Tuesday, June 14, 2016

Law of conservation of misery

In my previous post, you could read that I was going to focus on the vultures view (offence) and the standard defences against whatever attack the vultures view might come up with.
Every time I fix my failures, the used time grows. Due to my new focus, I make less errors, but at the expense of using way more time.

Although I have the feeling that my approach works, it can't be denied that it is going to take an awful lot of time to make some serious rating progress. Somehow I absorb the newly discovered patterns too slow. I suspect that the reason for that is, that the frequency of occurrence of the same pattern is too low. There are quite a lot of patterns to absorb, but it takes often days or weeks before you encounter a certain pattern again, and that is not an efficient way to learn.

I selected 100 failures. Since the amount of new patterns is vast yet finite, I'm going to analyze these 100 puzzles more deeply. With the emphasis on categorization of the patterns. Somehow I must recognize these patterns vaster.


  1. If you spend twice as much time on a problem your performance will increase by 100-200 Elopoints ( depending on your k ).
    On the other hand the mayority of the puzzles you get at chesstempo are duplicates now, difficult to tell how much you would be rated on problems you did not see before.
    As a premium member you can filter your attempts and compare your performance on duplicates and non-duplicates.

    As older we get, as more repetitions ( and in shorter time ) we need to learn something. I did try to copy positions into Anki and learn them by spaced repetition but my impression was that the effects have been negaitive for me.

    The most important skill in chess/tactics seems to be the "speed of patternrecognition". As more pattern you accumulate as more time it takes to recognise the right one. As we did see at the M1 problem, that is a difficult task.

    1. I said in my post of may 30:

      These 18 themes can combine with each other in 18 x 18 = 324 ways. In practice, not every combination occurs equally frequent. A thorough knowledge of, say, the 50 most frequent combinations, should make a lot of difference.

      So far, I have identified 25 combinations with each 1-4 examples. When I gather more examples, I will end up somewhere between 50 and 324. I want, say, 8 examples of every combination. Studying of 8 examples per combination will hopefully create some extra cues that help me to speed up the retrieval of the pattern. That is the idea. . .

    2. I did use ANKI too for quite some time.
      The use of ANKI encourages the memorization of both the position and the move sequences. That are the wrong patterns, since only one pawn difference can change the meaning of every piece on the board. We need functional patterns. These functional patterns must be discovered first. ANKI can be used while skipping this necessary step.

      I use Freemind, which is an ideal tool to toss concepts around. It is very easy to link those concepts to chess positions at CT.

    3. GM Smirnov is suggesting to use Mind Mapping software to improve in chess. I will have a look at freemind, thanks for the hint

  2. The tags at chesstempo are bad.. but the problem search still might help you.
    If you are looking for Tactic t1 in combination of Tactic t2 it would help to search for t1 (or t2) and then step through these puzzles to find the right ones. If this combination is statisticaly "significant" you should find "quick" "several" examples

  3. To gain the positions is not the problem. In one evening I can make 25 errors easy ;)

    The time consuming part is to analize the positions and to subsume them in a consistent model of a combination that is simple, clear and practical, with as less redundancy as possible. Doing that work myself is necessary, since it is part of the learning process. Besides that, I'm no longer a premium member.

  4. When I was young I have been solving a lot of puzzles (tactics/combinations). I tried to focus on the motifs (themes) and use these to solve the puzzles. After I solved about 800-1000 tactical chess positions - I could see much more key (crucial) tactical points. I am not sure how many positions (rated 1500-2000) I have solved so far, but it is probably close to 2-3 thousand.

    What I found out is the same as Tempo's findings. Some of the tactical motifs occurs very frequent and a few of these are VERY rare ones. That's why I would advice to focus on the most frequent motifs and learn these very well.

    If there really are 324 ways (18x18) of themes combinations... I could set up the limit at about 180-220. You can check out the contents at the chess tactics (combinations) manuals/workbooks - they mostly contain about 10-12 chess motifs, not all of the possible ones.

    If we could help you at the process of finding valuable things (for example: some ideas refutation) what would they be? Could you list the topics and specific questions you want to test (and find out if they works or not). I do not mean doing the work in your behalf, but instead of it - try to see if the same work give the same results (not to mention curiosity how could I feel doing some similar things).

    1. Some tactical motifs manifest themselves in the most simple form. For instance a double attack against two pieces. The targets aren't always two pieces though. It can be a piece plus a build up for a discovered attack, for instance. Or the targets are a piece and a pin.

      A combination is chaining main tactical motifs together (duple attack, trap, promotion). In the examples above, the main motifs are stitched together directly. But often, the connection is indirect. The main motifs are glued together. Less is known about this glue. The geometrical patterns that are accompanying the glue are less well known. One thing is known though: the connecting move must be a subset of CCT. Glue can be for instance: coercion, clearance, sacrifice, deflection, attraction, zwischenzug, capture, etcetera.

      If you want to help, you can do so by focussing on captures. A capture often is the glue between motifs. There can be all sorts of reasons to capture a piece. It can be the final execution of the combination, a clearance move, capture an attacker, capture a defender, exchange a problem piece, get the right target in place, deflection of an overworked defender etcetera. I'm interested in all the different reasons to capture.

      You can help by providing me a list of problem-ID's from CT, with captures and every different reason for the capture you can think of. 1-4 examples per reason is enough.

  5. Great idea! I will try to do it soon as I will have more free time (or to be exact: I will not be overloaded by the amount of work).

    Building the list with all the captures motifs is an interesting idea! It will be a good reason to refresh my mind with some hundreds of puzzles related to tactics and captures! :)

    I am curious If I will be able to discover some new reasons (ideas) behing all the captures motifs presented in chess literature.

    What do you think about writing the article related to the LIST of ALL the captures motifs (reasons) to see if we gathered every type we can think of? :)