How to buy stock - searching for patterns

So now lets get down to the nitty gritty, of how I found a stock and index trading system.

A systme that can be demonstrated would not only have produced the return on that graph - but would have made continuous gains for the last 20 years.

Moreover the method I am about to describe can find other trading schemes too.

The first key issue it is important to understand is simply a way of thinking.

Now the scheme that produced that graph, has NOTHING to do with moving averages or any conventional indicator system: but a moving average system will do wonders for illustrating the point I have in mind.

PROGRAMMING IS NOT THE ANSWER

Historically what people do: is write a program to test a system.

If their idea is buy when average1 crosses average2 and sell when average2 crosses back the other way. That is simple isnt it?

Theres only two possibilities - if A is greater than B or B is greater than A. Simple to say and simple to test in one case you are long, the other short or out of the market
Sadly of course if fails every time - nothing that simple works.

So then you decide that may be it will work if I only buy is not only if A is greater than B, but also the gradient of A and B are both above zero, that is they are both increasing.

So then your purchase criterion depends not on one thing - but that 7 possibilities - thats
A is the 5 day moving average lets say
B is the 12 day moving average maybe
Here are the possibilites

gradient A> 0 Gradient B>0 A>B
1 yes yes yes
2 yes yes no
3 yes no yes
4 yes no no
5 no yes yes
6 no yes no
7 no no yes
8 no no no

One of those 8 conditions is used as a buying signal, another one to sell, one to go short and one to buy back.

It isnt obvious which should be used. So there is lots to test and change in the program, as well as changing the average durations.

Reality is nothing as simple as that works either

But there are already a LOT of combinations of possible buying and selling criteria to chose.

It would take many hours to do a rigorous test - when stop loss criteria and trading costs are also included.

ANY TESTER WILL TELL YOU IT QUICKLY GETS OUT OF HAND!!

Then you decide that it is actually the stock price that matters..

So it is when the stock price rises above both of the averages , and both of the gradients are positive.

So then the number of possibilites are already 16 for buy or sell.
And the number of permutations is huge.

The more factors you put in , the more the testing gets out of hand.

And then of course you want to include - whether or not the stock opens higher than yesterdays center of trading.

And eventually you have too many possibilities to test
Too many to test using software changes.

THE ANSWER IS TO THINK IN A COMPLETELY DIFFERENT WAY

And now my military signal processing background kicked in.

The truth is computers are not the best things to find patterns.
The human eye and the human mind are far more powerful.

So what most military systems do, is not take the decisions , instead it presents the information to the human in a way that it is easy for the human eye to take over.

So let us go back to that example again.

If there are 8 different possibilites for buy signal, and 8 possibilities for selling lets say, then there are 64 possibilites total for trading schemes.

Now lets introduce the concept of SCATTER - it is scatter that has found all my trading schemes.

So now consider you have 64 bins.
Each bin is what is called a scatter graph. You put a spot or a cross or a tick, for each time that condition was triggered: at a height on the graph according to how much it won or lost.

Then you let the eye take over . Most of the bins have random spots. Just one or two bins have a cluster of spots of mostly wins or mostly losses. The eye tells you whether there is a pattern or not.

And in this way you can test 64 trading possibilities all at the same time.

It is these scatters which I have used todecide how to buy stock

As clear as mud? Probably!
In the next post I will give you a simple example.
If you are interested in this, please stumble or digg it, or click it to netscape!!!

That way I will decide whether to continue!