Monday, January 18, 2016


Latest update Jan 19, 2016

To get a little order in the mess, a few definitions are given. The goal is to simplify the discussion somewhat. There is no pretension of correctness or completeness. The definitions are only applicable for chess improvement. In broader domains these definitions are not usable. Please feel free to add suggestions or corrections. I will add this list to my sidebar.

Aura = the squares that are covered by a piece. The lines or geometrical figure that emanate from the piece.

Chunking = is a term referring to the process of taking individual pieces of information (chunks) and grouping them into larger units. By grouping each piece into a large whole, you can improve the amount of information you can remember. It is especially used to enhance the STM.

Combination = combination of motifs

Cue = A Retrieval Cue is a prompt that help us remember. When we make a new memory, we include certain information about the situation that act as triggers to access the memory.

Duplo attack = when multiple (usual two) targets are attacked by just one move. There can be one or more attackers (discovered attack). Examples: pin, skewer, X-ray attack, double attack, discovered attack, fork.

Motif = A combination exists of combinational motifs. The function motif and the geometrical motif seem to be the most important.
  • Function of the pieces. Role and status. Tells you the vulnerabilities due to obligations.
  • Geometrical motif. Gives you a glance into a possible future. Prevents tunnel vision.
  • Undefended piece. LPDO
  • Encircling. Mate and entombing.
  • Assault.
Pattern = Patterns come in a lot of flavors. Geometrical pattern or a tactical theme are used the most here.
Tactical themes = For the sake of simplicity these are the tags from CT:
  • Advanced Pawn
  • Attraction
  • Avoiding Perpetual
  • Avoiding Stalemate
  • Blocking
  • Capturing Defender
  • Clearance
  • Coercion
  • Counting
  • Defensive Move
  • Desperado
  • Discovered Attack
  • Distraction
  • Double Check
  • Exposed King
  • Fork/Double Attack
  • Hanging Piece
  • Interference
  • Mate - Anastasia's
  • Mate - Arabian
  • Mate - Back Rank
  • Mate - Balestra
  • Mate - Blackburne's
  • Mate - Boden's
  • Mate - Damiano's Bishop
  • Mate - Damiano's
  • Mate - Double Bishop
  • Mate - Dovetail
  • Mate - Dovetail - Bishop
  • Mate - Escalator
  • Mate - Epaulette
  • Mate - Greco's
  • Mate - Hook
  • Mate - Kill Box
  • Mate - Lawnmower
  • Mate - Lolli's
  • Mate - Morphy's
  • Mate - Opera
  • Mate - Pawn
  • Mate - Pillsbury's
  • Mate - Railroad
  • Mate - Smother
  • Mate - Suffocation
  • Mate - Swallow's Tail
  • Mate - Triangle
  • Mate - Vukovic
  • Mate Threat
  • Overloading
  • Pin
  • Quiet Move
  • Sacrifice
  • Simplification
  • Skewer
  • Trapped Piece
  • Unpinning
  • Unsound Sacrifice
  • Weak Back Rank
  • X-Ray Attack
  • Zugzwang
  • Zwischenzug

Vision = Mental awareness of something. You know what it is, without the need to go into detail. Can be visual or mental. Vision comes in a lot of flavors. Board vision and tactical vision are the most well known.
  • Board vision = This encompasses the roles and tasks of the pieces, and the squares they are covering.
  • Tactical vision = vision of tactical themes.

Visualization I = seeing the course of the future positions before the minds eye.

Visualization II = the conscious mind tries to communicate with the unconscious mind by means of visualization II.


  1. Under Motif:

    "The function motif and [???] seem to be the most important."

    Not trying to quibble; just wanted to know if "and" was unintended, or if there was another motif that seemed most important.

    Overall, I think a very good summary of definitions. I'll try to comply with them in future comments. Thanks!

    1. Something went wrong with copy and paste. Thx for the correction.

  2. Tactical Themes:

    I would divide them into :
    Tactical weakness , like Exposed King, Advanced Pawn, Hanging Piece...
    Tactical Method , like double attack, blocking, pining,...

    The weakness "exposed king" can be used with the method "double check"
    then tactical vision becomes the skill to spot tactical weaknesses ( by the fusiform face area?? ) and the automated process to find the related method is a motor skill

    To say it in the terminology of the object orientated programming: an instance of the class tactic has attributes and methods ;)

    While to see that a Bb2 attacks a Nf6 and to see that Nf6 has no defender is board vision, both together is tactical vision ("hanging piece") so i dont see no substantial difference between board and tactical vision.


    usually the term combination is used as : "a sequence of forced moves" and the americans like to add: "a combination starts wih a sacrifice"
    If i remember correct Lasker did say that a combination is a sequence of forced moves

    But i like the idea that a combination would be a combination of tactical weaknesses used by a combination of tactical methods resulting in a sequence of forced moves

  3. @Aox:

    A small correction in attribution: GM M. M. Botvinnik in his classic book one hundred selected games provides a small section "What is a "combination"? at the end of the book. He first quotes GM P. Romanovsky from The Middlegam (which accords with GM Lasker's definition):

    "A combination is a variation (or group of variations) in the course of which both sides make forced moves and which ends with an objective advantage for the active side."

    GM Botvinnik takes issue with that definition, providing his own (self-described as "objective") definition:

    "A combination is a forced variation with sacrifice."

    I know we Americans like to take credit for anything and everything whenever possible, but let's give credit to the Patriarch of the Soviet School where deserved.

    Personally, I agree with the working definition provided by Temposchlucker. I don't see any real value for chess improvement in splitting hairs over semantics.

  4. I just did want to point out that Tempos definition of combination is not "common":
    And for people who eventually are interested in chess history, i found this link

  5. @Aox:

    Thanks for the links. Somehow, it reminds me of the story of the 6 blind men and the elephant. (See: Blind men and an elephant)

    It would appear that a "combination" has many shades of description:

    "I don't know what it is, but I know one when I see one!"

    I was NOT trying to provide a precise definition; I was merely trying to provide the background for proper attribution of the inclusion of a sacrifice in the given definition, removing that attribution from "americans." (Not that it really matters for the topic under discussion!)


    BTW, I like your analogy between chess and object oriented programming. I've used many similar analogies in the past between chess and programming. A given tactical concept does indeed include general attributes and methods as a class, which then must be instantiated at the instance level. (Non-programmers, on the other hand, will probably think this particular digression has gone completely off the rails!) I think that is one of the reasons we find it so hard to embed the relevant patterns/processes into our subconscious. We focus our attention on the specifics of a particular problem (instance), in the hopes that automagically the higher-level relevant general "pattern" (class) will embed into our subconscious; somehow, I don't think that will or does happen in most cases. When analyzing, we tend to separate the attributes (data) from the methods (actions based on data), losing the holistic sense that is required to retain the required class structure and functionality: identification of the potential and realization of that potential, in new positions (instances) that bear little or no actual visual resemblance to the pedagogical position(s) (instances) that we have studied. Put another way: it doesn't matter how many individual trees you study intently, you will never "see" the forest until your focus changes. This is why I do not think that solving massive quantities of tactical problems will ever raise the chess playing level significantly, as into or above master class. As you pointed out, the available themes can be subdivided into categories using different criteria. That does not mean that any particular categorization is either "correct" or "incorrect" in some abstract "objective" sense. I tend to agree with the idea that "Whatever works for YOU is a good way for YOU to look at it!" Abstract theoretical concepts about how humans learn are good areas of investigation for cognitive scientists, but pretty much useless for practical chess players. I've gotten some insights regarding teaching protocols from the cognitive science literature, but nothing that would directly benefit my chess playing skill. As GM Rowson put it, "Know HOW is more important for good chess playing than know WHAT." I think that is the most important point of Temposchlucker's "driving a car" analogy.

  6. Robert said "This is why I do not think that solving massive quantities of tactical problems will ever raise the chess playing level significantly, as into or above master class."

    Here we get close to the core of the problem "improvement in chess"

    The average player might improve by .. say 30 elo-points if they improve 100 points in tactics ( measured for example at chesstempo ). A player who is already relativly strong in tactics will only gain a fraction of these 30 points if they improve an other 100 points in tactics.
    This is the principle of "Diminishing returns"
    or ( in other words ) the principle of "A chain is only as strong as its weakest link", with the strategys of a 800 player you cant reach GM titles even if your tactics is "perfect"

    But the improvement situation is even worse: solving massive quantities of tactical problems does not raise the level in tactical problems significantly either. After several thousand problems the further improvement in tactics is virtually 0.

    And here the principle of the weakest link backfires. If we cant reach the level in tactics of a master, then it is probably not possible to become a master either.

    The master with the lowest chesstempo rating in blitz mode (blotto) does have a CT rating of 2115, we dont know if blotto has still the strength of an FM, he/she might be weaker now. But for the sake of simplicity lets say that blotto is the tactical weakest FM ( of the world )
    Then we need to get at least a rating of 2115 in tactics to have a chance to get FM by ourselfs ( this is a simplification !! )

    Temposchlucker has a blitz-rating of ~~1700 so if he cant gain an other 400 points.. he can forget about to become a fide master

    So IF we want to reach the level of an FM we have to be able to improve in tactics to a certain level.. and seemingly even this tiny little subtask of the general improvement task seems to be impossible.

    While on the other hand some kids can reach such a level of tactical skill without ever solving any tactic puzzles ( at CT ) at all!

  7. The learning-how-to-drive-metaphor suggests that those 400 points at CT could reached in 50 hours or so. I mean, if you can learn how to drive in 50 hours . . .
    After all, if even kids can do it, it can't be that hard.

    IF we can find the precise method . . .

    1. THIS observation brings us full circle back to Dr. Lasker's (UNPROVEN) assertion that by following a "good" method (whatever THAT means), in approximately 200 hours, one could raise the playing capability of a young(?) person to the level that no master could give any odds and still expect to win. So it appears that there IS a method (or perhaps a collection of methods) that can achieve the goal. The quest for the "Holy Grail" continues! Salim Belhaddad provides some very intriguing ideas to investigate, along with Aox. The most promising "key" still seems to be: Keep It Simple (for) Stupid (adults)!

  8. you already know "how to drive a car" , you want to get better and drive like "Mr. Hamilton". "No" problem to teach anyone chess in 50 hours.
    Lets look at it in a different way:
    How many students are the best in their class?, how many best in their school? How many best in their town, how many best in their district? I quess that a FM ist equivalent to a "best of the district".
    Do you realy think you can get the best of the district in "what ever you like" in 50 hours?

    1. AoxomoxoA: I do neuroscience research and I've come to develop and algorithm to learn using chess tactics (it's for therapeutics purpose but chess was the best way to achieve it). I've tested it with my 4 years old boy pretty successfully. I'm still analyzing the dataset built from EEG logs and it seems that you have right on a lot of points. Repeating tasks help the brain shift the skills and behaviors from the cortex (that's why beginner have hard time learning chess) to procedural memory (basal ganglia, the emotional/automation system of the brain). It makes you repeat it faster.

      However being in contact with people from and Convekta I noticed that chess tactics training have a problem with chunking. While it's good for broad vision, training from a position based on a real game add far too much sabotaging processes (heuristics) : chunking, filtering, cross knowledge etc. This is a big obstacle I've seen from EEG.

      What you want it to help the "visual memory" store the core concept as a face. By adding too much details you don't let it learn that new kind of face properly. The same happen for westerners (for instance) who don't know a lot of Africans or Asians. They say most of them look the same. Because the brain didn't has the opportunity to learn the now face algorithm from that very macroscopic phenotype.

      When I adapted this with my wife back in 2008, she was able to increase the grade of all her pupils (she is a teacher) in a short time without working more one only on making knowledge more bio-compatible and easier to grasp.

      So a good solution would be to train from endgame tablebases or from easier positions. A higher number of positions than 1 move would be useless for beginners btw. then adding move numbers for a specific motif would only be OK when mixing motives in the same problem.
      Really AoxomoxoA, I think you are making awesome advances. Hold on :)

    2. To store chess information "as a face" is necessary to gain several seemingly important skills of a master. We know from research that Masters can decide if a position is mate or not in far less than 200ms. The speed shows that this "decission-process" was made in the fusiform face area. The fillm about Susam Polgars "brilliant brain" did show that her brain was analysing chessposition with the FFM too. But it seems to be hard to improve exactly here ( for an adult ). My experiments with the "mate in 1" tools was very disapointing and if we cant improve in "Mate in 1" then how can we improve in "Mate in 2" ? Maaaybe the task "Mate in 1" ist still to complex? But i suspect that at a certain age the neuroplasticity is to low to move tasks into the FFM.
      It is possibe for an adult chessplayer to improve the processes past the FFM though. A organised thinking process will ( by stubborn repeating the same thinking over and over agin ) implement new or better motor skills ( to some extend).
      But i think i might try to simplify the task "Mate in 1" in different ways and test if this is "improvable"?
      Hmmm.. saltmines are not my favorite :/

    3. Although the statement is partly with tongue in cheek, I do belief that it is possible to get much better at tactics within a reasonable amount of time. When I analyse my errors at CT, it is actually unbelievable how bad I am. There is an enormous potential for improvement in this area. And if that will be 400 points in 50 hours or 300 points in the next 4 years doesn't really matter. Once I only was able to see simple problems as simple.

  9. Well you have a good point here. Of course mate in one was just an example but actually it's not a good start. Mate in one involves a large amount of of different goal. The only fact of mating is an extra goal that makes you difficult to manage.

    The problem is probably not your brain but the thinking habits you have.
    Hopefully habits can be changed but you need gradual exposure like in Cognitive Behavioral Therapy with the Systematic Desensitization.

    Lets say you are doing a discovery attacks training.
    Imagine the problem you are trying to solve involves another motif like double attack. Let's say a rook behind a bishop is targeting a queen and by moving the bishop you check and profit from this forced move to take the queen. This is a really bad start to learn because too many goals are squeezed and the brain has to give a meaning for every single goal so build heuristics.

    The problem is that we are unaware of most of goals/ you are conscious of the discovered attack, the double check etc but our brain spots far more. So it's important to disassemble the exercise into unique tasks as much as we can.

    Ultimately we want to avoid the bishop to take or attack anything at start. In a first step we want to look at several similar basic positions of a rook behind a bishop attacking a anonymous piece/square/line. Then we want to move the bishop to an arbitrary square in order to attack the queen with the rook.

    I've built such a progressive series of tasks successfully.

    The problem is that it requires a huge amount of chunking work and I had to write a program to adapt the difficulty without me to spent too much time on it.

    In 3 years I've come to an algorithm that can chunk a problem into 20 progressive tasks. Without EEG I would never spot the thresholds and I would project my beliefs on what is better according to best practices from the the chess literature.

    The good thing is that paradoxically, the higher the number of task, the faster the training and the longer the retention lifetime.

    Did you try improving your discovered attacks before discovered checks ?

    BTW is there a way to access your blog please, it really look interesting according to google searches ?

    1. Actually i did not start with Mate in 1, i did start with "Board Vision" ( which is improvable ( to masterlevel )).
      I did close my blog because i had the impression no one is interested anymore. Now its open again.
      In my opinion we have to show that "Mate in one" is improvable ( to master level ) for an adult or we can "forget-it". According to your thoughts and findings we have to reduce the M1-problem further to the point where it become improvable. This would be reducing the number of pieces and/or reducing the size of the board to a 5x5 instead a 8x8 for example.

      So i start thinking about a program which will simplify M1-problems automatically..

    2. Forgotten knowledge is often a treasure. Thanks for re-opening it. I appreciate it.

      To simplify, you may want to script something that will load a position from a set, ten analyze it with a limited amount of nodes and then tell you which pieces are subject to be of utility (for instance if you want a mate in x, a given piece which don't oppose to the mate can be removed). If you build a tree of variations, forced moves are the pieces you definitively want to keep, then pieces that allow counterplay. Then once the simpler version is mastered, for the next rounds, you may want to add the pieces you removed
      Adding noise to have complicated positions will be probably a final step before moving to mate in two or to problem with 2 or more valid variations you have to find and sort.

      I also found that immediate lookup for counterplay will avoid overlooking potential blunders but it's outside the scope of my goal.

      Are you at a master level ?

    3. @Aox:

      I also thank you for re-opening your blog. I have been a "lurker" in the shadows on your blog for a pretty long time, and was very disappointed when I saw your 13 JAN 2016 post, re "Thats it - at least for now". I had the same sad feeling when Temposchlucker stopped blogging so long ago. You both have contributed so much research and time to the investigation of methods for adult chess improvement! I know it MAY seem like it has been a failure because the "Holy Grail" method has not (YET!) been discovered.

      I remind you of this Sherlock Holmes quote:

      "How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?"

      An apropos Thomas Edison quote:

      I have not failed 10,000 times. I have not failed once. I have succeeded in proving that those 10,000 ways will not work. When I have eliminated the ways that will not work, I will find the way that will work.

      A final quote of dialog from Star Trek II: The Wrath of Khan (1982), which embodies my own personal philosophy:

      Saavik: Admiral, may I ask you a question?

      Kirk: What's on your mind, Lieutenant?

      Saavik: The Kobayashi Maru, sir.

      Kirk: Are you asking me if we're playing out that scenario now?

      Saavik: On the test, sir... will you tell me what you did? I would really like to know.

      McCoy: Lieutenant, you are looking at the only Starfleet cadet who ever beat the no-win scenario.

      Saavik: How?

      Kirk: I reprogrammed the simulation so it was possible to rescue the ship.

      Saavik: What?

      David Marcus: He cheated.

      Kirk: I changed the conditions of the test; got a commendation for original thinking. I don't like to lose.

      Saavik: Then you never faced that situation... faced death.


      I thank you ALL and heartily encourage you to continue the search for the WIN-WIN scenario!

    4. ... and to think, I could have started this week without discovering that awesome group of people!

  10. With regard to the simplification of the tasks for EDUCATIONAL PURPOSES:

    There is a chess education blog that explores these issues:

    Momir Radovic's Chess Blog

    Here's a couple of sample links that will give you a good idea of the discussion:

    How to Get Far in Chess: Get the Basics Right

    Superiority of Visual Thinking: Look, Don’t Think! [Where have we "seen" THAT topic?!?]

    (I suspect that you might already have this site bookmarked. For some reason, it was not available for some time, but does appear to be back online.)

    I would think that simple "contact" training (Mr. Radovic's terminology) using no more than 2-3 pieces initially could be set up similar to flash cards but using the computer. A "contact" relationship could be "flashed" to the subject, and then removed, and the subject would have to choose a "contact" relationship (multiple choice) as a recall "test." The speed could be variable, allowing the subject to control the speed in one mode, and the computer to automatically change the speed depending on success/failure in another mode. After a specific number of repetitions of the flash cards, a "real" chess position (utilizing that specific relationship) could then be flashed, with the subject identifying the appropriate "contact" relationship that is being studied by identifying the participating pieces. Obviously, this could be elaborated in many different ways.

    I know that sounds extremely simplistic, and most chess players do not want to do this kind of fundamental drill very much. However, I think it has significant potential for improving "face recognition" much more than trying to solve an infinite number of much more complicated tactical problems.


    Are you willing to divulge the 20 progressive tasks? Is there any literature that you have written about these tasks and the program you have written?

    Thank you for participating!

    1. Momir Radovic and i did exchange a few ideas some while ago. His Contacts ( which are the "Contacts" of Awerbach in his Middlegame book ) are more or less identical to the "board vision" of Dan Heisman. At my blog i have a set of links to board vision training programs, a few of them have been written by me.

    2. Would you mind elaborating on what strategy you use for board vision ?

    3. At my blog you find links to Board Vision Exercises which are often of the type: find all pieces, which are attacked , find all pieces, which can give check and so on.

      But i think that several of the visualisaton exercises should help to improve board vision too ( for example ).

      But now i think that the best method to implement "vision" is to implement a thinking process , may it be as simple as a mantra . Example: we may ask at every position: "which pieces might be forked any soon?" Doing this for several weeks at any game we play, at any position we analyse,... we will develop a good fork vision.

    4. One additional reference for the Contacts Method:


      There also is an article written by GM Aron Nimzovich in that PDF file, titled The First Chess Lesson, from “How I Became a Grandmaster”.

      You will have to scroll down through the PDF file to get to Mr. Radovic's two articles.

      It may seem hopelessly "simple" to start with basic contacts, but I think there is considerable merit in beginning AGAIN at the beginning. Many chess books that I have read all emphasize that the primary reason we amateurs (particularly adults) fail to progress beyond a certain (amateurish) level of chess skill is because we have rotten foundations. If that is true (and I have every reason to believe that it is), then let us "blow up" that crappy foundation and begin again from the beginning with "Beginner's Mind."

      That "Beginner's Mind" conveys a Zen concept that I think is very important for chess improvement. If we are stuck in our existing thought processes and assumptions, then we will NEVER "see" how to improve nor how to change our thinking! We need to "think outside the box" of our current thinking paradigm (except when we are trying to keep the opponent's King INSIDE the "box" to checkmate him).

  11. Robert,
    Thanks for sharing, the site is awesome. The process looks like Convekta's CT-ART software. When you fail at answering it shows you a 5x5 board with a minimalist version of the challenge. My approach is somewhat different however.

    As for the speed it's of a relative importance. Calibration have to be done first with something like a set of initial measures (reaction time, cognitive process etc. in form of a short IAT survey -

    a 40yo person would find a solution three times longer (or progress slower) than a younger person but both still being measured as good because initial (and continuous adjustment) calibration would have been done. For personal non academic training there a software like Anki or Mnemosyne which could do the trick.

    Anyway, speed is kind of an unstable external measure. EEG help me to "see" and sometimes predict what kind of idea is being built before the subject actually is aware he/she is thinking about it. I get better metrics and I can calibrate the variable almost in real time to select the next task.

    Some problems can require less than 20 tasks and some more, but the minimal number of tasks is not less than 5-10 even for the most basic position. I will need to make more research with more subject (I only have 4 now) to be able to define a more universal sequence of tasks.

    The research is still in progress and draft state as I fund it with my own money. As soon as when I complete the milestone, I will be happy to share results, several examples of progressive tasks and some Python code, or hopefully Javascript so that it can run directly from the browser.

    However let me give you an idea. Let's say that the common base position counts 2 or 3 pieces. According to the discovered attack example, you may want to start with a task that consists in displaying a series of position with rooks hidden by a bishop + a few random positions. The positions are rotated and you have to click YES if the position is a rook behind a bishop.

    When you are able to recognize the position from at an angle of 90/180/270° you can increase difficulty with more complex positions by adding more pieces (eg an attacked piece, or "decorating" different pieces like pawns, knights etc.) or changing colors (B+B, W+B, W+B).

    Speed increase and pseudo randomness are variables which depend on continuous calibration.

    I tried hundreds of different algorithms but one of them is very efficient for most of tactical motifs - as long as they do not involve sacrifices for reasons linked to cognitive biases and probably emotional biases. My apparatus is not as flexible as I would to measure the limb system efficiently, I will invest as soon as possible :)

    Rotation is probably one of the 10 most effective variables in the algorithm.
    Speed is one of the least.