Fast Fourier Transform - Cycle Extraction - page 68

 

thanksgiving

Hi mladen,pip

Thanks you so very much for your understanding ,help and enlightenment.

Kind regards,

irena

 
someoneinjapan:
Pip,

Can you point me to a post where you talk about figuring out cycle lengths?

Thank you!

Someoneinjapan

Hi Someoneinjapan,

This entire thread is devoted to the simple question that you ask But, i will try to summarize the methods mentioned in it.

The exercise involves two aspects; the first aspect is determining cycles present in the historical period that you are analyzing, while the second aspect is determining the most dominant cycle.

As for the former, there are several ways to determine the present cycles the most basic of which is visual estimation. This is done by measuring the distance between peaks or troughs and determining the most repeated distance. This will tell you the available cycles and also the most dominant. Another method is by performing "phase analysis". There are several tools in this thread and in the "Digital Filters" thread that perform this task. By performing phase analysis you are simply identifying all repetitive cycles in the historical period you are studying.

As for the latter, there are several ways to determine the dominant cycle in the historical period that you are studying. I already mentioned one of them above, which is eye balling the distance between peaks and troughs and determining the repeated period. The second method is done through mathematical computation to determine the dominant cycle.

After determining the present cycles and identifying the most dominant you still need to test whether the identified cycles are actually there and persistent (i.e. will most likely continue in the future) or not. This is also done via statistical means.

Do bear in mind that i have over simplified things for each of explanation. For more details you will need to read this entire thread.

Hope this helps.

Cheers,

Pip

P.S. I am by no means an expert in cycle analysis and hopefully others will add to my explanation.

 

Thanks, Pip. That makes a lot of sense. I appreciate your feedback.

 
someoneinjapan:
Thanks, Pip. That makes a lot of sense. I appreciate your feedback.

My pleasure.

Mladen, i sent you a PM

 

Cycles

Hello Pip,

I thank you very much for this thread.

Obviously time trading can be separated in 3 phases: past, present and futur.

About past: It is evident that all indicators re-draw (in a certain manner) or draw past periods. One can have the best algos , indicators are based on past period Maths formulas. (No one can trade in the past).

About present: Price action can be operated via straddle method, (that's where I am know in the time scale). Price action gives me good results. I am happy with that as I have developed edge tools to monitor price action. (all trades are operated in that period except few ones due to lag, latency, order routing, delay of the broker, spoofing...)

Present milliseconds ahead: HFT are now at this stage or milliseconds ahead...1 step farther (that's sound barrier)

Now about futur: and this is why I write on this thread. Clearly, this address is right for you Pip:

Time-Price-Research: The Kondratieff Cycle & its Subdivisions

A good description of different kind of cycles is clearly explained and can en-light some of us.

Cycles can obviously help to speed up to "light barrier".

Any news of MiniMe ?

@ Mr Tool or @ Mladen: In advanced sections, are there "cycles thread" groups developed ? can be worth?

Not much links about that...Thats' a field to digg in...

Anyway take care and evaluate risk. Risk = frequency of events * probability of a damage (a loss). All these evaluations require a closed monitoring loop to lower "risk"

Trying to speed up...

Tomcat98

 
Tomcat98:
Hello Pip,

I thank you very much for this thread.

Obviously time trading can be separated in 3 phases: past, present and futur.

About past: It is evident that all indicators re-draw (in a certain manner) or draw past periods. One can have the best algos , indicators are based on past period Maths formulas. (No one can trade in the past).

About present: Price action can be operated via straddle method, (that's where I am know in the time scale). Price action gives me good results. I am happy with that as I have developed edge tools to monitor price action. (all trades are operated in that period except few ones due to lag, latency, order routing, delay of the broker, spoofing...)

Present milliseconds ahead: HFT are now at this stage or milliseconds ahead...1 step farther (that's sound barrier)

Now about futur: and this is why I write on this thread. Clearly, this address is right for you Pip:

Time-Price-Research: The Kondratieff Cycle & its Subdivisions

A good description of different kind of cycles is clearly explained and can en-light some of us.

Cycles can obviously help to speed up to "light barrier".

Any news of MiniMe ?

@ Mr Tool or @ Mladen: In advanced sections, are there "cycles thread" groups developed ? can be worth?

Not much links about that...Thats' a field to digg in...

Anyway take care and evaluate risk. Risk = frequency of events * probability of a damage (a loss). All these evaluations require a closed monitoring loop to lower "risk"

Trying to speed up...

Tomcat98

Tomcat, THANKS for the link!!!!!!!!

 

half_spectrum_FIR Indie

Dear mladen & Pip,

Need Help education for this Indie, please...

Thank's in Advance

 

Tsar,

Take a look at these 2 posts :
https://www.mql5.com/en/forum/178842/page45

https://www.mql5.com/en/forum/178842/page45

You will find some more info about it at those 2 posts

 

Nice!

Many thanks

 

Hi Tsar

it's a low-pass FIR filter.it uses from past data until present data and didn't use from future data.

traditionaly low-pass filter uses from past data and future data. for example by using SINC() function that produce coefficient of low-pass filter , everyone can produce filtered signal .

in google search "FIR Low-pass filter introduction".

in this indicator i use from past data till now and coefficient of SINC() function for low-pass filter

and convolve they(CONVOLUTION FUNCTION). wn is cut of the low-pass filter.

by using of this parameter you can make (harsh - smooth) trade of (no-lag ~lagged) output

good luck