Fourier Transform (How to design more adaptive indicators )

[Grey1]

The subject of measuring the dominant cycle and the settings for indicators is not some thing new,, In fact if you go back to some of my posts in 2000 you can see why I have always been against those who are after a meaningful trading system using a fixed cycle based TA indicators such as RSI or similar.

Those of you who are interested in the design of more advanced indicators might find this article interesting ,

Grey1

 

[Grey1]

The subject of measuring the dominant cycle and the settings for indicators is not some thing new,, In fact if you go back to some of my posts in 2000 you can see why I have always been against those who are after a meaningful trading system using a fixed cycle based TA indicators such as RSI or similar.

Those of you who are interested in the design of more advanced indicators might find this article interesting ,

Grey1

opps the article

 

[Grey1]

more

 

[grantx]

Grey1,

I have often believed that static indicators need to be constantly re-adjusted to reflect ever the changing nature, ie volatility, of the markets.

For the non-mathematical (me) these transforms seems an awful lot work. By contrast, assume you're using 5- and 10-day ma's for to generate buy and sell signals.

Their effectiveness will vary on a daily basis. Therefore one simply needs to adjust accordingly by entering, eg 4- and 9-day ma's on the previous day's data to see if the signals have improved. Through trial and error, more appropriate indicators will be found. However, for those with multiple screens/multiple markets/mutiple charts it could be be somewhat laborious.

Grant.

 

[Grey1]

Grey1,

I have often believed that static indicators need to be constantly re-adjusted to reflect ever the changing nature, ie volatility, of the markets.

For the non-mathematical (me) these transforms seems an awful lot work. By contrast, assume you're using 5- and 10-day ma's for to generate buy and sell signals.

Their effectiveness will vary on a daily basis. Therefore one simply needs to adjust accordingly by entering, eg 4- and 9-day ma's on the previous day's data to see if the signals have improved. Through trial and error, more appropriate indicators will be found. However, for those with multiple screens/multiple markets/mutiple charts it could be be somewhat laborious.

Grant.

Grant,

Moving averages do not reflect the market cycle how ever OB/OS such as RSI,CCI and others do ,,As a result most Buy LOW sell High Indictors use OB/OS oscillators to calculate the top and the bottom and not any MA.

You are spot on about INDICATOR'S effectiveness based on daily volatility ( or even more accurate " the daily dominant cycle " ) , and hence the more adaptive cycle based indicators such as those designed by Dr Ehler.

These indicators are not difficult to design and in fact you can copy and paste the code into TS( they are all over the internet ) if you search for them .The message as you correctly point out is this ,, Technical indicator's settings need to use market cycle as a variable to be off any good other wise they are as daft as a broom .

In the seminar We did on friday Frank explained how to use previous cycle NODE to identify a possible turn around on the future TOP or BOTTOM correctly .


Grey1

 

[grantx]

Grey1,

Fair enough, point taken.

Funny you should mention CCI. Only today I read an article which mentioned this. I've placed it on a chart and it will be interesting to see if makes a contribution.

"daft as a broom". Excellent.

Grant.

 

[Sirlosealot]

I've tried loading the "EasyLanguage Code to Display the Spectrum Derived from a Filter Bank" into my Tradestation 2000i but get errors when I try to verify it. It does not recognise "RGB" in the "Plot" statements and does not recognise "Plot" statements greater than "Plot4". I am kind of assuming that this is becausec these reserved words have been added to Tradestation 8 and I will have to upgrade. Does anyone know if this is correct.

Gerard

 

[MarkB]

Hi Sirlosealot,

I'm sorry I don't have an answer to your question but if anyone else wants to try these indicators I have uploaded them. It will save you the trouble of cutting and pasting and tidying up.

 

[Trader333]

Gerard,

TS2Ki is limited to a maximum of four plots and because of this then spectral analysis is not possible. I did try this myself a couple of years ago but I couldn't get it to work.


Paul

 

[Sirlosealot]

Paul & MarkB

Thanks for the replies. I've just received my march Tech Analysis of Stocks & Comm and notice that eSignal have uploaded versions at:

www.esignalcentral.com/support/kb/efs

The efs files are named "ChRSpectrum.efs" and "DCTunedBypassFilter.efs".

eSignal say "Both studies contain formula parameters that may be configured through the Edit Studies option in the Advanced Chart to change the color and thickness of the indicators. The spectrum study also has a parameter to turn on/off the dominant cycle display"

I haven't downloaded them yet and will be too busy and away from broadband access till late Friday but I'll try and have a look then -providing eSignal is working properly.

Gerard

 

[Ultramarine]

Paul & MarkB

Thanks for the replies. I've just received my march Tech Analysis of Stocks & Comm and notice that eSignal have uploaded versions at:

www.esignalcentral.com/support/kb/efs

The efs files are named "ChRSpectrum.efs" and "DCTunedBypassFilter.efs".

eSignal say "Both studies contain formula parameters that may be configured through the Edit Studies option in the Advanced Chart to change the color and thickness of the indicators. The spectrum study also has a parameter to turn on/off the dominant cycle display"

I haven't downloaded them yet and will be too busy and away from broadband access till late Friday but I'll try and have a look then -providing eSignal is working properly.

Gerard

Hi

I've installed and run them. Seems to look like what's shown on the can - but interpreting the output is taking me a bit of time!

Al

 

[belflan]

Hi

I've installed and run them. Seems to look like what's shown on the can - but interpreting the output is taking me a bit of time!

Al

Can these codes be used with TS2ki I've try to cut and paste the codes.. doesn't seem to like it?

i take it you have to change EFS codes to run in TS2ki?

belflan

 

[belflan]

Can these codes be used with TS2ki I've try to cut and paste the codes.. doesn't seem to like it?

i take it you have to change EFS codes to run in TS2ki?

belflan

I've typed the code into EasyLanguage PowerEditor and it passes ok

however when try to open it within TS2ki it gives this error. i've tried to re write the code to suit, without success.

anyone had the same problems? can anyone help

cheers
belflan

 

[iq200]

beflan, Its basically saying that you have a divide by zero. this is obviously coming in from either the line where you calculate Value1 (possibly MaxH equals MinL) or where you calculate Fish (possibly Value1 equals 1) or maybe both! Comment out the line where you calculate Fish and see if the error still comes up. If it doesnt then it means the Fish calculation is causing the problem. Then make the correction similar to what the error dialog suggests.
I havent looked at the code logic in great detail but one other error I saw is that you are making a check on Value2 but never actually setting the value of it anywhere. That is most likely your error.
Cheers,
Imran

 

[belflan]

beflan, Its basically saying that you have a divide by zero. this is obviously coming in from either the line where you calculate Value1 (possibly MaxH equals MinL) or where you calculate Fish (possibly Value1 equals 1) or maybe both! Comment out the line where you calculate Fish and see if the error still comes up. If it doesnt then it means the Fish calculation is causing the problem. Then make the correction similar to what the error dialog suggests.
I havent looked at the code logic in great detail but one other error I saw is that you are making a check on Value2 but never actually setting the value of it anywhere. That is most likely your error.
Cheers,
Imran

Thanks Imran,

will try this

cheers, belflan

 

[Grey1]

Thanks Imran,

will try this

cheers, belflan

I wonder if some one could help me,, My Tradestation has a GAP in the data because of the MONDAY holiday and it is affecting my analysis, I used to know how to fix it but some how cannot do it ..

any idea

thanks

 

[Glenn]

I wonder if some one could help me,, My Tradestation has a GAP in the data because of the MONDAY holiday and it is affecting my analysis, I used to know how to fix it but some how cannot do it ..

any idea

thanks

View - Chart Options
Under the Symbol tab -
Uncheck whichever of these two you want:-
Show empty Daily trading periods
Show empty intraday trading periods.

Glenn

 

[Glenn]

hi

is that in global server ?

thx

No sir, it's in TS. There is no such option in GS
Glenn

 

[Ultramarine]

Can these codes be used with TS2ki I've try to cut and paste the codes.. doesn't seem to like it?

i take it you have to change EFS codes to run in TS2ki?

belflan

Sorry for providing no response but i've been on holiday for the last week and have only just seen your post. Not that I can be much help. I use esig as my data provider so just ran the EFS codes on their charting package. I think that Trader333 has posted somewhere about incompatibility between TS2ki code and later versions of TS. Hope you have had some success in the meantime and will be interested to hear any views on the usefulness of the analysis.

Al

 

[belflan]

Sorry for providing no response but i've been on holiday for the last week and have only just seen your post. Not that I can be much help. I use esig as my data provider so just ran the EFS codes on their charting package. I think that Trader333 has posted somewhere about incompatibility between TS2ki code and later versions of TS. Hope you have had some success in the meantime and will be interested to hear any views on the usefulness of the analysis.

Al

na, played around with it for a while, not much luck,,.

will try again when i get some time

ta
belflan