Monday, September 26, 2005

How to play long lines a tempo.

Working with CTS has given a whole new meaning to the expression "a tempo".
CTS shows that the slightest hesitation costs you already a few seconds.
To get the full reward at CTS you have to give the total solution within 3 seconds.
That is really short.
Below you see a typical situation.

Black to move.
The last move of white was Ne5, attacking the black queen.
The first thing that came to mind was 1. ... Nxe5 but that loses the queen.
Then I saw 1. ... axb5, winning a piece.
I did the move, then came a shock: 2. Qa8+
I hadn't seen this check. But with 2. ... Kc7 blacks problems are solved.
Total time used: 6.2 seconds.
About 3 seconds per move.
That I could solve the check was pure luck, I hadn't seen it coming.
That is an extra on CTS what you don't have OTB, there is always a good solution somewhere.
The question arises if it is actually possible to see the whole line within 3 seconds.
The details of the problem revealed that there were only 6 persons out of 60 who were able to solve the problem in 3 seconds or less.
Those 6 persons had an average rating of 1774, that is 250 points higher than mine.
Working with CTS is as if I have to learn chess again.
But now a tempo.
I have to learn all the one-movers a tempo first.
Before that, it is impossible to learn the more-movers.
Even with a problem like above I see actually two seperate one-movers and not a two-mover.
So there is a long way to go.
The fact that there are now 4 masters playing at CTS and that they all have a high rating is enough prove for me that I'm on the right track. They all are extremely fast in more-movers.
The most important thing is to realize how bad we are in one-movers.

Doing 300 - 400 problems a day should do the job to adress this problem.

circle 0: 1470
circle 1: 1500
circle 2: 1520
circle 3: 22,000 /70,000
Highest rating 1567


  1. Hi there,

    I must first say great, active blog you have. It really helps me thinking about my own chesstraining.

    The reason I write this comment is because I have a question about the CTS. I'm sorry its not related to your post "how to play long lines a tempo" and maybe it has been answered somewhere else but I like to address this topic nonetheless.
    I was wondering if the set of problems is not too big at the CTS to effectively absorb the problems. I mean I have access to 8000 problems or so and while doing the tactics training, I mostly recognise and not really memorize the patterns, its just like playing blitz to me. Blitz doesn't necessarily increase your database of patterns while the same patterns occur over and over again. What are your views on that?



  2. Jeroen, that is a very important question. Actually this is the main reason why I work as a madman at CTS. Because if the pace is too low the work has not enough momentum. If your rating is 1500, the rating window is from 1360 - 1640 (+/- 140). This contains roughly 10,000 problems. The current problem is selected randomly from that window.
    I have done 22,000 problems. This means that at average, I have seen all problems 2.2 times. But some problems I have seen allready 5 times. This means that there must be
    problems too that I haven't seen at all. I estimate that in practice I have seen 10 - 15% of the problems before.

    1 circle costs me about 1 month. So from a learning point of view you are probably right.
    It is not ideal to learn 10,000 items a month, and to repeat them the next month. But the problems are so basic, that it is always possible to reconstruct the answer, given enough time. Ok, your rating will drop from this, but that's no problem.
    You only can go forward by learning to do the lower problems a tempo. Your rating will rise and the window shifts upwards. Until you know all 28,000 patterns by heart.

    What simplifies matters is that certain patterns are recognized as being the same. See my previous post of sept. 21. Say that there are 200 queensacs followed by a knightfork within the problem window. Say that every group of 10 problems is recognized by the mind as being of the same pattern. This means that effectively the database of 10,000 shrinks with a factor 10.
    But I don't know the real figures here.

    Now I really have to stop rambling about the work, because that is not going to do the job for me. Back to CTS. . .

  3. Tempo,

    Thanks for your answer,I understand your reasoning now better and I
    hope it works out for you. I know it probably wouldn't for me because
    I need to go over a set of problems atleast 20-30 times to remember
    the patterns(solve the problems in less than 5 sec), that would mean
    300k tries per window *eek*

    please keep up the good work.

    Kind regards,


  4. In more-than-one-movers it seems you get additional time for free with every move. In the first move you get 6 seconds for free before your clock begins to run. Is it another 6 seconds for the following moves? In order to test this one should log in as guest with a stop watch. I did not yet.

  5. Mouse, good point! I measured it with a metronome. For the first move you get 6 seconds for free, for every additional move you get 3 seconds. The total of the free time is subtracted from the total time you use for the problem. The value after subtraction is displayed by CTS as your used time.