Can tactics be learned / improved ?


Thinking about how we think(!)
Training our tactical vision

Thinking about how we think(!)

Computers are very good tactically, because they can analyse variations very well and very quickly!. They will look at many more candidate moves than human beings, and consider even the craziest looking moves.

A budding Fischer, Kasparov or Tal must get very good tactically, and be able to find combinations easily.

Kotov in [1] suggested a policy of establishing candidate moves, and being systematic in analysing them. He made an interesting assumption with respect to the way computers calculate tactics. He suggested that we could take some of the methods that computer's employ and use them within the context of our own thinking processes. For example systematically analysing positions, thus producing variations quickly.

Being able to see moves ahead and assess the positions is a vital ingredient to combinational vision. Seeing ahead is a skill that can be evolved through practice. Kotov suggested looking at very complicated tactical situations, and trying to analyse them and compare the analysis to the actual game notes. He noted several very important tips related to looking ahead and finding the resources buried in the position. These included:-

These tips need some intelligent qualification.

The way computers "think" is completely different to the way humans think.

Humans have the ability to abstract and form plans. It is easier to implement computer chess programs with alpha-beta pruning, rather than teach them about planning concepts.

Finding really important variations is helped by experience. Experience will act as a filter to find the most relevant moves quickly that meet the needs of the position. An experienced sicillian Dragon defence player for example, will look at variations involving the thematic Rxc3 quite often for example. An experienced Kings Indian player will be familiar with the tactical themes of the Kings Indian.

Computers have no emotions and fear. Their programs do not become "tired" and make elementary blunders.

Being systematic and not going over variations twice is an idea borrowed from how computers analyse positions. It seems slightly risky when for example you are going to sacrifice your queen or heavy material. It is re-assuring to try and check the soundness of these types of combinations! Computers have no fear and do not have to suffer the emotional implications of unsoundly sacrificing a queen!

This type of human re-assurance has to be balanced with Kotov's fundamental point about discipline in terms of not jumping from one variation to another and back again. This could have very bad consequences such as running out of time.

Having made these qualifications:- in tactical situations, it would be nice to produce a computer like analysis. The structural framework and discipline that computers employ to analyse variations is therefore a very interesting idea in principle.

Training our tactical vision

The following set of tips may help train tactical vision:-


Click Here!

This page belongs to:
Barnet Chess Club