining the local club and getting down to it. This I must do shortly if I am serious about wanting to improve for sure. Also in terms of getting a coach, though I have had several session with one.

I know that there is a lot of junk posting on the forums, that may be true, and it is not as if I spent all my time on them. However the forum is also a way to touch base with other chess players many of which have sensible advice to pass on and knowledge for which I am grateful.

In terms of looking for a chess database I am after something really simple that I can flick through key positions with. Something like flashcards online. In this day and age you would have thought that there would be an efficient program for this, but I have spent the last couple of hours looking and searching online, downloading a few and found nothing satisfactory. I would print them out on card but surely in this day there is something better than that?