Random Flow Theory and FOREX - page 44

 
Mathemat >> :

If you subtract the linear regression from the prices, you will, of course, no longer have prices but something else oscillating around zero.

I would like to ask one more question, what can be deducted from prices (for example, MACD or something else) in order to form a new price series? I thank you in advance for your reply.

 

Anything you like, Alexei. But it won't be prices (quotes), but something else. And this other thing can be displayed on the screen - that's what the indicators are for.

P.S. Or do you need prices? Well, for example, if you multiply closing prices of EURUSD by closing prices of USDJPY, you will get closing prices of EURJPY (with some error).

 
Mathemat >> :

Anything you like, Alexei. But it won't be prices (quotes), but something else. And this other thing can be displayed on the screen - that's what the indicators are for.

P.S. Or do you need prices? Well, for example, if you multiply closing prices of EURUSD by closing prices of USDJPY, you will get closing prices of EURJPY (with some error).

My main task is to get the prices without noise. That's why after reading this topic and some other (I don't remember anymore) I came across the method of noise removal, when something is subtracted from the initial quotes flow.You suggest to subtract the linear regression, others suggest MACD. But you upset me when you wrote that it will never be possible to obtain a new quotes flow. Is this approach a dead end?

 
prostoleha >> :

My main aim is to get prices without noise.

Well, it depends on what one understands by noise. As many traders are, so many interpretations. Some people think that there is no noise as such. It's quite a logical point of view.

The trick here is that if we still assume that we have some clean signal, but in reality it is noisy, then the signal-to-noise ratio is much less than 1. So we have to extract the useful part from the extremely noisy signal.

And don't get upset here, the problem is complicated and cannot be solved in one go.

 
Mathemat >> :

Well, it depends on what is meant by noise. As many traders understand it as much. Some people believe that there is no noise as such. It is quite logical viewpoint.

The trick here is that if we still assume that we have some clean signal, but in reality it is noisy, then the signal-to-noise ratio is much less than 1. So we have to extract the useful part from the extremely noisy signal.

And there is no need to get upset, the problem is complicated and cannot be solved in one go.

But is it possible to get new prices from the initial ones through deduction of MACD, say, not linear regression?

 

Of course you can, Alexey. That's what the language is for, and it's powerful enough.

But it will no longer be prices in the original sense. If you like, you can call what you get "net prices" and work with them as you like.

 
prostoleha >> :

But is it possible to get new prices from the initial prices by taking away not linear regression, but MACD, for example?

Before to start taking something from prices, we should understand the reason of it. The price difference between the quotes and the regression is the price cleaned from the trend component. For the same purpose all kinds of MAs are subtracted from prices. If you want to remove the noise component, you can, for example, use lowpass filter LPF, but it will lead to phase delay. In any case, you should first define your understanding of noise, once you have a clear definition, the rest is a technical matter.

 
Mathemat >> :

Of course you can, Alexey. That's what the language is for, and it's powerful enough.

But it will no longer be prices in the original sense. If you like, you can call what you get "net prices" and work with them as you like.

>> Thank you for putting up with me for so long.) One last question Mathemat. If according to your description the calculated value resulting from the operation of linear regression deduction from prices (the same situation will probably happen with MACD) is about 0, I would like to know for sure if it is worth it.

I work with AUD/USD.

If the initial prices of this pair are around 0.8000, and after the conversion they will be around 0.1000 or even lower (about 0), I do not see the point of performing this operation.

And vice versa.

If after transformation the prices are around 0.5000-0.6000, this operation may be called noise elimination.

 
prostoleha >> :

I'd like to know if it's worth it, if it's worth it.

You won't get a net zero anyway, no matter how you look at it. There will still be a residual, which is the basis for decisions to open and close positions. In fact, it's noise too, but that's the point :)

They do different things with prices - subtraction, division, clever filters. And each author believes that all this is appropriate in each case. There is no universal recipe for removing noise, of course. Because there is no general consensus about whether or not prices are noisy.

If you do not want to do anything with them - do not do it. There is an opinion that you don't have to do anything with them and the simpler calculations are, the better for the system. That's what I'm getting to.

 
Mathemat >> :

You won't get a net zero anyway, no matter how you look at it. There will still be a residual, on the basis of which decisions are made to open and close positions. In fact, it's noise too, but that's the point :)

They do different things with prices - subtraction, division, clever filters. And each author believes that all this is appropriate in each case. There is no universal recipe for removing noise, of course. Because there is no general consensus about whether or not prices are noisy.

If you do not want to do anything with them - do not do it. There is an opinion that you don't have to do anything with them and the simpler calculations are, the better for the system. I'm getting to that opinion, too.

Thank you very much for the chat. Anyway, it turns out that our conversation ends on such an intriguing note after all, which means we have to try and determine whether this subtraction operation will be useful.Thanks again.