Monday, June 27, 2016

Elaborating on the initiative

The position from the previous post is one of seven positions of 80 of my failures at CT. The seven positions have in common that there are mutual captures, and I feel I have difficulty to grasp exactly what is going on in them. That feeling is highly personal,and it is difficult to explain. But that feeling indicates there is something wrong with my approach what needs to be corrected. The effects of what is wrong are not limited to this position, it has a much broader effect. Which means that correction will have an impact on my capability to solve other positions of the same type, whatever “type” might mean here. That is exact what we want to learn, skills that transfer to other positions.

We agreed on the idea that we cannot improve our calculation if we cannot improve at M1. No one so far managed to improve at M1 in a convincing way which is usable as template for further improvement. This means we must be smarter than that, and seek improvement elsewhere. One area where we might improve, is that we learn to calculate the right thing, whenever we calculate. Everything that can help us to avoid calculation will make us better. When Tomasz became enthusiast by the position below, he started to calculate all kinds of stuff that have only a far relationship with the diagram, and it is only a matter of time before he would have ended up by Adam and Eve, I guess. Which means, without some method we can end up everywhere. Personally I didn't see the difference between 2.Bxg7 and 2.Bxc5, so I played the latter without further ado. Robert at least had a gut feeling that 2.Bxc5 might give black some counter play.

When we want to prune the tree of analysis, we must be careful not to prune the promising branches. What makes a branch promising? What makes that we know we can prune a branch and be sure we don't throw away the win? The story of the initiative might give an answer to that.

I don't know where thinking about the initiative will lead me of course. And it is probably possible to use other semantics to express the same ideas as well. For instance in the language of weaknesses, which Aox advocates. But I don't want to anticipate on that. Let's have a little chitchat about the initiative first.

For me, the term “initiative” is rather vague. Let's see if we can add some precision. An important factor, I think, is the amount of tempo's we or our opponent need, and how many tempo's we have. A duple attack works because it adds two tasks on the shoulders of your opponent, while he has only one tempo to do so.

After 1.Bxd4 Nxc5. White to move
3qr1k1/2r1ppbp/p5p1/2n3P1/1p1B1P1Q/1P6/P4PBP/2RR2K1 w - - 0 2

The discovered attack 2.Bxg7 works, due to the threat against the black queen. Black wants to do two tasks, capture the bishop and save the queen, but he has only one tempo. The capture Kxg7 doesn't provide a spare tempo for black.
If white plays 2.Bxc5 in stead, what is different?  There is still a threat against the black queen, and black still needs two tempo's to save both his queen and to recapture on c5. Where does this spare tempo come from? The move 2. ... Rxc5 doesn't save the black queen. It introduces a new threat against white. White can still execute his old plan and capture the black queen, if he likes. But he must abandon the rook on c1 to do so. Normally, that would be no problem, since it leaves white with a queen for his rook.

After 2. ... Rxc5, there is a sort of interbellum. An interwar period. White needs a tempo to execute his discovered attack. When he uses that tempo, he hands over the initiative to black. If he doesn't use this tempo, he returns to the previous more or less balanced position, where black is slightly better. According to mister Stockfish.

After 2.Bxc5 Rxc5 3.Rxd8 Rxc1+ white is a queen against a rook up. But he needs a tempo to save his rook on d8. Since he is in check, he is not going to get that tempo, he must first relief the attack on his king.

What has happened, is that the initiative didn't swing right away from white to black, but that there was a balanced position in between. With 2. ... Rxc5, he got his lost piece back, while adding a new threat. The new threat against white was based on the following  factors:

  • The white rooks are mutually overworked. Together they have to accomplish three tasks: Take on c5, take on d8 and protect each other. They can't do all three.
  • The white king on g1. If it wasn't there, white had time to save his rook on d8.
  • The protection of the black queen. If the black queen wasn't protected, white would need no tempo to save his rook after he captured the black queen.
  • The black rook on e8, standing between d8 and the king. If it would be on b8 for instance, Rxd8 would be with check.
Finally there has come some clarity in my mind. Under normal circumstances, the initiative does not swing directly from one side to another, but there is a balanced moment in between, where you can opt out of the attack. I assume that a duple attack is needed to swing the initiative to the other side in one go. I will test that idea.

There is still hanging a whiff of magic about the swing of the initiative, but at least it is clear now how it works, theoretically. I'm going to look at the initiative more now, during the post mortem.

A whiff of magic


  1. We can improve our calculation without improving at M1. M1 is pattern recognition ( mainly ) while calculation contains elements like "memory of already calculated positions and move sequences and their evaluation", "Visulaisation of deep lines", "Organising the calculation in the "roight" order, not calculating in circles", reasoning, knowledge of essential endgames and so on . (Long ) Calculation is k = spending many thoughts while patternrecognition is a fast or even very fast process . Patternrecognition is an important element of calculation estimatingly THE important element, but not the only one.
    Usually the chessplayer have the same ratingperformance at puzzles with fast medium or slow "average time", they perform acording to elos formula if they solve simple or complex puzzles. But there are exceptions.
    people who are used to play bullet and nothing else are getting ( relativly ) weak if the problems get deep and complex, they are not used to do deep long calculations and / or they lose "sight". While others, able to play blindfolded sometimes perform better and better ( relatively ) as more average time and / or as more complex the problem is.

    While i have strong doubts that the pattern recognition can be improved decisivly, i think that k can be improved. Such an improvement would be especially beneficial in slow games and not in blitz or bullet.

    It might sound as if the said above would be "proven" but that is not the case. I would be happy if somone could prove me wrong. Its is seemingly at least wrong for kids.

    Presently i do some "Strategy 3.0" Training. These puzzles are organised by positional aspects. Funny: the most puzzles are more or less tactic puzzles. So some ( many?, most?) tactics can be seen as positional meneuvers too ( as robert pointed out ). Just sad that there are not enough puzzles to realy burn these "positional" pattern into the brain. I will repaet them at Strategy 3.0 several times, but my experience with repetition was definatly negative in the past .

  2. In addition to CCT, perhaps we need to "see" (or at least look for) EST: Equal or Stronger Threat as potential counters to our grand "idea."

    A simple example:

    [FEN: 5k2/6p1/p2q3p/b7/2QB4/7P/P5P1/6K1 w - - 0 1]

    Bruendtrup - Budrich, Berlin 1954 (Understanding Chess Tactics - Martin Weteschnik)

    A series of EST: 1. Bc5 (pinning the Black Queen; protected by the White Queen) Bb6 (counter-pinning the White Bishop, apparently winning the White Bishop because it is attacked 2:1) 2. Qf4+! (creating a 2:1 attack, winning the Black Queen with a fork)

    A relatively simple example, demonstrating a sequence of threats and counter-threats based on pinning and a fork. The initiative seems to belong to White, shifts to Black, and then finally back to White.

    Another good example is the game Adams - Torre, 1920. Link: Take my wife. Please!

    After 17. ... Bxf6: [FEN: 2r1r1k1/1p1q1ppp/3p1b2/p2P4/3Q4/5N2/PP2RPPP/4R1K1 w - - 0 18]

    There is the "obvious" Black back rank weakness. The Black Queen is the protector of that weakness. The first White move is also "obvious": 18. Qg4. The White Queen cannot be taken without allowing checkmate on the back rank. So Black moves the Black Queen to a "safe" square: 18. ... Qb5. At this point, I suspect that most readers will have already seen the continuation. If not, enjoy working it out!

    There is a continuous series of "attacks" which have a very limited number of responses at each move. Throughout, White maintains the initiative by continually targeting the Black Queen, until (finally) the Black Queen has no place left to hide AND simultaneously protect against the back rank mate.

    Another classic example is Bernstein - Capablanca, Moscow 1914. Link: Phi Beta Capa

    After 26. ... Rc5: [FEN: 3r2k1/p4ppp/1q6/1Nrn4/8/2p1P3/P1R1QPPP/2R3K1 w - - 0 27]

    White calculates that now is a good time to remove the c3 Pawn. However, he did not look far enough ahead in order to "see" the vulnerability of his back rank. The "sting" comes at the very end of the sequence. I suspect that Bernstein overlooked the "attack" of the Black Rook on empty square d1, thinking only of the exchange sequence on c3. I think it highly likely that if there had been a piece to be captured on d1, he would have seen that he could not capture on c3.

    I can think of (many?!?) other positions where this same kind of complexity of multiple pieces across multiple squares occurs. I think this is the kind of position that Weteschnik recommended doing a "status examination" of each and every piece (and I would add "squares" as well). But maybe even that is insufficient to "see' the solution as an organic whole; maybe it's just a "hole" in our vision.

  3. Diagram 1 perfectly demonstrates what I tried to tell you: a full swing of the initiative from one side to another, can only be accomplished by a duple move (pin, double attack in this case). All single threats give time to opt out.

  4. I can confirm this statement (a full swing of the initiative from one side to another, can only be accomplished by a duple move), but I could give a bit broader sense: every move what is a bigger threat intercept the initiative. It does not matter if it is a duplo move (attack) or triple one. It only HAS TO BE stronger than your opponent's. Simple examples:

    1. You attacked my Rook, and I answered by attacking your Queen.
    2. You attacked my Queen and I threatened to mate you (1 move checkmate threat).
    3. You threated to checkmate in 3, but I found the idea to mate you in 2.

    That's why so called single threats are not efficient after you reach 1500-1600 level of play. The more multitasking the threats - the better (stronger) they are.

    When I teach my student (a kid) why one move is better than the other I always ask him to COMPARE the moves and list all of the pluses and minues. If the latter wins it means the move is rather not so good.

    And comment to Aox: I agree - moves that accomplish mate in 1 are very fast discovered due to the pattern recognition. However when you have the series of 4-5 moves - the pattern recognition of back rank mate is only a helpful advice which way you consideration should be triggered.

    1. Matters are not very simple, alas. If you look at then you will find that all checks (7) and all captures (2) don't work. Only the right threat works. It doesn't make sense to scan all checks and captures, before you scan the less forcing threats, since you will waste your time.

      Btw, if you change "tactic" by "problem" in the link, like
      you get a button under the board where you can get the FEN of the position. Just in case I'm too lazy or forget to add a FEN, you can get it yourself ;)

  5. After a few dozens of analyzed positions, I come to the conclusion that "the initiative" or CCT is not a viable way to generate a list of candidate moves to ponder. It leads to way too much redundancy. CCT can be used to check a move for counter play. If your opponent hasn't a more forcing move, your move will work.

    It is much better to "think backwards". For that, it is necessary to recognize the main idea of the combination first, and then to find the moves that lead to it. I'm looking further for characteristics of such main idea.

    1. As I said (wrote) in a previous article - the initiative is best when we want to explain positional aspects of chess. However CCT is much better to check if there is any counterplay - or to say in another way - if we may be killed by the simple oversight or 1-2 move threat we cannot stop.

      I cannot agree move to your post. I am sure it is a good way to check out what is the biggest problem when we know there is a tactic, but we cannot discover how the things should be set up. When I was younger I tried to use this concept of "thinking backwards", but in a way to explain why some moves do not work. The simplest idea is to ask if the move reverse makes any difference. There were about 5% of positions where you could change the move order - in all the rest positions doing this simply killed realising the ideas (i.e. the opponent could defend his position and the goal could not be achieved).

      PS. Your analitical work and chess research remind me my own journey at my teenager years ;) :)

  6. FWIW: I found two books in my personal library that directly address the subject of the initiative.

    The Middle Game: Book Two: Dynamic and Subjective Features, M. Euwe and H. Kramer, ©1965, ISBN 0-7135-0432-3, G. Bell and Sons Ltd.

    Part VI. The Initiative
    The Activity of the Pieces
    The security of the King
    To make the exchange or permit the exchange?
    The avoidance of exchanges

    How to play Dynamic Chess, Valeri Beim, ©2004, ISBN 1-904600-15-8, Gambit Publications
    Chapter 5: The Initiative

    GM Beim defines the initiative this way:

    Well then — possession of the initiative means being able to create threats faster than the opponent, and the aim of developing your initiative is ideally to use your threats to forestall the opponent's activities, defensive as well as aggressive. From this it follows that fighting for the initiative always means trying to be ahead in a race! Thus we can see already that the concept of the initiative is inextricably linked to that of time and speed.

    I would repeat that for the player who has seized the initiative, it is very useful — indeed essential — to keep creating new threats to make the opponent's life as hard as possible. This demands inventiveness and quite often boldness too, because not infrequently the only way to sustain the initiative is by material sacrifices.

    As we shall see on several more occasions, the initiative can affect the play not only objectively, but subjectively too. When faced with his opponent's initiative, a player gets worked up; by no means everyone is able to keep cool before the daunting spectacle of threats assailing him constantly and often from the most varied of quarters! It's only natural that in such conditions you make more frequent mistakes; you tend to mistake imaginary threats for real ones, while underrating the genuine danger.

    Handing over the initiative unconditionally, without even trying to fight for it, is a sure fire way of heading for defeat.

    In the most general sense, the concept of the initiative means keeping ahead of your opponent. It follows that this concept embraces everything which increases, or at least maintains, the disparity between your own and your opponent's capacity for active play. So measures for reducing or wholly forestalling your opponent's activity are also a contribution to the initiative.

    . . .allow us to formulate one more conclusion that is important, although generally fairly obvious: the fight for the initiative can be conducted by the most varied of methods and can assume a multitude of forms. But once the initiative is in your hands, one strict rule applies: it is essential to sustain your initiative and endeavor to develop it. If you don't, it will pass to your opponent, and you will be in for a hard time.

    The initiative in chess can likewise be represented as a certain flow of activity directed from one side to the other. If the disparity in activity decreases or disappears, the initiative will be reduced or extinguished.

    1. That is what I meant with that the description of the initiative is usually pretty vague ;)
      Now you know how to swing the balance of the initiative: look for moves that are aiming at two (or more) positional goals. Only duple positional moves are sufficient.

      In positional play the same is true: it starts with two ideas. From there you try to find the moves that lead to the realization of the ideas. Backward thinking.

      Aox starts with weaknesses, which is one way to get on the track of ideas. It is a form of backward thinking too.

    2. The ever-present duality:

      (1) Start our vulture's eye view from our strengths and look for the opponent's corresponding weaknesses.

      (2) Start our vulture's eye view from our opponent's weaknesses and look for our strengths.

      Opposite sides of the same chessic coin.

      I think your conclusion regarding duple (combinational or positional) ideas corresponds to Yuri Averbakh's generalization of everything in chess to "double attacks." Obviously, great minds think alike!

  7. I am going to solve some (tactical) puzzles at lower level (1400-1600) and see if I can find any relevant elements we can discuss. The simplest concepts of captures and checks is probably not interested to anyone, but I will see if there are any hidden factors we can discuss or think over.

  8. I've been mulling over the comments to this post. I'd like to focus on the "interbellum" or interwar period. The chess literature refers to these types of positions as Zwischenzug or Zwischenschach (if the "in-between" move involves a check).

    It is the nature of such a move that it involves an Equal or Stronger Threat (EST); if it did not, then it could simply be ignored. That is so patently obvious that we often overlook the significance of it when trying to "see" the solution.

    In the position under discussion, there are various "in-between" alternatives potentially available to Black IF AND ONLY IF White plays 2. Bxc5; those alternatives are NOT available if White plays 2. Bxg7.

    There is the geometrical relationship between the WRc1 and the BRc7. With the Black Knight on c5, there is no possibility of Black playing an "in-between" ("discovered attack") move of the Black Knight as an EST: the White King, White Queen, and White Rook (d1) are out of reach of the Black Knight. On the other hand, with the White Bishop on c5, there IS the possibility of an "in-between" move: the Black Rook can capture it (regaining the material balance AND now the WRc1 is attacked, potentially with check (a Zwischenschach). This changes the dynamics in Black's favor.

    The interrelationship of the pieces (WRc1 vs BRc7; WRd1 vs BQd8) provide the necessary "clues' to consider the potentiality of EST as a counter. I think that was what gave me the "gut feeling" that 2. Bxg7 should be considered rather than 2. Bxc5. (Since all of that was spewed by my subconscious, I cannot be certain.)

    I don't think it is necessary to actually perform much concrete calculation to distinguish between NO potential EST versus actually creating the potential for EST for the opponent.

    1. The interbellum I meant, has nothing to do with a zwischenzug. It is more like the slack tide between ebb and flood. During one ply none of the parties has the initiative. After 2. ... Rxc5 neither side has the initiative. White had it, but lost it. So he hasn't time to execute his originally planned discovered attack. With execution of a duple attack, I mean to gain the wood.

      The execution of an attack costs a tempo. So when you win a piece, you loose the initiative. Remember the phrase "the threat is stronger than the execution"? That is because you loose the initiative.

      If white plays 3.Ra1, he accepts that he has lost the initiative. Only a set of pieces have been exchanged. Nobody won a piece.

      So to swing the initiative from one side to the other, two plies are needed. One ply to let white loose the initiative, and one ply to let black get the initiative. The reason I bring this up, besides that it was unclear to me so I had to investigate it, is that it implies the hypothesis "to swing the initiative from one side to the other, two plies are needed, and only a duple move can do it within one ply".

    2. I discarded the initiative as a viable method to find the idea of a combination. Once you think you have found the right moves though, you can (t)EST the moves. If the opponent has a stronger move than you, your combination will not work. EST for blunder checking.