Fast Fourier Transform - Cycle Extraction - page 46

New comment
 
mladen:
Q_Mouze

I am not getting a zero divide error when I call it from another code. This is the code I used for testing :

#property indicator_chart_window

int init()

{

return(0);

}

int start()

{

double testValue = iCustom(NULL,PERIOD_M1,"Spectrometr_Separate_Selective Cycle 2.04",0,1);

Comment("Spectrometr test value : "+DoubleToStr(testValue,5));

return(0);

}

and it returns correct value (it shows the "unknown subwindow number for object creation function" but that is because it would need to skip object creation when called from some other code, but other than that there is no zero divide error). This is how it looks compared to the indicator values when I test it :

Thanks mr mladen.

I've gotten where lies the fault. I think there is in my MT4 history database that I have not updated.

because after I update, no error message "zero divide".

but how do I clear the message "unknown subwindow number for object creation function"

thanks b4.

 

Hit Again

SIMBA:
@Fadi Thanks for the EA,.

GBPUSD 1.6051-Number mentioned at one of my previous posts- Hit near exactly,this is basically the 195 FLD stopping the countertrend reaction.

Based on that FLD I project a target area at 1.5920-1.5935 or below,to be hit or exceededtomorrow October 23rd between 08:00 and 15:00 EST.

Regards

S

FLD Target area 1.5920-35 was hit at 10:30 EST,nice hit in price and time,then it was exceeded.

Next target is 1.5870-1.5890+- to be hit -probably- 13:00-17:00 EST today,in a few hours.

Then it is very probable that we will see a pullback or even the start of an uptrend.

Based on FLD 205 which is the current dominant cycle for GBPUSD h1.

Regards

S

 

...

You can use this one : it avoids the "unknown subwindow number for object creation function" when called from external code

Q_Mouze:
Thanks mr mladen.

I've gotten where lies the fault. I think there is in my MT4 history database that I have not updated.

because after I update, no error message "zero divide".

but how do I clear the message "unknown subwindow number for object creation function"

thanks b4.
 

There is a small correction on the EA settings towards the end set the MAPrice =0 which is the close.

Currently its setting to look for the high of the bar.

So in short this is how the EA works , and you can always add your comments or discussions about the EA on RR thread on FF

If you attach a moving average with MA_Period=1 , and you notice the shape of the moving average on the past 20 bars . The EA will look in the history for the past 100,000 bars for this shape and will report the top 10 shapes with their dates that have a pattern that is at least 90% correlated to the shape you are seeing ( 20 bars from the current bar) . The EA will then take the pattern with the highest correlation , record the move that happened in the past after this pattern then project this move in the current bar ( currently will project 10 bars a head ) . The EA can trade based on this move and thats where the , SLBuffer , TPBuffer , MinRR setting for.

Personally I don't trade this EA nor RR at least not now as its still a work in progress, however if you set the risk to 0 , then you can use the EA as an indicator for pattern mining.

MiniMe:
This is an EA, I worked on the concept with Ronald Raygun from FF

The EA is for pattern data mining .

Iif you are looking for a pattern and want to see how far it happened in the history then you can use this EA , IMHO a cycle is nothing but a collection of patterns.

You can defined the length of the pattern from the current bar and how far in history you want to look back. You can also filter the pattern "smooth it " by increasing the rank of moving average and in this way you will find more patterns.

This the link to RR post on FF

Pattern - Finding EA @ Forex Factory

I decided to post the EA here as its related to the topic but its up to tsd to move it if they want to
 

found a bit of info - as reading 1200pages of the Jim Hurst book maybe too much..

Hurst Cycles: Learn JM Hurst Trading Cycle Analysis Methods

Hurst Cycles: A Weekly Hurst Cycle Analysis of the SP 500 Index

 

Does FFT repaint or adjust the line in past ????

 
SIMBA:
FLD Target area 1.5920-35 was hit at 10:30 EST,nice hit in price and time,then it was exceeded.

Next target is 1.5870-1.5890+- to be hit -probably- 13:00-17:00 EST today,in a few hours.

Then it is very probable that we will see a pullback or even the start of an uptrend.

Based on FLD 205 which is the current dominant cycle for GBPUSD h1.

Regards

S

The ultimate target was a (near) miss.

If you check the attachment you will see how the target area was projected(Blue rectangle) using the 212 FLD(any FLD 190 or higher would have given the same price area result).the target area was punctured with the 1.59125 low,so the move may be finished.Alternatively we still have approx 17 hours( till 01:00 a.m fxcm server time)-see vertical red line-for the lower end of the target rectangle to be reached.In trading terms this means no trades,at least FLD based.GBPUSD has practically the same chance of retracing into 1.6000+- area than of completing its move into the rectangle,so,there is no FLD based edge.

NOTE:Cycles change periodicity,the now dominant 212 periods cycle started as 192,then 205,now it is at 212.This obviously affects the timing of the end of the move,but it doesn't affect the estimation of the target price area,at least by using the method shown in the attachment,so,that is why I do not use FLDs for timing but I do use them for estimating end of moves potential targets.

Regards

S

Files:
gbpusd_fld212_3.gif  75 kb
 

Fld histo

Hello all,

GBP USD : in Hourly chart.

Fld 24 --FLD-Momentum histo 12 === 1 day (24 Hours) cycle

Fld 48 --FLD-Momentum histo 24 === 2 days cycle

Fld 96 --FLD-Momentum histo 48 === 4 days cycle

Fld 192 --FLD-Momentum histo 96 === 8 days cycle

Files:
fld04.gif  17 kb
 
 
altoronto:
5min gbpusd - 24.10 around 9-10am - BUY

It went up ..

1...394041424344454647484950515253...112
New comment