22 April 2012

Heuristics and Review

Read about Heuristic

Infer things from common sense and use it for solving a problem. The human generally don't believe in common sense unless they perceive a pattern (we usually term this as experience) and once they find the pattern the human hardly believe that the pattern can change. We start with a low confidence (that leads to frequent checks/reviews), build the confidence by perceiving patterns and become over confidence (through the knowledge gained). Over a period of time, we tend to ignore the signs of pattern reversal. It is not pattern reversal but it is mere lack of understanding the system (one theory is that we fail to understand the theory of large numbers. For example, the probability of getting a head is 0.5. In the current trial, if we get a tail, it doesn't mean that you will get head next time. It means that for (real) larger number of trials you will get 50% heads and 50% tails.).

For example, when you try to learn to drive bicycle, you are tend to be very careful (cautious). Then, why do accidents happen? (check out theory of large numbers, theory of medium numbers)

I think, this applies to everything and the failure of things is more pronounced in discrete systems.

So, it is absolutely critical that we review (to keep inferring things/environment) and anticipate failure.

Failures are inevitable (as most of the systems in this world are really medium and hence there will be irregularities)

One can rely on heuristics if he reviews the heuristics properly (keep inferring and never conclude).

More gyan can be found in Introduction to General Systems Thinking by Weinberg.