Hi. The manual re. HTDcPeriod says



without any description of why DominantPeriod is more accurate. They give obviously different results, but what makes DominantPeriod better?

Also, there is no DominantPhase as there is an HTDcPhase, etc. Do these not make sense for DominantPeriod?

Thanks.

Yes, we could certainly implement a DominantPhase if required. We just hadn't needed it so far. The DominantPeriod is more accurate because it filters noise spikes better.

Yes, please add DominantPhase to the list. In the meantime, could you post the code so I could implement it temporarily on my own?

Similarly, is SpectrumPhase a possibility?

Does Spectrum have ~10 bar lag like DominantPeriod?

Thanks.

I can not post the code as it belongs to the licensed part of the software, but I've added DominantPhase to the list. A SpectrumPhase is not possible because the spectrum algorithm does not deliver a phase. The Spectrum lag is frequency dependent and in the same range as the DominantPeriod lag.

Thanks jcl.

Can the Spectrum code be posted or is it also in the licensed part of the software?

Thanks.

Also, do I understand correctly from your post above that a Spectrum curve is best used in aggregate form over some time period (like the example of an entire month)? Using Spectrum on a particular bar would be a problem as each frequency in the Spectrum will be a little "off" - shifted by increasing lag - and possibly thereby misleading? I.e., different Spectrum TimePeriods will actually refer to different parts of the price curve?

Thanks.