.

Indicators by William Blau

Moderators: mntiwana, mrtools

Re: Indicators by William Blau

#71
Krelian99 wrote:
Tue Dec 25, 2018 2:33 am
MATLAB-documentation: Hilbert-Transformation returns a complex helical sequence, sometimes called the analytic signal, from a real data sequence.

The analytic signal x = xr + jxi has a real part, xr, which is the original data, and an imaginary part, xi, which contains the Hilbert transform. The imaginary part is a version of the original real sequence with a 90° phase shift. Sines are therefore transformed to cosines, and conversely, cosines are transformed to sines. The Hilbert-transformed series has the same amplitude and frequency content as the original sequence. The transform includes phase information that depends on the phase of the original.

The Hilbert transform is useful in calculating instantaneous attributes of a time series, especially the amplitude and the frequency. The instantaneous amplitude is the amplitude of the complex Hilbert transform; the instantaneous frequency is the time rate of change of the instantaneous phase angle. For a pure sinusoid, the instantaneous amplitude and frequency are constant. The instantaneous phase, however, is a sawtooth, reflecting how the local phase angle varies linearly over a single cycle. For mixtures of sinusoids, the attributes are short term, or local, averages spanning no more than two or three points.
The answer for you how to use it is on slide 28/29 of the pdf. The osci period should have 1/2*dominant cycle to be in phase, so set PAcycle to 0.5 for scalping/swing trading, 1 to 2 for trend tranding. Maybe use 2 instances with different cycles for better riding the momentum.
Thank you very much Krelian99. Makes a lot of sense. So should we always set Phase Accumulation PaFilter to the same value as PaCycle and if so, why is that? Usually filters eliminate noise but just to be sure, exactly what is PaFilter filtering out in this case? Also are there maximum and minimum values (50 for example) recommended for the DzLookBackBars (Dynamic zone look back) parameter or is 70 the optimal value and if so why 70? I prefer to use 0.05 for the Dz Probability options.


Re: Indicators by William Blau

#72
global wrote:
Tue Dec 25, 2018 8:45 am
Thank you very much Krelian99. Makes a lot of sense. So should we always set Phase Accumulation PaFilter to the same value as PaCycle and if so, why is that? Usually filters eliminate noise but just to be sure, exactly what is PaFilter filtering out in this case? Also are there maximum and minimum values (50 for example) recommended for the DzLookBackBars (Dynamic zone look back) parameter or is 70 the optimal value and if so why 70? I prefer to use 0.05 for the Dz Probability options.
No, as all-pass with the Filter you filter out high and low frequencies. The art of filter designing is to filter out disturbance and frequencies that you don't need or want to see. Low frequencies have to less energy and too high frquencies is disturbance. You can filter with Hilbert Transformation, for proper filter design I recommend you other types of filters.

NTL, values around 80 for PAfilter give good results.

Re: Indicators by William Blau

#73
Krelian99 wrote:
Tue Dec 25, 2018 11:33 pm
No, as all-pass with the Filter you filter out high and low frequencies. The art of filter designing is to filter out disturbance and frequencies that you don't need or want to see. Low frequencies have to less energy and too high frquencies is disturbance. You can filter with Hilbert Transformation, for proper filter design I recommend you other types of filters.

NTL, values around 80 for PAfilter give good results.
Thanks for the info but although do some coding even if I know of better filters to use I really won't know how to add them to replace the iHilbertPhase function within the indicator. Anyway, thank you very much for your explanation. I now have a much better understanding of how the indicator works. I'll try around 80 for the PaFilter and see how that goes.

Re: Indicators by William Blau

#75
tommo78 wrote:
Fri Jan 04, 2019 12:02 pm
is it possible to upload the MQ4 version of this indicator please. Cheers
Tom, you've been a member here since we've started. As Mladen, Mrtools, Mntiwana and everyone else has repeated zillions of times, EX4's are provided as is and if there's no MQL4 file, we won't provide it to prevent theft etc.
But as a reminder, here are our rules (in case you've missed them for the past 2 years :|).
Myfxbook live trading results (new 2019 account coming soon)
Need custom MT4 coding? Send a PM to Mrtools or post here for a quote!


Re: Indicators by William Blau

#77
Krelian99 wrote:
Tue Dec 25, 2018 11:33 pm
global wrote:
Tue Dec 25, 2018 8:45 am
Thank you very much Krelian99. Makes a lot of sense. So should we always set Phase Accumulation PaFilter to the same value as PaCycle and if so, why is that? Usually filters eliminate noise but just to be sure, exactly what is PaFilter filtering out in this case? Also are there maximum and minimum values (50 for example) recommended for the DzLookBackBars (Dynamic zone look back) parameter or is 70 the optimal value and if so why 70? I prefer to use 0.05 for the Dz Probability options.
No, as all-pass with the Filter you filter out high and low frequencies. The art of filter designing is to filter out disturbance and frequencies that you don't need or want to see. Low frequencies have to less energy and too high frquencies is disturbance. You can filter with Hilbert Transformation, for proper filter design I recommend you other types of filters.

NTL, values around 80 for PAfilter give good results.

Hi Krelian99,

I was just revisiting your response to my post quoted above where you said:

You can filter with Hilbert Transformation, for proper filter design I recommend you other types of filters.

From you responses you seem to really know a lot about cycles, filtering, etc, so could you please let me know what other types of filters do you recommend for proper filter design?

Thank you very much.

Regards,
global

Re: Indicators by William Blau

#78
global wrote:
Mon Feb 04, 2019 5:49 am
Hi Krelian99,

I was just revisiting your response to my post quoted above where you said:

From you responses you seem to really know a lot about cycles, filtering, etc, so could you please let me know what other types of filters do you recommend for proper filter design?

Thank you very much.

Regards,
global
I actually missed it till now to do it properly. I had so many other interesting stuff. I followed the filter nothing approach, since for proper filter design you need a program like Matlab and I hadn't the time till now to charting Matlab with the MT4 data feed. Financial data also follow other rules than technical data. There are natural rules, so clear boundaries are given like damage in material the available amount of energy. Here are no natural rules. Here, a fat finger of a big player and the price jumps 50pips up or high, we also had 150pips. The boundaries in financial markets are S/R, singularities though and a proof that markets are human driven and the algos also behave like humans.

When you want to filter NTL, you need to know what to filter out regarding to your concept. Disturbances or the shakings or high till low frequency. Scalping don't need the higher frequencies, while trend following needs no lower frequencies. It must be clear that Online-Filtering also always has a delay, otherwise it's intended repainting (for offline use). You must do a system identification. Bode-Diagram is normally a good tool, but I doubt it helps in financial markets. They are too dynamic. In FX is a calm Asian Session, here it could help maybe. Of course there are much more powerful SI-tools, but I never checked them for FX. Neural network machine learning maybe could also be interesting, but here I know too less.

An often used simple FIR is the EMA or SMA. As dynamic S/R it is pretty good. Cause and effect are changed here, so IMO people who say an EMA is mathematical manipulation of price are not wrong but they use it the normal way. Here, the price tests the EMA whether bulls or bears come in or stay out (I have tons of examples), so some candles on the chart follow the EMA. This is a failure in the causality (seeing the future, e.g. a pit hole in the street you avoid, math jumps in) and never happens in natural or technical systems.

Re: Indicators by William Blau

#79
Krelian99 wrote:
Mon Feb 04, 2019 9:33 am
I actually missed it till now to do it properly. I had so many other interesting stuff. I followed the filter nothing approach, since for proper filter design you need a program like Matlab and I hadn't the time till now to charting Matlab with the MT4 data feed. Financial data also follow other rules than technical data. There are natural rules, so clear boundaries are given like damage in material the available amount of energy. Here are no natural rules. Here, a fat finger of a big player and the price jumps 50pips up or high, we also had 150pips. The boundaries in financial markets are S/R, singularities though and a proof that markets are human driven and the algos also behave like humans.

When you want to filter NTL, you need to know what to filter out regarding to your concept. Disturbances or the shakings or high till low frequency. Scalping don't need the higher frequencies, while trend following needs no lower frequencies. It must be clear that Online-Filtering also always has a delay, otherwise it's intended repainting (for offline use). You must do a system identification. Bode-Diagram is normally a good tool, but I doubt it helps in financial markets. They are too dynamic. In FX is a calm Asian Session, here it could help maybe. Of course there are much more powerful SI-tools, but I never checked them for FX. Neural network machine learning maybe could also be interesting, but here I know too less.

An often used simple FIR is the EMA or SMA. As dynamic S/R it is pretty good. Cause and effect are changed here, so IMO people who say an EMA is mathematical manipulation of price are not wrong but they use it the normal way. Here, the price tests the EMA whether bulls or bears come in or stay out (I have tons of examples), so some candles on the chart follow the EMA. This is a failure in the causality (seeing the future, e.g. a pit hole in the street you avoid, math jumps in) and never happens in natural or technical systems.

Thank you very much for your response Krelian99. I will ponder what you shared to see if I can practically apply something new.



Who is online

Users browsing this forum: aikgio, billbroch, CommonCrawl [Bot], heispark and 22 guests