The main question is, what is the best way to learn pattern recognition? Via learning a zillion patterns (Tempo), or via learning a relatively small alphabet of patterns (Heisman) around which your skills accrue?
Until now I haven't given these items much attention because I thought I had not the time to develop something my self. So I accepted a less than optimal method as CTS. But let's have a close look on this issue, maybe some good idea's arise.
First I would like to reformulate the second question as follows:
What is a better methaphore, to compare pattern recognition with learning the tables of multiplication or with learning the vocabulaire of a new language?
The tables are limited to 120 items, a vocabulaire of a new language is 20,000+
Prof. Adriaan de Groot mentions somewhere that a (grand-?)master knows 50,000 positions by heart. Other sources speak of 100,000 positions.
An article that especially inspired me was about the five stages of skill acquisition. They talk about 50,000 positions for a chess master.
So the comparison with learning a new language seems to be the most appropriate.
Accrueing skills around a few basic patterns has never worked with me.
Back to the main question:
What is the best way to learn 50,000-100,000 patterns?
I'm not too far from finishing circle 4 at CTS.
This means that at average I have seen every problem of the window (=10,000 problems) 4 times by now.
My average rating at CTS is increased with 50 points.
What does this mean?
If you look at the bell curve of the rating distrubution of the problems you can see that my problem window has shifted 1,300 upwards.
This means that I ingrained 1,300 patterns at the downside of the problem window in my system and that I got 1,300 new problems at the upside of the window.
Summarizing: it took me 6 months to gain 1,300 patterns.
That means 50,000 patterns will cost me another 19 years.
The price of being objective bears heavily on me, thanks, Blue Devil:)
There are three marginal notes:
- I took a break of 3 weeks what costed me 2 months to recover.
- I am almost finished with circle 4 now so I'm expecting an accumulated effect the next circles.
- Since a few weeks I altered my method. This seems to work well.
So I have to find a faster method. Idea's anyone?