Thursday, February 25, 2016

Order in the salt mine

There are of course a whole bunch of subtasks we can think of. Subtasks of the 1st order, the 2nd order etcetera. But can we get some order in the chaos? I will give it a try.

Tactics consist for an important part of the combination of 4 elements and 4 actions.

  • Attacker
  • Attacking square
  • Target
  • Defender

  • Attack
  • Restrain
  • Block
  • Protect

Sorry Dan, I'm not ready to use your terminology.
The subtasks of the first order are related to the direct attack.

Direct attack
 The following subtasks spring to mind:
  • Identify the attackers
  • Identify the targets
  • Identify the pieces that can block the attack
  • Identify the escape squares
  • Identify the protectors (of the target)
All these subtasks are directly related to the attacker and the target. So I categorize them in the subtasks of the first order. I don't know beforehand whether such categorisation will proof to be useful, but we will see about that.
The subtasks of the second order are related to the defenders and the escape squares.
  • Identify the attackers of the defenders (the blockers, the protectors, the defender becomes the new target)
  • Identify the attackers of the escape squares (restrain)

The next step is to look at the indirect attack.
Indirect attack
This introduces the attacking square. New subtasks:
  • Identify the attacking squares
  • Identify the blockers between the attacker and the attacking square
  • Identify the protectors of the attacking squares.
Maybe we should call this the subtasks of the second order too?
Subtasks of the third order:
  • Identify the attackers of the defenders of the attacking square (the blockers, the protectors, the defender becomes the new target)

Are there other skills that have no relation to this classification? For instance, Troyis is meant to train such unrelated skill (knight manipulation in a confined space).


  1. All the mentioned skills in the post start with the word "identify". When I'm salt mining, that is the continuous challenge: to recognize what is going on, on the board. As both Aox and Tomasz mentioned, sometimes you even have to look 10 seconds before you even know where the black king is.

    When I do blitz tactics at CT, I'm continuously wondering why it takes me 30-50 seconds before I notice that a piece is overworked, or a square is protected from a distance, or a piece is pinned and the like.

    Aox mentioned that good chessplayers have "hacked" the fusiform face area (FFA) for recognition of important chess characteristics. I think it is good that we look at skills this way. We don't acquire the skill to mate in one, but we learn to recognize the features of the position instantly, and the result will be the mating move. Just to put everything in the right perspective.

  2. @Temposchlucker: "When I do blitz tactics at CT, I'm continuously wondering why it takes me 30-50 seconds before I notice that a piece is overworked, or a square is protected from a distance, or a piece is pinned and the like."

    The reason is that a "surface" impression of the relative importance of a specific feature (a certain combination/constellation of pieces that is familiar to YOU) grabs your attention initially, and then you have to work to free your intuition in order to grasp the "essence" of the position. Given the number of pieces and the potential interactions between them (attack, retrain, block, protect), it is natural that your attention will be "hijacked" by the subset of pieces that YOU perceive as important on "first impressions." Obviously, that will vary from individual to individual, based on the totality of the individual's prior experience.

    There is a fascinating investigation of the roles of analogy and categorization in Douglas Hostadter and Emmanuel Sander:

    Surfaces and Essences: Analogy as the Fuel and Fire of Thinking

    1. In this post I describe how I bide my time in those 30-50 seconds: Essentially, I'm trialling&erroring cluelessly all the time. The thought process that I need here will be geared around the pictures of the direct and the indirect attack, as shown in my post. The 11 identification exercises I propose in my post, are at the very base of this thought process.

      Imagining to following: all these subtask usually take 30 seconds to pop up = 2 clicks per minute. Which is quite normal for a complex subtask. If I could only let them pop up in 3 seconds (= 20 cpm), I would be 10 times faster!

    2. While I don't disagree with your categorization of the subtasks, I do think that it may be more fruitful to prioritize them into a hierarchy of sorts using your inherent chess knowledge. In essence, I am suggesting a knowledge-directed search rather than a brute-force approach. I assume you do that already.

      I'll use the given position as an example, recognizing that it is extremely simple so that it illustrates your points without obscurity.

      We "know" that an attack on the King is a higher priority than anything else related to material gain. My usual first question upon seeing a position is to ask myself: is the King safe? If not, then everything else takes a back seat to figuring out how the King is unsafe, and what possible measures can be taken to make it safe (if it is my King) or to make it even more unsafe (if it is the opponent's King). In the given position, neither King is unsafe, and cannot be made unsafe in the foreseeable future. This requires a quick glance at both Kings and the available material (pieces, not Pawns) on the board.

      The next question is: who has the move? A mere glance at the position given shows an asymmetrical relationship between the White Bishop and the Black Knight. The subtask "Identify the attacker(s)" is merged with the subtask "Identify the target(s)" into a single perception of an "Attacker/Attacked" relationship. Since it IS asymmetrical, the "who has the move" question becomes relevant for understanding that relationship. If White is moving, 1. BxN ends the investigation, simply because there are no other potential captures/recaptures.

      In order to continue the investigation, we have to assume that Black is to move. That brings into play the next three order-1 subtasks. "Identify the pieces that can block the attack" is immediately eliminated; there are no Black pieces with which to block the specific attack under investigation. "Identify the protectors (of the target)" is also immediately eliminated; there are no Black pieces with which to protect the Knight. By process of elimination, that leaves the final order-1 subtask "Identify the escape squares."

      Now we focus on the safety of the available squares to which the Knight can move. In the given position, that is simple: any one of the 8 available squares is "safe." Additionally, there are two squares (e4; d5) which turn the tables, allowing the "Attacker/Attacked" relationship to be inverted.

      I think that there is a fuzzy hierarchy of priorities involved, that runs concurrently with the identification of the relevant subtasks vis-a-vis the various relationships either existing on the board in the given position or the potential for existence inherent in the given position. That hierarchy is coarsely dependent on the specific internalized chess knowledge of various "goals" such as mate or material gain. In other words, that internalized knowledge has to "bubble up" to the conscious surface whenever the appropriate "triggers" exist (however deeply hidden) in a given position.

      Yeah, I know: that's what all of this investigation is about!

    3. Your story let me longing back to the development of a balanced slick thought process all the more (working on the K-factor, according to Aox, if I'm not mistaken). I detest the salt mines, and I hate it that I cannot invalidate the arguments with their ugly logic behind it. So I have to settle this vision stuff once and for all first.

    4. Given the considerable difference that exists between various people as they progress (or not) at chess, I firmly believe that there is no such thing as a universal "balanced slick" thought process that can be used by all players effectively. I find that my own thought process has changed as I've learned more over the years. Sometimes what I have learned has been directly chess related; sometimes it is something that is far away from chess. Insight is gained by a process of making analogies between whatever it is that I am studying currently and what I have already learned in the past. Apparently that is a process that we all utilize to make sense of the world, but we don't often think about our thinking processes (i.e., metacognition). I think that's why it is so hard to get a handle on it. It is so easy to slip into and out of metacognition that we don't even realize (most of the time) that we are doing it. If only it was as easy to slip into and out of "chess expertise!"

      I eagerly "look forward" to "seeing" the outcome of the "vision" investigation using the subtask tools.

  3. An interesting experiment would be to modify the various tools to provide a subliminal "snapshot" of the correct answer(s) for a very short time (less than one second; perhaps a period that can be controlled by the user within a certain narrow range of values), to see if that would provide sufficient focus on the appropriate features of the given position. Perhaps that is an entry path into the FFA for the desired "vision" to be trained. The subtasks would remain the same: find all X in a given position. The difference would be that (subliminally) there would be a certain amount of "training" (or conditioning or "bias" or whatever you want to call it) focused toward the solution but without any verbalization or logical justification. I think that after a certain number of exercises, the subliminal "training wheels" could then be removed to see if the training was successful. I think this would train the subconscious toward seeing the relevant piece constellations/relationships (across the entire board). Since the subconscious is the actual target of this training (not the logical reasoning conscious), I think it worth exploring. Unfortunately, I do not have the necessary programming expertise in Javascript to make the changes, nor the time to learn how to do it. So, consider it a "food for thought" experiment.

    Another suggestion for modification: provide a capability that (similar to Chess Tactics Server) allows the user to select the capability to "break" (effectively, pause after a problem is "solved" correctly or incorrectly) on failure, success or both. This allows for a period of reflection on a particular problem, and consolidation of the "lesson" of each problem. This will necessitate a programming change to the timing functions, so that only the time from initial display until final solution is measured. I'm betting that someone (perhaps several someones!) here knows how to do THAT!

    1. We certainly haven't found the best way to hack our FFA. Maybe we haven't even found a bad way yet. I know nothing about the FFA, so I will definitely have a look at your suggestions. More accessible knowledge about the FFA is needed.

      From time to time I do the exercises in slow motion, just to be able to reflect on the positions.

    2. FFA of Susan Polgar at 41:00
      But i dont think its so important where and how the brain improves.
      Fact is that the performance in tactics is a matter of speed. A 1600 spending 5 hours at a tactical puzzle performes as good as a master spending 3 min at the same puzzle, and a GM playes better 1-min bullet then me in a 5 hour tournament game.
      Many researcher think that the fast skills are the decisiv skills in chess
      ( )
      The question is.. how to improve them?
      What i found is: there are chess related exercises where we can improve , like the traditioanl "board vision" exercises.. and there are exrecises where we cant improve ( that much, that quick ) like m1.
      The difference of these task is their complexity.. as more easy as lower the level.. as more easy to improve.
      So i think we should try to make things easy somehow, for example by improving the/some subtasks. Deeper understanding could help too of course, but i found no book which could tell me something new about m1 lately.

      So i dont care if something does change in the FFA or somewhere else.. I know what should get better: my MPM
      If i cant improve here.. i can forget about getting "decisiv" better.

    3. @ Aox: Thank you for the references. I had seen the NatGeo video on Susan Polgar before. I had not seen the Burns paper previously.

      As you correctly note:

      Many researchers think that the fast skills are the decisive skills in chess. The question is.. how to improve them?

      As Burns noted, beyond a certain skill level (perhaps 1600-2000 USCF?), the search process for a good move tends to become "standardized" (in some sense) among all players, and no longer differentiates playing skill. There is very strong correlation (0.78-0.90) between skill at blitz and non-blitz, with "up to 81% of the performance variance in nonblitz chess was shared with blitz chess, which strongly suggests that variance in the effectiveness of fast processes (such as recognition) accounts for most of the variance in chess skill."

      It becomes obvious that the "solution" for improved chess skill lies primarily in developing the "fast processes" such as pattern recognition. I note that the "fast processes" (pattern recognition) occur quickly, at a more or less subconscious level.

      The current vision exercises are predicated on gaining skill through rapid exposure and repetition (massed practice). In essence, the assumption is that it is massive amounts of fast (high speed) training that will improve the fast processes (recognition). So far, I am not convinced that this assumption is correct.

      I will stipulate that speed training (especially at higher and higher speeds) WILL provide short-term improvements in those specific subtasks. It is an open question as to (1) whether that gain can be retained long-term, and (2) whether that training will produce a corresponding increase in chess skill.

      A question that has not been addressed (to my limited knowledge) is this: are there effective training exercises that are NOT "high speed" that can positively impact the fast processes (recognition)?

      In short, is high speed massed practice the only way to increase the fast processes (recognition), or can the fast processes be improved significantly by slower training methods?

      My personal experience leads me to believe that fast (high speed) training does not lead to fast processes (recognition) with subsequent increase in chess skill.

      But, what do I know?!?

      A most cogent question is:

      What training methods do the chess expert trainers recommend?

      I have seen no recommendations regarding high speed subtask training tools. Perhaps that is because such tools did not exist, although flash cards have been around for a long time.

      On the other hand, maybe high speed training IS the only way to significantly improve chess skill in the modern computer age.

      We should soon have a tentative answer on fast subtask training from the salt mines, at least with regards to the Mate-in-1 skill.

      As for the FFA: I also do not care where in the brain the skill resides, as long as it IS there and it is ACCESSIBLE when required. The "hijacking" of the FFA for chess skills is only of interest if a method can be developed that will cause the "hijacking" to occur for chess. So far, (other than Papa Polgar's methods), I have not seen much of relevance in this regard.

    4. It is not so easy to apply a slow method for mate in one. Because, well, it's just a mate in one. You can stare for 15 minutes at a mate in one, but hey, after you solved it in 20 seconds, where are you staring at? It doesn't make much sense to stare any longer at the position. And before you know it, you forget to stare.

    5. @Tempo
      i think that is the diffeence between the future GM and us so far, they think a lot about a mate in 1, they use every position thanks to a thinking process as board vision exercise. They pay a lot more attantion to fin, forks, attackers asf then we do. We simulate the lack of attention we did pay to the tiny little aspects of a position by high quantitis of clicks now.

      @Coble Nigel Short says that adult chessplayers reach usually very fast a ceiling in tactics and should focus on other things.. like pawnbreaks..

      WE try to break that ceiling, we invent a new method.. or .. we dont get decisiv better

      And you misunderstand the role of speed, we need to become fast, but not to make a fast training AND the tasks we perform are extreme simple. There is no need to check if a piece is attacked for several minutes.
      We try to make high quantitys to develop a skill. Skills remain, what we forget is knowledge. You dont need to swim or drive a bike for centuries.. you will still be able to perform quite well after a short initial moment.
      See tempos experience with troys. The ability to handle Knights did stay alive for years

    6. PART I:
      A small excerpt from Charles Hertan’s Forcing Chess Moves: The Key to Better Calculation, pg. 14:
      My peak FIDE rating was over 2400, yet I am not ashamed to admit that some particularly devilish mate-in-two problems have stumped me for as long as an hour!
      In light of the supposed superiority of the master class in solving problems of mate, that seems astonishing for more than one reason.
      (1) Given the skill of an International Master level player, why on earth would someone who has internalized 10,000 to 300,000 positions NOT be able to solve ANY mate-in-two problem almost instantly? Doesn’t Master Hertan “know” that working the salt mines (100,000 problems or more) is the pathway to success and a higher level of skill (reflected in higher rating)?
      (2) Why on earth would someone spend an HOUR working on a relatively simple mate-in-two problem? Why not just look up the answer after a few seconds of attempting to solve the problem and then continue on to the next problem? After all, it is the QUANTITY of positions observed and stored that counts, not the QUALITY and number of “lessons” extracted from any one problem, right?.
      I think there is a significant clue in the description of a problem as “particularly devilish.” My OPINION is that M-1-H is HARD because of the “particularly devilish” nature of those problems, compared to the relative simplicity of M-1-E problems. The M-1-H problems (usually) do NOT reflect typical chess positions that are encountered in chess games. The configuration of many of these problems simply could not and would not occur in any normal chess game. Consequently, the domain specific “expertise” is effectively nullified, requiring a prolonged look at all of the insane interactions in order to find the needle in the haystack solution. Dr. Lasker addressed this issue in his Manual of Chess, and recommends against using this kind of position for training chess skill. (I note that not as an appeal to authority, but as a pointer to wisdom regarding chess skill training.)

    7. PART II:

      My training in martial arts (and in other fields) began with SIMPLE things done S-L-O-W-L-Y and consciously, until I could subconsciously perform those simple things at ever increasing speeds. I experienced the slowness of Tai Chi Chuan exercises, simple movements done at a glacial pace. I have then observed a trained Tai Chi martial artist respond at lightning speed to a formidable attack. I observed the "speed" training of other martial arts, and the eventual skills demonstrated by both methods of training. The "speed" advocates have no use for simple, slow exercises; they want to get to the "good stuff" of bopping somebody else with a punch or kick in sparring. The interesting thing is that the "slow" speed students do take longer to get to a particular belt level (correspondence to chess rating?). On the other hand, there appears to be no upper bound on their eventual skill levels. The "fast" students often do not grasp the fundamentals (simple stuff), plateau at a fairly low level, and often just quit because they cannot progress. A foundation built on sand will not stand against a storm. I don’t think I have to explicitly lay out the correspondence to chess training.

      IMHO, M-1-H is HARD not because we do not have the requisite patterns internalized and available. It is hard because it DELIBERATELY (by intentional design) exceeds the complexity of relationships (attack, restrain, block, protect) that could ever conceivably exist in a normal chess game. Consequently, even a Master has difficulties with recognizing all of the (too many) relationships in that type of problem, not because no patterns are recognized but because too many chess specific patterns are triggered.

      De Groot found that randomly placed pieces were sufficient to almost eliminate chess pattern recognition as a skill, reducing the experts to the amateur level. (Subsequent research showed that there is still a residue of chess pattern recognition available to the experts but not a significant amount.) I submit that the obfuscation of M-1-H problems DELIBERATELY exceeds the pattern recognition skills of almost all human beings. As a consequence, I do not consider that type of problem to be useful for training pattern recognition, regardless of the speed. YMMV, of course.

      I have seen the same problem when working with pilots in a flight simulator, with the goal of determining the best clustering of instrument panels to allow the trained pilot to scan and interpret the information about the state of the airplane. After the instrument panel reaches a certain level of complexity, requiring eye scans across an ever widening field of vision (perhaps not in an optimum configuration to allow the critical information to reside close together), the pilot inevitably begins to lose the thread of what is happening, and they lose control of the plane. The problem is not the pilot but the complexity of the problem and the presentation of the information required to solve the problem.

    8. Papa Polgar used to put his daughters on a diet of composed chess problems, so it can't be too bad.
      His daughters had no innate talent for chess, so they come closest to us. Besides the fact that they were young and we are not so young.

      The subtasks we focus on, are very common subtasks which will be useful for all kinds of tactics. The composed problems just provide them in a concentrated way. If we need an unconcentrated way, we will soon find out.

  4. There is the need to analyse the position and the characteristic (features - its elements). However when the element is TOO simple - it does not make any sense to slow down "more than necessary". The best approach will be to think over the problems and the methods (ways) how to overcome it :). That's my point of view!

    1. @ Tomasz: I am reminded of Einstein's dictum: "Everything should be made as simple as possible, but not simpler."

      The fundamental training question IS simple: What is the simplest and fastest method of training chess pattern recognition and therefore chess skills?

      We are engaged in a process of discovering at least some answers to that question.

      I will accept your training suggestion at face value: "The best approach will be to think over the problems and the methods (ways) how to overcome it :)."

      As resident curmudgeon, I respectfully suggest that little "thinking" is going to occur when attempting to increase the speed at "solving" various problems - and that IS a problem, IMHO. Obviously, I hold a distinctly minority opinion on the subject of fast training for fast results. I am not offended nor do I have any intention of offending anyone holding a contrary view.

      Time will tell if the current approach will yield the Holy Grail of permanently increased chess skill. If the empirical results are obtained, then I will be the first to congratulate the Knight Errants who have labored so long in the salt mines.

    2. There is no need to bungle around with opinions about things I cannot possibly know beforehand. I opt for the simple way: I'm going to find out. Later will be soon enough to form an opinion.

  5. @Robert
    If you know a method how to improve in tactic.. then demonstrate it. Show us how you get decisiv better in tactics, we will apply your methods.
    I am not shure if we will have success with this method in real tactics. But we make experiments and try to find out if it works and how.
    Our main goal is to improve decisiv and not to discuss intelectual.
    If you have the Philosopher's stone.. demonstrate it.
    We dont have it.. but at least we make some progress , empirical proven.

    1. Hmmm, a "put up or shut up" challenge.

      The most significant (recent) method I suggested was to stop trying to arbitrarily locate and apply tactical themes/devices to a specific position, with no idea of any goal. Instead, (prior to "looking" for tactical themes/devices) look for motifs ("motives") as an indicator that one or more corresponding tactical themes/devices might exist which lead (if applied properly) to an advantage. I gave two of the most important motifs, the "function" motif and the "geometrical" motif. I made no claim for originating the concept; I gave the reference sources so that others can explore the context and application of those ideas. I recall that Temposchlucker "saw" the value of the concept and began to "look" at positions using those two motifs in particular. Q. E. D.

      Hopefully, I am still allowed to observe; I will refrain from commenting further. My apology for interrupting the experiment in progress.

    2. We have working on the K-factor, and working in the salt mines. The first is aiming towards understanding, developing a thought process etcetera, while the second is aiming at improvement of board vision. I sense that you use arguments that are applicable to working on the K-factor, for reasoning about the salt mines. That doesn't seem logical to me. But maybe I don't understand you well.

      Maybe you remember that I complained about "triggers that are not sensitive enough" while looking around the board, during solving tactics at CT. We need an inner "Robert Coble-module" that shout at us "Qf4+!!" at the right moment. There are two lines of work necessary, as far as I can see, to develop such module. The first approach is initiated by the conscious thinking mind AKA working on the K-factor, the second approach is initiated by the unconscious mind, which has become more sensitive for very low level elements in the position like "this square is protected", "this is an escape square", "this attacker is pinned".

      The salt mines already give me a better sense of which square is an escape square for the king. An unconscious knowing, a "gut feeling", about where the king can go safely, and that pops up even before I have seen the details of the position. That's what we are looking for.

      In an ideal world, the K-factor melts together with the board vision, and together they find the best move in no time. But the methods of the K-factor don't apply to board vision.

      Is at least our hypothesis. Which we try to falsify or proof in a rabid manner. (Can you actually proof an hypothesis? Never mind, you get the idea).

      After the salt mines, that is, after having found a method how to improve board vision and, hopefully, having improved it, I continue to work on the K-factor, which I only consider temporary postponed. The more rabid we work, the sooner we can go back to the grassy pastures of the K-factor.

    3. @Robert

      "Hmmm, a "put up or shut up" challenge." - it does not have to be this way. You can participate in our tests, exchange ideas and comment. However it would be a REALLY helpful if you could verify our methods and efforts by some kind of additional conclusions or excercises.

      I know it may look strange that we are "mining salt" close to MDLM style. Anyway I see some light in the tunnel. It is not a BIG light, but the visible one.

      I recommend (suggest) you DO NOT refrain from commenting further. Just try not to "force" your ideas and everything will be ok. At least it is my point of view.

      Your comments are GREATLY appreciated. I love reading them even if I could not proove them. Anyway they give me the inspiration to see the problems we are working on - from the other perspective.

      And back to your main idea. There is NO need to think over and analyse deeply the reason you miss the Knight attack on the King - if there are ONLY two pieces left on the board. You can think of it for a while, find a method to fix and... and TRY it on the board (in practice).

      What else? I have recently found out that most games against strong amateurs (1900-2200) finish in a very simple scenario: one of the player makes BIG mistake (in comparison to his strenght) because he is in a time trouble and cannot see or mix the idea (or move order). And that's the reason there have to be some correlation with the speed of recognition and fast process of info. If you would have to solve 40 math tasks and you will be given just 120 minutes... the chance you could solve them all correct is NOT high - unless you learn the simple mathematics (addition, division, multiplication, etc.) really efficient (fast). Of course at the NEXT level you have to optimize the equations and try to improve each step at the way to reach the final goal. Anyway it is the NEXT step, not the first one.

      In a drastic way: if you HAVE to spend 30 seconds to RECOGNIZE pieces (due to their weird shape or your screen very dirty) it is not any good if you could solve the puzzle withing 5 seconds. Why? Because if you clean the monitor and you will be able to recognize all the pieces in 2 seconds... your overall (total) time is just 7 seconds! Compare the first case (35sec) with the last one (7sec). The difference is HUGE. And if you would like to optimize this result (7sec) you can do it AFTER you shortened the first phase (piece recognition) to the minimum. If you recognize pieces in 2 seconds and improve the second part to 1 second - it is in total 3 seconds ONLY!

      Compare the following case studies:
      1) BAD phase: 30sec + GOOD part: 5 sec = 35 sec
      2) GOOD phase: 2sec + GOOD part: 5 sec = 5 sec
      3) GOOD phase: 2sec + VERY GOOD part: 1 sec = 3 sec

      Now the comparison: ORIGINAL (inital) time - 35 seconds

      1st case: 5 seconds. 7 times faster
      2nd case: 3 seconds. 10 times faster.

      Imagine you could find a way to do it - and there is NO difference between you and any IM or GM at blitz. The only difference would be: knowledge, skills and other elements (nervous system, motivation, perseverance, dedication, healt, etc.).

      You can comment my vision and show any refutation (wrong points) if you wish. As much as I understood this is how our friends: Aox and Tempo think about it (with minor differences, but the main idea remains).

    4. @Tomasz: Thank you for your kind words. I am not offended by anything that has been written by anyone. I am responding to you so that (hopefully) there will no misunderstanding about why I am not going to comment any further. A grand experiment in "salt mining" is underway. I do not want to distract from that experiment with irrelevant comments from a "non-laborer." I have nothing constructive to add to the experiment at this time. Consequently, I will not comment any more until the experiment is concluded and the results are summarized. I will continue to observe from my glass house. ;-)

      I think AoxomoxoA's "challenge" for ME to come up with something that works for improving chess skills is interesting, to say the least. I am not conversant with Javascript, and I do not have the time (nor inclination) to become conversant with it. I retired from the programming wars 12 years ago. I also have family issues that must be dealt with immediately: a 7-year old granddaughter who was just diagnosed with Double Cortex Syndrome, whose father just lost his job and health insurance. [I AM NOT LOOKING FOR SYMPATHY; just stating the facts of life.] I cannot work in the "salt mines" at present.

      However, I will be exploring (time permitting) the idea of a program to evaluate the difficulty level of mate-in-1 problems in terms of interactions (attack, retrain, block, protect) between pieces. I'm curious to see if there is a significant difference between the number of actions in M-1-E and M-1-H. If I do write such a program, it will not be in Javascript, so I don't know if it would be useful for digging in the "salt mines" through the Web. Perhaps it could be used to analyze and rate various mate-in-1 problems, and provide a (more objective?) distinction between the two categories. It MIGHT shed some light on WHY M-1-H are "hard," even for the experts.

      I also have some other (half-baked?) ideas for part-task chess skill training programs, but will not discuss them at this point. I have no ingredients available for baking "pie in the sky."

      Good luck to EVERYONE!

    5. Oops! I note a typo in the preceding post: the various interactions should read "(attack, reStrain, block, protect)." Only those of us who desire to improve should be "retrained" (whatever that may mean).

    6. @Robert, your comments are always highly appreciated by me. The second goal of this blog always has been to become better at English, and your erudite and eloquent linguistic usage is of great help. Besides that, I tend to get carried away by euphoria when starting a new experiment, but at the same time I'm very sensitive for critical background humming, and on a regular base I calibrate my euphoria by it. So please keep humming, if life permits.

  6. There are 2 alternatives

    -We try to develop a method where we already have a proof that it does work.. at special cases.

    -Or we do the same we do since years without any improvement.

    If someone does have a better method.. show it. Most likely we already did try it and did not improve with it.

    But theoretical discussions without any empirical background..are not interesting as long as we HAVE a promisiong idea with some small but positive results.
    As soon as we see the idea does not work.. we may search for a next one. But to stop now as we see some light at the end of the tunnel of several years of non improvement..just because other nonimprovers said this or believe that..

    We will see if this idea can work within a few months.
    We will not see if it can work if we stop now.

    So why does anyone want us to stop?

    There are billions of people who know it better but we need 1 person who DOES it better.

    Robert you are programmer by yourself. Generate exercises which are better than ours
    DO it better.