Tuesday, July 11, 2017

Initiative redux

All scenarios that win a tempo have two targets involved.
Legend:
• T = Target
• D = Defender
• A = Attacker
• C = Capturing attacker
• PoP = Point of pressure
• Green = enemy
• Yellow = winning us
• Arrow = attack or defense
• Lightning = gained tempo

The initiative group has two families, the captures and the double attacks. There are four scenarios of defense:
• Recapture with tempo (additional threat)
• Save target with tempo (desperado)
• Recapture with defending the other target
• Save one target and defend the other target in one move

1. I am not sure if it is helpful for you, but anyway I want to share it with you. You can use it to explain your theory/system at real chess example.

[4b2r/Rrp2k2/4pqp1/1pPp1p2/1P1P4/4PPQ1/4BKP1/R7 b - - 0 34]

And now Black played: 34...Bc6? and white gained a decisive advantage:
35. Rxb7 Bxb7 36. Qxc7+ Qe7 37. Qxe7+ Kxe7 38. Ra7 Rb8 39. c6 wins a piece.

In a few moves played we can say such motifs like: double attack, pin, destroying the defender (guard), etc.

Good luck in your future discoveries and findings! To be honest I cannot understand your ideas without having chess examples. Anyway it does not mean they are bad or not valuable! I keep the fingers crossed on you!

2. I hear what you say. The diagrams have more meaning for me than for you, because for me there are all sorts of memories and emotions connected to the diagrams. For you it seems theoretical, for me it is highly practical. Since every diagram is related to a series of failures at Chess Tempo, and they are designed to address these failures.

The most important part of the current tree of scenarios for me, is probably the initiative. I have been blind for the subtleties of the battle for tempos. With the use of the diagrams in this post, I'm able to think more logically about the initiative. I haven't found all there is to know about the battle of the initiative yet, nor have I internalized it completely. Yet I'm making progress in replacing trial and error with instant scenarios.

I'm gathering example positions for every branch of the tree of scenarios. Just as one would gather suitable pictures of birds for the composition of a field guide for bird determination. That is a time consuming process.

About 1/3 of my failures at Chess Tempo are initiative related. 1/3 is about not finding the invasion and or the mate. The last 1/3 is related to the shenanigans around promotion.

The seven scenarios around the box should give a clue about mate. I haven't tested if the diagram of the box with the seven scenarios is actual helpful. I'm working on that. I'm gathering a series of failed mate and invasion problems.

I haven't written anything about scenarios that are related to the promotion until now. Right now, I'm gathering relevant positions which I can use for the extraction of promotion related scenarios.

As you see, there is a lot going on. But it is difficult to write a coherent and readable post about it at this stage.

3. I think the addition of colors and symbols help clarify the intent of the diagrams.

I'm going back through Momir Radovic's blog posts and comparing to Averbakh's concepts regarding the Theory of Contacts, looking for additional insight. Radovic left the promotion (queening) contact out of his taxonomy (Attacking, Blocking, Restricting [Controlling], Protecting [Defending]). Averbakh includes Promotion (queening threat) and Stopping (the queening threat) as the final two categories of the six elementary contacts.

I share Tomasz' desire for examples directly related to chess diagrams. AS you have noted, the abstract diagrams work best for you but not necessarily for others. I understand the concepts embodied in your diagrams, but because I don't have the memories and emotions to give them meaning, it is hard to use them directly as a "coat rack" to hang my own mnemonics on.

I've got a couple of ideas in work, but it may take some time to get them to the point that I want to share them. (I'm sure you can understand!) It involves using GIF files with chess diagrams that change, showing the same ideas you have used in the more abstract diagrams with multiple scenarios per diagram. I'd love to do some graphics programming to create a dynamic tool that would extract the appropriate information for tutorial purposes, but that doesn't seem likely, given my lack of time to learn an appropriate language and then do the programming itself. In another lifetime, perhaps. . .

4. I should post some problems from Andrei Volokitin's "Perfect Your Chess". Those are the hardest ones I've ever seen, and I don't believe the lines are as short and tidy as Chesstempo.

Right now, I've been going through CT-Art 6.0 here and there, mainly for the pattern recognition, and I believe it has helped, though only going through the easy section so far, haven't completed it yet. The stronger players do spot the hanging pieces more quickly than the lower-rated players, I've found that consistently to be the case.

5. Okay, the first problem from Perfect Your Chess. Find Black's 31st move, it's a drawing combination. http://www.chessgames.com/perl/chessgame?gid=1164838 It's a good example of initiative. :-)

6. That was a fantastic visualization of various line of attack! Thanks!

Here's another example of maintaining the initiative using the assault motif:

5rk1/pppqb2r/2np3p/7Q/3P2Np/2P5/PP1N2P1/R4RK1 w - - 0 1

This is position 781 from Dave Couture's book Progressive Tactics - 1002 Progressively Challenging Chess Tactics. The positions are from his own games.

I thought about the position a lot, and could clearly "see" (in my mind's eye) the intermediate positions but had problems evaluating the merit of various alternatives as the attack progressed. I put it into Fritz 11 (Stockfish), and then found different suggested alternatives, depending on which move I plugged in at each stage. It's disconcerting to find Stockfish re-evaluating the move order once it is given the position part of the way through the combination, especially since I set the search depth to 23 ply to start with, and let it run!

7. PART I

In the interest of continuing the discussion, I'm going to circle back to an earlier topic involving chess improvement, based on re-reading GM Jonathan Rowson's Chess for Zebras.

GM Rowson introduces the topic of chess improvement in Part 1: Improving Our Capacity to Improve.

"How can I improve? . . . Improvement begins at the edge of your comfort zone."

I previously referenced the excerpt from GM Nigel Davies' article on Chesscafe.com titled The How and the What in Chapter 1, What to Do When You Think There is a Hole in Your Bucket, pg. 25:

'The HOW [SKILL] is more important than the WHAT [KNOWLEDGE].' . . .It really doesn't matter WHAT you study, the important thing is to use this as a training ground for THINKING rather than trying to assimilate a mind-numbing amount of information. . . . The reality is that you've got to move the pieces around the board and PLAY with the position[independently explore concrete IDEAS in the specific position] . Who does that? AMATEURS DON’T, GMs do...

GM Rowson continues:

“. . . If you want to get better at chess you need to place much less emphasis on ‘study’ whereby you increase your KNOWLEDGE of positions, and place more emphasis on ‘training’ [SKILL], whereby you try to solve problems, play practice games, or perhaps try to beat a strong computer program from an advantageous position.”

“In any case, I believe this distinction between knowledge and skill gets at the heart of the matter concerning why younger players tend to find it RELATIVELY easy to improve. . . . Paradoxically, the problem seems to be while junior players tend to put what they learn into practice WITHOUT ANY REAL CONSCIOUS INTENT, and thereby improve steadily, ADULT PLAYERS STRAIN IN AN EFFORT TO UNDERSTAND WHAT THEY ARE LEARNING, and this leads to all sorts of problems because rather than gaining in tactile skill, this skill is adulterated by our attempts to formalize it into knowledge. It seems that the gap between knowledge and skill is somehow smaller for juniors than it is for adults. But why is that? This is harder to explain, but it probably has something to do with younger players having greater neural plasticity and fewer prejudices getting in the way. In any case, once you grasp the significance of the distinction between KNOWLEDGE and SKILL, the ground is clear to ask WHY adult improvement is so elusive.”

“ To my mind, the problem can be distilled into two main parts:”

“1) MOST PLAYERS SEEK TO INCREASE THEIR KNOWLEDGE BY LEARNING NEW POSITIONS, and tend to study by “reading and nodding” as Nigel Davies put it. WHAT THEY SHOULD BE DOING MORE OFTEN IS HONING THEIR SKILLS, however meager, by forcing themselves to think ["think" has multiple shades of meaning; System 1 or System 2?] through training and practice.”

“KNOWLEDGE often gets in the way of SKILL, because it is not ‘innocent’—and has to be CONSTRUCTED. This means that there will be limits to what you can learn by passive absorption and that you are more likely to make progress by unlearning some of your existing ideas through the honest and rigorous analysis of your own games.”

8. PART II

Pg. 26:

“Consider that at any given moment, our five senses are taking in more than 11,000,000 pieces of information. . . . Our eyes alone receive and send over 10,000,000 signals to our brains each second. . . .The most liberal estimate [of our CONSCIOUS CAPACITY] is that people can process CONSCIOUSLY about 40 pieces of information a second. Think about it: we take in 11,000,000 pieces of information a second but can process only 40 of them CONSCIOUSLY. What happens to the other 10,999,960? . . .”

“. . . one of the reasons it is so difficult to improve at chess is that it is impossible to know precisely what you are thinking during a game. You can have a rough idea, based on memory, but even during the game you can’t really know what you are thinking, for two main reasons: First, when you stop to think about HOW you are thinking you are no longer thinking in the same way. Second, we have introspective access to the CONSCIOUS PRODUCTS of thought, which we have to construct, but we have absolutely NO ACCESS TO THE PROCESSES OF THOUGHT, and it is only when these PROCESSES start to improve that we become better players.”

“Improvement comes about through improved thought processes.”

Pg. 46:

[I think the following statement has important implications for training to “see” from the vulture’s eye viewpoint. It gives a clue as to how to go about “looking” at a position WITHOUT PRECONCEPTIONS.]

. . . images observed long enough will cease to be random or disconnected and will [SELF-]organize into symbolic dramas, narratives, or PROBLEM-SOLVING PROCESSES.

“. . . think and try to decide what to do. . . . the main thing is to get into the habit of sitting down, selecting a problem, setting it up, setting the clock [for about 20 minutes], THINKING, stopping, comparing your analysis with the source analysis … and keep doing this. It will be hard, and you will not really be “learning” anything through this. But, as I’ve said before, CHESS IS ABOUT SKILL—WHAT YOU NEED IS N-O-T ‘KNOW-THAT’ BUT ‘KNOW-HOW’.

9. PART III

Well, that’s certainly a lot of “word salad”! What the heck is the point of all this verbiage?

We all have taken in a considerable amount of KNOWLEDGE in the form of solving tactical problems exhaustively in large quantities (perhaps 50,000 or 100,000 or more?). We have (to a large extent) been UNSUCCESSFUL in improving our SKILL level (perhaps measured by rating).

Are those positions lurking (unrecoverable) somewhere in our unconscious? I speculate that the answer is definitely YES! As Tempo and others have pointed out, the problem is NOT getting the information IN TO memory; it is getting the information OUT OF memory. But we have NO CONSCIOUS CONTROL over the retrieval process from our subconscious System 1 . It either recalls the relevant information (cues, patterns, whatever) or it doesn’t. And if it doesn’t, there is nothing we can do CONSCIOUSLY to force the retrieval.

As Tempo began his interlude, working on the Tree of Scenarios, I went to work trying to CONSCIOUSLY solve tactical puzzles using the PoPLoAFun approach. I tried to CONSCIOUSLY and conscientiously think using the elements to bring into mind what I should be looking for as a solution. Perhaps not surprisingly, if I didn’t force my CONSCIOUS thinking to consider ALL possible elements, I still missed “obvious” clues as to what should be done. While re-reading GM Rowson’s book, something about this passage (repeated for emphasis) kind of “jumped off the page” at me:

. . . images observed long enough will cease to be random or disconnected and will [SELF-]organize into symbolic dramas, narratives, or PROBLEM-SOLVING PROCESSES.

Perhaps I was wasting CONSCIOUS effort! Perhaps if I just looked at the image of the position, without trying to force any process CONSCIOUSLY, the “image(s)” I needed to solve the problem would get thrown up into the conscious by a hidden System 1 process. So, I stopped trying to follow a CONSCIOUS thought process and just looked at the position, without trying to do anything particular with it. In a sense, it was almost like meditation. Think of it as BEING rather than DOING.

I just started this approach a few days ago, and (so far) it has been simply amazing! Rather than getting stuck trying to force a solution (and often CONSCIOUSLY picking the wrong initial “clues” and then getting stuck in a cul-de-sac trying to “make it work”), I just patiently wait for the “clues” to bubble up from System 1. It only take a short amount of time, and then System 1 WILL begin to make sense of the position based on what has been stored in memory. Instead of rushing into CONSCIOUS thought at the first sign of an idea, I wait for the position itself to “tell its story.” I only begin to “calculate” after waiting for System 1 to sort out and present the significant parts of the position. The best part is that (usually) it eliminates from consideration those things that are NOT relevant.

It’s not about following a well-thought out formal process, which we already know does not give us access to the hidden System 1 “knowledge.” It’s about trusting that the marvelous system that can store bazillions of pieces of information CAN recall the relevant items when needed IFF we do not try to force the process. Given that System 1 operates in massively parallel mode, this “wait and SEE” approach is an extremely fast process.

10. my thought about improvement is at the moment along this route

we want to improve in chess
chess is a sequence of moves in a sequence of positions
so we have to improve on the very large set of "relevant" positions ( millions, billions..)

we do a training on a small set ( thousands )
and even there we fail to improve

i used the last 7 months the software strategy 3.0 with ~1100 exercises. every exercises starts with a position of a master-game and usually explores several lines int the type of "guess the move". I did start with 5+ repetitions of each exercise in ~2 weeks. Now ~7 months later i check my improvement on these exercises... its small, very very small.

this paper http://chessok.com/?p=21207 is talking about 3 time hours of training 4 days a week ( but several courses?? ) and an improvement from 40 to 55% in strategy. I do 4 hours a day, 7 days a week and my improvement is from 40 to 51%( ? test just started ) . My learning speed is way slower, therefore my forgetting speed faster. I think we reach our plateau when the learning quantity equals the forgetting quantity. Skills don't get forgotten, so the trick is, to make knowledge skill. But how?

11. PART I

I think the answer is: skill is not easy to acquire. As adults, we focus on acquiring knowledge (KNOW-THAT), (incorrectly) assuming that a sufficient quantity of knowledge will (eventually) equate to or translate into skill (KNOW-HOW). It doesn't. Skill lies in USING available knowledge but is NOT the same thing as simply RETRIEVING stored "facts" (knowledge) from memory.

I'm sure that you are familiar with the concept of the opposition in the endgame. Given a knowledge of the "rules" of the opposition, you still may be unable to USE that knowledge to play a particular endgame requiring application of that knowledge. Why not? Because your thoughts may be focused on some other aspect of the position, and you may not even attempt to recall and use that knowledge until it is too late. Let's take an example from GM Rowson.

4b1k1/1p3p2/4pPp1/p2pP1P1/P2P4/3B4/1P6/1K6 w - - 0 1

GM Rowson devotes pgs. 69-74 with a dialog between himself and a student regarding this position. He notes that it is a difficult position to analyze, even for GMs.

There are 4 "ideas" involved:

(1) The White King must be activated. (Obviously)

(2) A sacrifice of the White Bishop for a Black Pawn may be necessary at some point. (Obviously)

(3) Black must be prevented from creating a fortress, stopping any invasion by the White King. (Not so obvious)

(4) The eventual "solution" MAY involve the opposition and Zugzwang. (TOTALLY OBSCURE! It's too far in the future to calculate concretely.)

You can SEE the contours of the first two ideas "on the surface", so to speak.The third idea takes a bit of time to grasp, usually after trying some "obvious" moves, and finding out the the way into the Black queenside has been permanently barred to the White King.

It's the fourth idea that is the key to the solution. So, how do we "SEE" it at this stage of the game?

It is a Capablanca-inspired "technique" (skill, if you will) of simply removing pieces or rearranging the pieces (mentally) to a desired position without regard to legal moves. It does NOT entail calculating moves (at least not in the initial stages of arriving at the right overall idea).

12. PART II

Let's jump directly to idea (4). What changes would you make in order to make this position dependent on opposition and Zugzwang? First of all, chop off the Bishops and all a-file and b-file Pawns for both players, with White still to move. How do you gain the opposition? 1. Kb2 Kf8 2. Kb3 Ke8 and what MUST White play to win? If you know the "rules" of the opposition (both direct and distant), you know that the correct move is 3. Ka4! But that contradicts the idea of trying to rush toward the opposing King to set up the direct opposition. If 3. Kb4? Kd8! and Black now has the distant opposition and will be able to draw. If you cannot "see" this without calculating, then I suggest that your skill in UTILIZING the opposition is deficient.

FWIW: I prefer Euwe and Hooper's method for determining distant opposition. It is easy to apply to virtually any position without calculating moves.

Back to the drawing board. The Black King is confined to the square d7-d8-h8-h7 as long as the White Bishop remains on the board AND there is a possibility of sacrificing it for either the g6 or e6 Pawn when Black is unable to capture with the Black Bishop or when that Bishop is not on e8 square. Again, this must be “seen” rather than calculated. Why? Because the f6 Pawn could then advance to promotion. So the Black King’s movements are restricted, whereas the White King is unrestricted. More freedom of movement offers more possibilities for winning. That dispenses with idea (2).

Let’s look at idea (3), building a fortress. If the queenside can be rearranged in such a way as to prevent the entry of the White King, then the Black King can be utilized to prevent entry on the kingside, establishing the fortress. Knock off the two Bishops, and rearrange the Pawns on the queenside to prohibit the White King’s entry. As the position is initially, there is already a blockage in place: the a5 Pawn and d5 Pawn prevent the entry of the White King. All Black has to do is maintain that blockage and White cannot penetrate via the queenside with the Bishops removed.

That leaves us with idea (1), activating the White King. One try is to get the White King in on the kingside. Ain’t gonna happen with the Black King over there! So, White must figure out how to combine all these ideas and penetrate via the queenside with his King.

13. PART III

Timing of the various "strikes" must be exact, so it's not good enough to just play any old moves. This is where concrete calculation comes into play.

Without going any further, what conclusion would you now reach?

If you “guessed” 1. Bb5?, you would be in good company! BUT, that allows Black to set up the fortress with 1. … Bxb5 2. a4xb5 b6 and the fortress has become impregnable. An eventual b4 will be met with … axb4 and the fortress holds.

I’ll cut to the chase here, and just give the initial moves: 1. b4! axb4 2. Bc2! And White prevents the fortress, activates the White King (advancing to recapture the Black b4 Pawn), eventually either forcing off the Bishops or putting Black into a position where White can sacrifice the White Bishop for a Black Pawn at the appropriate time when Black can no longer protect all of the various points because he has insufficient running room and/or he loses the opposition and is in Zugzwang.

So after all that verbiage, what conclusion do we draw? SKILL involves connecting the various pieces of KNOWLEDGE into a coordinated plan of action. Some of the ideas enable and some constrain what is possible. It is the PROCESS of “seeing” ALL of the potentialities AND meshing them together coherently that represents skill. IMHO, this can only be done by (1) acquiring the KNOWLEDGE and then (2) APPLYING that knowledge repeatedly in various positions until the processes involved (NOT THE SPECIFICS OF ANY GIVEN POSITION) become subconscious (second nature). Until that occurs, we have knowledge without skill. It's not what we know but what we know that we can use that is important. This is a large part of why I have opined previously that memorizing 50,000 or more positions will NOT ever result in skill beyond a certain minimal level (certainly NOT master level or above).

14. Here about the training-improvement-method of Silman : https://www.chess.com/article/view/getting-past-plateaus
500 games in say 20 h, that are 25 games an hour, that are less than 3 min per game, thats bullet speed.

15. The key is System 1 (the subconscious):

". . .the hundreds of positions, structures, and patterns were drilled into my subconscious and began percolating quietly, until the time that everything would suddenly make sense!"

If you cannot move the knowledge from conscious to subconscious, you will NOT improve in skill. In spite of the heroic efforts (Silman and MdlM, for example), there has to be something more than mere exposure to a huge quantity of games going on. Why? Because there are any number of people who have accomplished the exposure, and still not improved in skill.

16. yes , i think the same. I wonder if something like conscious thinking really exist.. i guess its just a focus of attention. So i am not sure if we really need to understand anything consciously first. Skills often cant be "explained".
What irritates me most at the moment, is the learning speed of better player. They learn ( memorize ) chess related things faster. Subconscious can only work on things which are still in the memory somehow. When i watch Chess masters like Ben Finegold at youtube i am always astonished how many games they know.
I hope to improve my learning speed by learning game-fragments sorted by (strategic..) themes and keep this knowledge active for a long period of time.
Silman did learn "master chess" with many signals but only short duration of each signal, i try the opposite, small amount of signals long duration ( = many repetitions ).

An other shocking example of learning is https://chesstempo.com/chess-statistics/cmuroya17
He did start as an 1800, now his performance is 2700, but he "solves" the problems he learned now at hyperspeed.. he simply learned all puzzles at chesstempo. He almost never fails a puzzle twice. His speed at new puzzles is slow but his performance is now very high at them.

17. "I wonder if something like conscious thinking really exist.. i guess its just a focus of attention. So i am not sure if we really need to understand anything consciously first. Skills often cant be "explained"."

You certainly have a few good points here. What we call conscious thinking mainly consists of simple associations brought to mind by subconscious processes. The key is indeed the focus of attention which is the narrow keyhole through which we look to the world.

I don't think that acquiring skills is the problem. Since that is a process that happens just magically since unconsciously. We might even have acquired already enough skills. And if not, acquiring the omitting skills is not a problem.

What I see that the problem is, is revealed by measuring my time usage while solving problems. It lies in the realm of the focus of attention. There is nothing wrong with tunnel vision, as long as you have entered the right tunnel.

The problem seems to be twofold: entering the wrong tunnel and overloading the STM.

The tree of scenarios is a help to find the right tunnel. Overloading is prevented by pruning the tree of analysis. That pruning is based on the tree of scenarios.

The knowledge we need isn't rocket science. It is actually very simple, and even a child can understand it. The problem is that the area is so vast, and the mind is so easily overwhelmed. I have unearthed a few ideas about the initiative. And already I feel I'm better at problems that revolve around the initiative. Before my latest distraction, aka break, that is. But then I get a series of problems that need knowledge of promotion, and I'm lost.

It takes time to plug the main holes in my tactical bucket. The main holes are: initiative, promotion and mate. There is a lot of systematic work needed to unearth the relevant knowledge. But once unearthed, it isn't rocket science. And the transformation to skill happens before you understand what you are doing.

Guide the focus of attention.

18. my pass 2 at "Strategy 3.0" is frustrating. it seems as if i found a new non-improvable ( for me ) set. i suspect that any set ( of chess puzzles any types ) is non-improvable if it is just big enough.

I did experience "non-improvability" now at 3 sets : very large set of mate in 1, 4000+ simple ( rated<1800 ) checkmate in n, and possibly(??) 1100 tasks of strategy 3.0.

Definition:: non-improvability of a set ( of puzzles/tasks): Its possible to improve at a small subset, but after a while ( days, weeks, months ) the improvement at such a subset vanish to 0!. No matter how the training is organized: there remains always a large subset ( not tackled for a while ) without measurable improvement.

19. Perhaps we rush too quickly to choose which tunnel to run down.

1. @ Takchess:

Definitely! System 1 will throw up an answer (ANY "close" answer, which may or may not be exactly what we should be looking for) for System 2 to examine more critically. However, System 2 is somewhat lazy, and (unless focused attention is brought to bear) will just run with the System 1 "suggestion" without any deeper consideration. Think of the aphorism, "Good enough for government work."

I've often thought about why GMs take so long over moves IF they can (almost instantly) "see" everything through the lens of pattern-recognition. The only conclusion is that they have learned NOT to trust System 1's first impression, and they wait for all System 1 "suggestions" to occur before focus their attention (using System 2) on exploring ALL options thrown up for consideration. In this sense, it corroborates GM Kotov's "thinking process" of FIRST listing ALL the candidate moves BEFORE diving down into a specific tunnel. Only when System 1 s no longer tossing up impressions do they note all of the possible suggestions. It's NOT a totting up of all possible moves, but a complete acknowledgement of ALL System 1's "suggestions" instead. In this regard, my excerpt from GM Rowson above makes sense. Just "looking" at the position and allowing the process to unfold naturally without rushing to judgement:

. . . images observed long enough will cease to be random or disconnected and will [SELF-]organize into symbolic dramas, narratives, or PROBLEM-SOLVING PROCESSES.

I can state quite clearly that waiting for System 1 to make appropriate "suggestions" can be nerve-racking. It occurs quite quickly but it seems (at least at first) that it is taking a long time.

I'm finding that my first impression may cover SOME of the right stuff, but (usually) it does not cover ALL of everything that needs to be considered. I still have to consciously override System 2's "rush to judgement" (i.e., the tendency to dive down into the first available tunnel that consciously opens up). It's getting easier to "hover" longer over the scene like a vulture, but it's still not EASY (meaning: it's not YET a skill).

2. I think a number of us System 2 guys have become reenergized becoming system 1 students. Cheers

20. Promotion (some random thoughts)
Two connected pawns always can promote (I heard this said, Valid?)
A pawn nearing promotion demands attention and causes immobility
A pawn promoting causes a imbalance in power and sacrifices of material often occur to
make it happen. (often on the eighth rank with a pawn recapture)
The pawn movement toward promoting is often a form of initiative
when a puzzle has a pawn on sixth rank promotion may be a factor
special cases
-pawn which promotes with check
-pawn which promotes with double attack
-pawn which promotes with a pin of a piece to the king
-pawn which promotes whose capture removes a guard
-Pawn which under promotes to a Knight

21. I came across the following excerpt in Victor Henkin's 1000 Checkmate Combinations, Before you open the book. . .:

YOU SHOULDN'T TAKE 'TACTICAL MEDICINE' IN LARGE DOSES. Even the most beautiful combinations can set your teeth on edge if you swallow them with the greed of a hungry pelican. . . . You should get to know the combinations carefully, thoughtfully, without distractions or hurry, returning again and again to the examined positions. Best of all, LIMIT YOUR DAILY RATION TO TWO OR THREE 'DISHES'.

That certainly seems to contradict our 'usual' improvement idea that practice in huge quantities will (eventually) produce quality (skill). GM Henkin seems to have been highly admired by GM Mikhai Tal, who wrote the Foreword, titled Don't reinvent the bicycle. I quote GM Tal:

When I'm studying a specific position, above all I note its peculiarities, the reciprocal positioning of the pieces, their connections. And suddenly (in the majority of cases this occurs intuitively) somewhere nearby the indistinct features of some new teasing and appealing position becomes faintly perceptible. It isn't on the board yet, of course, but everything points to the fact that it may arise. The hunt for the blue bird begins. Often the calculation of variations turns out to be a Sysyphean task. The position in your mind's eye hardly ever comes about, even if your opponent joins you in a 'cooperation'. Some piece is on the wrong square, some pawn is getting in the way...

But it can happen that the tedious calculation of variations brings real results. Move for move you get exactly the same position that you saw from afar. And if the circle of variations has been exhausted, then ... Then you can start the combination.

Can we find improvement through slow and steady study of a small number of prototypical positions, rather than by consuming enormous quantities of tactics until our eyes bleed?

1. There are several suggestions how to study the games of masters. Silman and others suggest to study hundreds of games per day in "hyperspeed" like 2 or 3 minutes per game. Others suggest to play : Guess the move from the perspective of the winner of the game. And there is the suggestion to memorize games of masters.
I suspect all of them work, you find for each method several gm's supporting it.

2. Can we find improvement through slow and steady study of a small number of prototypical positions

I think we can. All other methods failed anyway.

3. GM Rowson (again): . . .It really doesn't matter WHAT you study, the important thing is to use this as a training ground for THINKING rather than trying to assimilate a mind-numbing amount of information.

I'm not sure we have a good definition for the kind of THINKING that should be practiced. My "gut feeling" is that Mr. Linhares gives us some valuable clues that point toward the investigations currently under way here.

I think the difference between successful (so very few!) and unsuccessful (the vast majority) adult improvers is not in WHAT is studied but HOW it is studied. A relatively small field of study coupled with regular deliberate practice will produce more benefit than massive cramming "more stuff" into LTM, MdlM and IM Silman notwithstanding.

Yeah, I know: Physician, heal thyself. I'm working on that!

I think that most GMs/IMs who began training early in life can effectively utilize the various recommended methods. Obviously each of those methods works, since it did work - for THEM! What I think they overlook is that none of these methods is a universally applicable panacea. What worked for them as rapidly improving children may not work at all for older adult improvers.

FWIW: IM Pruess' "method" for memorizing games is based on spaced repetition. Play over short games at first, several times until the game can be recreated without error at least three times in a row. Repeat this process the next day. If still successful, skip for one week, and try it again; otherwise, start over again whenever you fail to recreate the game. After skipping several weeks and still successfully recreating the game, it is considered to be "memorized." Gradually add in longer and more complex games, rinse, lather and repeat.

Surely we can agree that the evidence supports spaced repetition as an aid in recall/remembering via reduction in the "forgetting" curve that seems inevitable in humans.

I note with interest that after he goes through the memorized moves of the game, he then starts "thinking" about the ramifications of the individual moves. Here is where I think the most value lies for increasing SKILL, rather than merely the rote memorization of games per se. Thinking about alternatives and WHY those alternatives may or may not be good is an essential part of the THINKING that must be practiced.

4. Surely we can agree that the evidence supports spaced repetition as an aid in recall/remembering via reduction in the "forgetting" curve that seems inevitable in humans.

I don't think there is a significant role for (spaced) repetition in the transition from thinking to skill.

It really doesn't matter WHAT you study, the important thing is to use this as a training ground for THINKING

It DOES matter what you study. If the feedback tells you that you fail at promotions, you must study promotions. The most time consuming part is to unearth knowledge that you don't know. You must unearth it because it is not known by your opponents of equal strength either. The actual ASSIMILATION of knowledge should take very little time and effort. Just like assimilating the knowledge how to drive a car did request only very little time and effort. The knowledge based thinking should cause no difficulty. Once you know WHAT to think about

The most stupid thing you can do by the way, is to publish your findings on a weblog. Since you can only get better when your opponents remain the same. Luckily this blog is unreadable for most people. :D

22. PART I:

On a similar front, regarding the current Tree of Scenarios investigation, based on PoPLoAFun:

I have been re-reading from Alexandre Linhares' paper, Decision-making and strategic thinking through analogies.

How do humans perceive similarity at a strategic level?

Strategically similar chess positions may be explained by looking at the configurations from the point of
view of sets of abstract roles. For deep understanding, that abstraction is the most efficient level of
description. While a skeptical reader might consider that the processing of both trajectories and of roles
may be redundantly inefficient, that is really not the case. It is crucial to process both trajectories and roles.
The distinction is an important one. Because a particular trajectory is bound to a set of squares, it is not an
adequate structure to either perceive the situation in abstract terms or to generalize to other superficially
different but strategically similar scenarios. Trajectories lead to chess relations, but are different from
them. Chess relations are independent of locations and of particular squares of a trajectory. And chess
relations are subtly different from abstract roles: In Capyblanca’s evolving interpretation of a position, roles
emerge gradually from the set of chess relations, by the process of continuous re-evaluation of the intensity
of each relation.

This is a distinction, not an argument. By structuring information in terms of abstract roles it is possible to
obtain abstract representational invariance. Consider, for example, the notion that, “under a double check,
the only response is to escape with the king”. This notion is generalizable to an immense set of situations.
It would be unthinkable—or at least computationally intractable—to try to describe it in terms of particular
pieces, particular squares, and particular trajectories. The most efficient level of description is the
abstraction “piece A has a role of attacker towards piece X”, “piece B also has a role of attacker towards
piece X”, and “X is a king”. It is irrelevant for understanding, and inefficient for processing, what type of
piece A and B are, what their mobilities are like, or even which squares they occupy. Therefore, it is
imperative, for an architecture with the ambition to model human understanding of a chess position, to use
different representational structures to account for the trajectories involved in chess relations, and separate
structures for the abstract roles that are found to be relevant in a particular position. Deep understanding
arises by recognizing a familiar, previously seen, configuration of abstract roles. For understanding a
scenario, the most efficient level of description is given by a combination of abstract relations representing
the most important roles objects play. By reasoning at this level, the system is able to perceive analogies
between dramatically different scenarios (Linhares 2008). This enables the eventual emergence of a global,
meaningful, understanding of the constraints and key issues underlying the strategic situation one faces—
and leads us to our final claim concerning the architecture.

23. PART II:

Unless I am badly mistaken, this is the very essence of what we have been exploring. Aox asked a simple question previously: Skills don't get forgotten, so the trick is, to make knowledge skill. But how?

I think the answer lies in Mr. Linhares observation above: sets of abstract roles. I add a caveat: the identification and absorption of those sets of abstract roles must be performed by each individual from scratch. Mere familiarity based on someone else's hard work will NOT suffice to ingrain those abstract roles into the subconscious, thereby producing skill. Trajectories are Lines of Attack (LoA); roles are Functions (Fun). But according to his analysis, this still does not get us where we need to be-at a high skill level!

Trajectories -> chess relations (not the same thing)

Chess relations are independent of locations and of particular squares of a trajectory.

Chess relations are subtly different from abstract roles.

Roles emerge gradually from the set of chess relations during an evolving interpretation of a position, by the process of continuous re-evaluation of each relation.

When we have worked out the trajectories, specific roles and relationships between the pieces and squares in a specific position, and then formulate them in our own terms as abstractions, then we (possibly) have gained in skill.

Somehow, I am reminded of Dr. Lasker's ideas once again. Different terminology, perhaps a different emphasis, but still the same ideas.

If one of the greatest world chess champions and a modern cognitive scientist agree on the approach, perhaps it merits our attention and careful study.

It will be interesting to see what direction Temposchlucker takes his investigation of chess improvement when he has time to resume his studies and blogging!

24. In my opinion the problem of complexity lies in the nature of chess. There are some positions we humans have great problems to navigate, but the engines can play these without any effort. Sometimes their evaluations simply shock me and I have to see the variations to try understanding why the hell is such an eval, not the opposite (or at least much higher/lower).

And the more I think the more I am convinced everything starts with the familiarity of the position. If you analysed the A-type positions pretty deep end extensive, you will succeed and the chance to go wrong diminishes significantly. However there are more elements we have to take into consideration. We have to probe and visualize some variations to see what TYPE (kind) of positions is at the end of these. And think of simple positions with big advantage - 95% of decent players can steer the game to the point of victory, do you agree? Why is that? It is because they have simple plans and nothing can be done to stop. However the simple plan comes from the lack of compensation and therefore we call some positions as "one-sided". And when you reach so called complex positions we are simply blind to the best scenarios and we lost track. What would it be if we reach a complex position 001 we had analysed and practiced extensively before? The chance we succeed is much higher and it is not because we can calculate well, but because we can recognize the position, we know what are the plans for both sides and what should be done (or avoided).

To sum it up: the more TYPES of position we understand really well, the higher chance not to go wrong. That was Lasker's idea (invention) when he stated that method is the only what counts.

And one of the reasons we cannot play chess as best as we wish... is lack of mental energy. Try to imagine being focused for 4-5 hours straight with maximum performance required. If you have to count, evaluate and calculate "everything" it is close to impossible you will do it flawlessly... especially when your problems on the board are complex ones.

BTW. What I have learnt so far from engines play and evaluation is the extended range of shocking exceptions to the well-know rules of strategy. Tactics have been analysed and used at chess for a very long period, but computers showed us how much tactics and dynamic play can help at defending lost (or bad) positions. That's why I appreciate these so much - they are great tools to learn REAL (deep) chess.

25. PART III:

In his paper, Mr. Linhares gives 6 chess positions to illustrate his notion of the abstract nature of strategic similarity. These 6 positions are given below in pairs: A & B, 6 & 10, and 8 & 20. I have modified the positions slightly in some cases, to provide more of the “Aha!” experience of recognizing the similarities at a very abstract level. Superficially, the pairs of position bear no resemblance to each other whatsoever. Yet, the same abstract “idea” or “pattern” intimately connects them.

These abstract roles are the “patterns” that must be embedded into the subconscious in order to produce skill. It is my contention that the total number of abstract “patterns” that must be known is considerably smaller than the usual estimates of 50,000 (100,000; 300,000) that are needed in order to provide the basis for masterful play.

Position A FEN: rnbqkbnr/pppp1ppp/4p3/8/6P1/5P2/PPPPP2P/RNBQKBNR b KQkq - 0 2
Position B FEN: 1b6/8/8/8/8/8/2k1P3/4KB2 b - - 0 1

Position 6 FEN: 8/8/7k/4pPp1/3pP1P1/2pP3K/2P5/7B w - - 0 1
Position 10 FEN: 8/1k6/1Pp1K3/2Pp4/3P4/8/8/8 w - - 0 1

Position 8 FEN: 5r1k/5p1p/5P1N/8/8/8/4p3/K5R1 w - - 0 1
Position 20 FEN: 1kr5/ppN5/8/8/5Qp1/1r1p2P1/2b2K2/8 w - - 0 1

Note that in the last pair of positions, the classical literature may categorize and name these two positions as distinct and separate mating patterns. Yet, considered at the abstract level, there is the basic concept of the smothered mate. This is what we must assimilate into our subconscious in order to build the foundation for skill.

1. A master is aware what squares are attacked if a given piece is at a given square : 6 pieces 64 squares say 10 attacked squares = 3600 chunks. This are just the 1 piece chunks. I guess that just the board vision chunks are already in the region of 50 000

2. @ Aox:

Your math is correct, and I will not quibble over it. I personally haven't seen any literature where this is described or presumed to be the basis for "chunks" as I used the term above but I'm willing to consider it. As I currently understand it (perhaps I DON'T understand it at all!), a single "chunk" is composed of interrelationships of pieces and squares, pieces and opposing pieces attacking/defending/blocking, and abstract "functions" (trajectories plus roles) all combined into a very high level abstraction which is independent of the individual pieces/squares involved, i.e., "sets of abstract roles" as described by Linhares above:

"Consider, for example, the notion [an abstract "rule"] that, “under a double check, the only response is to escape with the king”. This notion is generalizable to an immense set of situations. It would be unthinkable—or at least computationally intractable—to try to describe it in terms of particular pieces, particular squares, and particular trajectories. The most efficient level of description is the abstraction “piece A has a role of attacker towards piece X”, “piece B also has a role of attacker towards piece X”, and “X is a king”. It is irrelevant for understanding, and inefficient for processing, what type of piece A and B are, what their mobilities are like, or even which squares they occupy."

On the same basis, one could extrapolate from the number of rods/cones activated every second to a number in the billions, if that was used as the basis for "chunks." Given the fine granularity at that level, I don't think it helps us to THINK better while playing chess. Numerical counts and descriptions (regardless of how fine-grained) are not prescriptive; "how to" is needed for improved skill, not "know that." We can each pick our own favorite way to categorize "chunks" (and NONE of us have direct access to "chunks" in our memory) but categorizing will NOT help increase our skill IMHO.

26. As far as I can tell THE ONLY unrefutable proof of making progress is studying the games of masters (or even grandmasters).

1. The opening part - what are the plans, ideas and concepts at specific opening and even inside a variation.
2. Middlegame - what is the position evaluation (assesment) and the plan for both sides. In between there are shown some hidden tricks (traps) and you can familiar with these.
3. Endgame - what is the plan of actions, which side has the advantage, how to calculate variations with a precision and how to convert the advantage into the win.

There are a lot of players who confirmed that studying master games helped them in making progress. And I think it is quite reasonable approach as you cannot ask for me. That's why all trainers, writers and coaches always underline the role of well annotaed (master) games from tournaments, World Championship matches or even team league games.

I wonder what would it be if anyone devoted himself an studied 500-1000 well annotated master games very seriously. I think in most cases the progress would be anavoidable!

1. I wonder what would it be if anyone devoted himself an studied 500-1000 well annotated master games very seriously.

What are you looking for if you study just one game seriously? You look for ideas that are common to all games. Since if an idea is only used in that very specific game that you are studying, then that idea will not be applicable in other games. Which renders that specific idea as useless for studying. Scenarios concerning the initiative, that get you an extra tempo, happen in every game. You will find points of pressure in every game. You will find lines of attack in every game etc.. That's where the tree of scenarios comes in. It are scenarios that play a role in every game.

Don't get mesmerized by numbers. It is not about how much master games you study but about how seriously.

2. It can be argued that a good master game and the thoughts behind it may be valuable as a continuous whole. The struggle from opening to middlegame to end. A number of grandmasters attribute a leap of understanding to a deep serious study of Bronstein Zurich tournameny book.

3. As long as I perform as bad as I do, it makes simply no sense to even try to think beyond tactics.

27. @ Tomasz:

"I wonder what would it be if anyone devoted himself and studied 500-1000 well annotated master games very seriously. I think in most cases the progress would be unavoidable!"

I think you are right: it SHOULD (NOT would) lead to a significant improvement in SKILL because the emphasis would be on THINKING through WHY the moves were played, and what alternatives were considered and rejected. Few of us are willing to put that kind of SUSTAINED EFFORT into it. It doesn't give us nearly as much feeling of immediate "improvement" compared to solving many thousands of isolated tactical problems at the speed of light.

GM Rowson, in Chess for Zebras: Thinking Differently about Black and White, Chapter 6: Why is Chess so Difficult? The Exponential Problem:

A few weeks ago, I advised a student of around 2000 strength to TRY TO ANALYZE ONE COMPLICATED GAME DEEPLY AND CAREFULLY. [Emphasis added. Note he did NOT advise to MEMORIZE one or more games.] I asked him to keep on honing and revising his findings until he felt confident that I wouldn't find any major mistakes in the analysis. I encouraged him not to use any analysis engines until he had given it his best effort. He was not allowed to make generalizations about the position and not allowed to stop with the assessment that the position was 'unclear' unless he could demonstrate the effort he had made to find clarity. He is a diligent student and keen to improve, but looking at a complicated game in depth until he understood it move by move, variation by variation, proved to be too much for him. He devoted a few study sessions to this task, a couple of hours at a time, but he came back to me and said that he was 'scared' to do any more. Whenever he tried to look at the position without prejudices he felt 'lost' and 'helpless' and this gave rise to a sense of despair. It was clear to me that he had glimpsed what Nabakov called "the abysmal depths of chess".

That should not (and does not) stop us from trying.

The more you can "see", the more you realize how much more there is that you still CANNOT "see."

Like most people who remain amateurs after many years of studying some isolated aspects of chess, I will continue to IGNORE what has been shown to work, while pursuing the will o' the wisp of "improvement by other means." After all, doing it 'my way' has resulted in such great progress - NOT! Some day, if I "wise up", I'll try that recommendation!

1. I suspect it dont matter how you study games, it matters how your learning speed at chess is.

Here some Words of GM Axel Bachmann:
"In my career I have seen close to 100,000 chess games, including most of the grandmaster-level games played over the past decade. The cumulative experience from [b] spending a minute or two [/b]on each of these games has allowed me to gain an excellent positional understanding. Staring at a position for a few seconds is often enough for me to see who is better, which plans will work, which pieces should be traded, etc.
Acquiring such a level of experience and positional knowledge requires many years. Going through
thousands and thousands of games takes a very long time, even if you only spend a couple of minutes
on each. Most importantly, being able to actually see the patterns does not come easily to everyone."

They dont think ( a lot ) .. they just watch and see and learn. Like kids
This "method" is very common
I suspect that here is a reason for the difficulty to improve for older folks. The wich for a "verbalized" explanation. I suspect that almost all of the chessknowledge cant be verbalized.

28. @ Aox:

"I suspect that almost all of the chess knowledge can't be verbalized."

An interesting hypothesis, with which I concur. The "obvious" stuff has already been defined, refined, and published many times over the years.

This may be the basis for IM John Watson's assertion regarding annotations in Secrets of Modern Chess Strategy: Advances since Nimzowitsch, page 103:

Description Versus Reality

Before entering into discussions of specific rules and principles, I should make a simple distinction which applies to my notes as well as anyone else's. ONE MUST ALWAYS KEEP IN MIND THE DIFFERENCE BETWEEN A DESCRIPTION OF PLAY AND THE PLAY ITSELF. FOR ALL I WILL SAY ABOUT REJECTING RULES, IT IS STILL TRUE THAT WE MUST USE THEM AS TOOLS WHEN ANNOTATING GAMES. Thus, for example, there is no substitute for saying something like: "and Black stands better because of his two Bishops and White's backward pawn on the open d-file." One simply has to bear in mind that such a statement has an implied subtext, for example . . . [very long example omitted].

Naturally we don't kill trees for the sake of such [verbose] explanations, which in reality are usually even more complicated and qualified than the one I have given. Instead, WE USE ABBREVIATED STATEMENTS OF PRINCIPLES AS INDICATORS TO GUIDE THE READER'S THOUGHTS IN THE DIRECTION OF OUR OWN. IT IS VERY IMPORTANT TO REALIZE THAT A PLAYER'S USE OF SUCH DESCRIPTIONS IN WRITTEN NOTES BY NO MEANS IMPLIES THAT HE HAD GIVEN [CONSCIOUS] THOUGHT TO THEM DURING THE GAME. I think that there is a great danger here for the student. He or she will pick up a book of annotated games by some world-class player and ASSUME from such general descriptions that "THIS IS THE WAY THE GREAT PLAYERS "T-H-I-N-K". In reality, most players are unconcerned with giving exact descriptions of their thought-processes; it is much easier to characterize a position generally, with hindsight, and ignore the gory details.

This passage, along with an understanding of Dr. Lasker's ubiquitous ceteris parabus was very enlightening to me. I had implicitly assumed (wrongly) that the great masters "thought" in terms of "rules" and "principles". If I could just absorb those rules ans principles, I would then play much better as a result. WRONG!

It is only in the last year or so (thanks in large part to the discussions on this blog) that I "see" the "obvious": most grandmasters and masters can "confabulate" why (after the fact) they did something, but would find it almost impossible to do so (honestly) during a game, for the very reason that you gave: "almost all of the chess knowledge can't be verbalized". We verbalize what we can consciously bring to mind, but the vast bulk of our playing SKILL remains hidden from us, encapsulated within System 1 and therefore not directly accessible.

When I stop trying to consciously rationalize my reason(s) for playing specific moves WHILE PLAYING, my SKILL level appears to rise significantly.

29. Another thought:

Has anyone here ever tried just working through an enormous number of games, using a tool that plays the moves automatically (like on http://www.chessgames.com/), only pausing for a very short time (less than 10 seconds) when something "interesting" (not immediately obvious) shows up in the game?

I know I haven't, because of the huge amount of time involved. BUT, I'm beginning to think it MIGHT be worthwhile to try it for at least 10,000 games (or more). The "idea" is not to try to memorize anything but just become aware of the "flow" of patterns between the phases of the game, somewhat in the same vein that Jim Takchess suggested elsewhere.

1. I am working through more or less longer lines and alternative lines of middlegames taken from ~1100 games ( strategy 3.0 of chessok ). Most of the moves are just shown, 30(??) % of the moves have to be found by the user (guess the move).
I am thinking about a "guess the move" program (like peshka) where only engine proved moves will be asked and the other moves will be just shown, maybe : as worse the move as shorter the display?
As an older person i need repetition, why should i look at something if it is forgotten after some days? So i would include a system of repetitions where i make sure that i improve my score on these games at every repetition a little. I would select games of my openings or games where i can find good explanations. (Kasparovs My Great Predecessors maybe with the sidelines? )

30. Robert,
http://www.chessgames.com/perl/chesscollection?user=Qindarka&page=7 These collections can be run through game by game and has an arrow button that allows it to play automatically. Chessgames.com has thousands of these type of collections. Cheers, Jim

31. @ tachess:

Thanks, Jim!

FWIW: the chessgames delay time for autplay can be set by clicking on squares c2 [1 second], d2 [2 seconds], e2 [5 seconds] or f2 [custom delay]. I personally prefer 2 seconds delay. Just run your mouse over the various squares on the board to see the available options.