Sunday, November 13, 2011

Going in circles, so I'm making progress

After lengthy discussions with mr. Z I decided to see if I could bring some organization in the list of concepts I created while describing the solution of hundred very complex tactical problems. These are the sections which emerged after sorting:

  • Mate in 1
  • Checks
  • Captures
  • Threats
  • Duplo attack
  • Clearance
  • Overworked piece
I didn't see that one coming!
Much to my surprise I'm back at CCT again.

I often said that this blog is going around in circles and talks about the same topics everytime.
I prefer to look at it as spirals, though. Everytime I reach the same point, but with additional knowledge. The facts that the same topics keep reoccuring time and again after carefull reasoning, means that the solution must lie within those topics. It is like a chess combination. You know the moves, but you have to find out the order. All this is thus very much unlike this picture above which is NOT A SPIRAL but consists of circles instead.

Memory lane.
The last time we talked about CCT was here. Why did I leave the concept in the first place? I made a dissection of CCT in CCT and visualization here and I went on looking for visualization. I found that visualization is greatly helped by guidance by means of concepts. And now those concepts show that they are organized according to CCT!
Remember that we are speaking here in the realm of tactics, which is the ultimate place where visualization will give you an edge.


  1. I find this validation of CCT (Checks, Captures, Threats) as the base thinking process for each move quite valuable. I've recently had to acknowledge that my thinking process has been poorly structured, if you could call it structured at all. Implementing a well-understood CCT process, with reinforcement through tactical exercises, should go a long way to cleaning up my thinking process and eventually strengthening my game considerably.

  2. me Munich:
    I did the "attack", "defence", "check" training with rybka (=Fritz GUI).
    Alltogether I invested maybe 20 hours? Maybe even less. At the very beginning I remember I could do 8 attacks per minute.
    After doing it for an hour or so I was allready much better. 20 attacks per minute I might have reached after 2 hours or so.
    From there progress was not so fast. However, if I remember correctly, my very best record was 47 attacks per minute, 45 defences per minute and 35 checks per minute.
    I believe it helped me tremendously to see immediatly for instance a hidden black bishop an a7 attacking f2.
    I am not sure if it makes much sence to train this ability further, but I think it is not wasted time. You only reached 20 per minute or so, so there should be plenty of room still left to get a quicker board overview. Some more hours doing this training is not a big sacrifice. It supports your CCT-Thinking.

  3. I did go the whole nine yards with Fritz defence, attack and checks.

  4. me Munich: Well, you might want to check again if you are still at 20 attacks/checks/defences per minute?
    From your blog "3 down, 27 to go" you stopped with 20 attacks. But if I managed over 40 (47 to be exact), at least you should be able to reach 40, too.
    Still, even 40 might not be the limit. I think one click per second could be possible, so I would not wonder if somebody manages 60 attacks per minute.

    Besides: "3 down - 27 to go" --> what about the rest?
    I did forks, discoveries, skewers, capturing defenders, overloadings&distractions (together), x-rays.
    These are 6-7 (depended if you count overload & distractions seperatly). Of the 39 tags (CT introduced some new tags) not all make much sense to train.
    For instance "Needs more moves" --> what kind of tag to be trained is that? Or take "avoid stalemate" - there are 10 puzzles in the range 1150-1475. That would not make much of a set that occupy you for days... :-)

    On top of these 39 you have the 3 boad vision trainings (Fritz GUI attack, defence, check training). You counted them together, so we have 39+3 = the 42nd street? Plus think about the possibility to train check mates (not a tag, but you can create a "mate only" tag, and enter the minimum/maximum moves the puzzle should have).
    I'd say, that there are only around a dozen of tags that make sense to be trained, plus the "mate-in-1", "mate-in-2", "mate-in-3", "mate-in-4", "mate-in-5-and more".

    Well, 3 down - many more to go: Come on, do them. You will learn lots of simple concepts. Those puzzles I believe are ideal candidates for beeing a prototype of a tactic (with a typical concept/tactical combination) I marked with my own created tag "pattern". Each tag has its own typical concepts. You wont memorize the puzzles, but rather the general concept of them (but true, as a side effect, you will also memorize many of them). I recommend to do these tag-sets rating sorted, because somehow typical patterns tend to have somehow similar ratings. The prototype of a "smother" (=queensac and then a check with the knight (=the "two main actors"), which kills the king by suffocation (=environment is typically some pieces around the king) you will find very often around 1300-1400.

  5. Munich said: "I am not sure if it makes much sence to train this ability further, but I think it is not wasted time"

    "Mental power, that means the capacity C of short-term memory (measured in bits of information) is the product of the individual mental speed Ck of information processing (in bit/s), and the duration time D (in s) of information in short term working memory"

    See f.e. here

    The Working Memory is a very important limiting faktor in Problem-solving ability (IQ).
    Experts can enhance the limited Capacity of the WM by creating Chunks.
    But mental speed is important too because the STWM's duration is short. So: As quicker you can think as more complex problems you can solve (=~ as more you see )

    I think the high speed easy tactics training produces chunks AND speed.
    IQ-Trainings-problems to improve mental speed are looking very much like the "attack", "defence", "check" training of Fritz: Easy problems to solve quick. See f.e. :

  6. Munich: I am currently doing a set of discoveries in the CT Blitz rating range of 1150-1475.
    Here is a typical tactic which you will see about 20 times in slightly different variations, sometimes on the left wing, sometimes on the right wing sometimes witch queen and bishop instead of rook and bishop. You will not memorize this puzzle, but after you did the whole set, you will know the concept and can solve in 4-7 seconds, depend upon whether it is a 2 or 4 mover.

    You wont see this concept as often if you didnt to a tag-themed set. I also noticed, that if you do them rating sorted, typical concepts have a typical rating. I doubt very much that I conciously do the CCT process. If I do it, then it is subconsciously. But I think I rather remember the concept and just "do it".

  7. Me Munich again:

    The last one is the "most difficult" of them (by rating), with a rating of 1421.
    But the similarity is plain to see. I call it "pattern", but I think tempo sees this a "concept"? I still cant really tell the difference between pattern and concept.
    In any case: You do not need to worry, that you simply memorize the puzzle and not memorize the typical tactic behind. If you look at these 5 different puzzles, and you did not know about this pattern/concept/typical tactic, then you learned a tiny little bit that that should result in higher elo. I learned a few hundred of such typical tactics, and many I did not know before.
    It is a huge difference if you do 100K puzzles that are not connected by a tag-theme, or if you do easy tag themed puzzles, sorted by rating. The latter will give you lots of examples of similar typical tactics (=concepts), and you will be able to reproduce them instantly in your games. But if you dont do tag themed puzzles, and if you dont repeat puzzles you did wrong (= puzzles where you did not understand the concept), and if you dont repeat puzzles that took you long (puzzles where you hardly understood the concept), but simply do 100K puzzles very fast and keep rushing through them --> then you wont learn much. Tempo proved it.

  8. I made quite a few discoveries lately which adress the points you make. I will try to post about this tonight.