I add here a screen shot on the 1 h time frame together with the silver trend indicator.

OK not bad at all. Look at the Andrew's Pitchfork from the previous shot and compare with the present, in fact this was nothing else than the manual driven linear regression.

OK now you can check the two shots, one is with the Andrew's Pitchfork, the shot here we are using the Linear regression indicator with the same lenght.

How to define the best period for the relevenant linear regression?

Well that is not an easy question, basically you need a visual inspection, not much different from plotting the technical analysis studies.

Finally, we have really a very close result. The Andrew's Pitchfork tools do not neead to compute the least squares in order to find the best fit line but the end result is roughly the same.

So why that matters?

It matters because all the oscillator readings (and oscillator in general) can be interpreted regarding the linear regression model.

By now we have a clear oversold level in the stochastic.

Does it mean that the price is oversold and is bound to jump back?

Not necessarely!

The reason to show you this shot is to make my point that the oscillator studies readings and their oversold and overbought levels have a meaning only within the context of the linear regression.

Look now we are exactly at the bottom of the linear regression channel and at that level we have an oversold indicator reading.

At the beginning of the channel we had the same situation. Bottom of the channel with oversold levels.

So that will make a perfect sense only in the case that the linear regression model will hold.

If the linear regression model will hold that will be really an oversold level. If the linear regression model does not hold anymore it will not work (the oscillator is showing a very strong downards momentum).

If you look at the computation models you will see that despite the fact that the formulas are different all the oscillators can be interpreted as deviation from a central tendancy.

Beong on holiday for quite for a while I was thinking about a holistic understanding of the whole technical analysis, the many different theories, approaches and methodologies that are available.

You are into an ocen of indicators and strategies some proprietary some free on the net. The indicators available on FF are really countless from my perspective.

As you can see on this shot, the cycle studies are completely different studies from the oscillator studies. Have a look at the linear regression model on the daily chart, it makes some sense with what happens today isn't it?

## Comments

anymore indicators with OB/OS

guys i want introduce this ea http://www.stevehopwoodforex.com/phpBB3/viewforum.php?f=34

/Not bad with filter on h1 work on m5/

I add here a screen shot on the 1 h time frame together with the silver trend indicator.

OK not bad at all. Look at the Andrew's Pitchfork from the previous shot and compare with the present, in fact this was nothing else than the manual driven linear regression.

OK now you can check the two shots, one is with the Andrew's Pitchfork, the shot here we are using the Linear regression indicator with the same lenght.

How to define the best period for the relevenant linear regression?

Well that is not an easy question, basically you need a visual inspection, not much different from plotting the technical analysis studies.

Finally, we have really a very close result. The Andrew's Pitchfork tools do not neead to compute the least squares in order to find the best fit line but the end result is roughly the same.

So why that matters?

It matters because all the oscillator readings (and oscillator in general) can be interpreted regarding the linear regression model.

By now we have a clear oversold level in the stochastic.

Does it mean that the price is oversold and is bound to jump back?

Not necessarely!

The reason to show you this shot is to make my point that the oscillator studies readings and their oversold and overbought levels have a meaning only within the context of the linear regression.

Look now we are exactly at the bottom of the linear regression channel and at that level we have an oversold indicator reading.

At the beginning of the channel we had the same situation. Bottom of the channel with oversold levels.

So that will make a perfect sense only in the case that the linear regression model will hold.If the linear regression model will hold that will be really an oversold level. If the linear regression model does not hold anymore it will not work (the oscillator is showing a very strong downards momentum).

If you look at the computation models you will see that despite the fact that the formulas are different all the oscillators can be interpreted as deviation from a central tendancy.

Beong on holiday for quite for a while I was thinking about a holistic understanding of the whole technical analysis, the many different theories, approaches and methodologies that are available.

You are into an ocen of indicators and strategies some proprietary some free on the net. The indicators available on FF are really countless from my perspective.

Are you still looking for the Holly Grail?

As you can see on this shot, the cycle studies are completely different studies from the oscillator studies. Have a look at the linear regression model on the daily chart, it makes some sense with what happens today isn't it?

I am very grateful for all. Now I understand how to use in practice Andrew's Pitchfork . I am try to use Zup indicator for Andrew's Pitchfork points.