All John Ehlers Indicators... - page 23

fxbs
3203
fxbs  
igorad:
Hi,

I've developed MTF MAMA with Signals, Alerts & Warnings.

It's correct version of this indicator. All previous versions have bugs including Godzilla's version.

Regards,

Igor

thanks 4 mama, Igor

btw. what was the bug in other versions?

hb2
74
hb2  

i hope this work...

Igor Durkin
7832
Igor Durkin  
fxbs:
thanks 4 mama, Igor btw. what was the bug in other versions?

Godzilla's version bugs:

1. Q2 = Q1 - jI; should be Q2 = Q1 + jI;

2. if (period < 6.00*period) period = 6.00;

if (period > 50.0*period) period = 50.0;

should be

if (period < 6.00) period = 6.00;

if (period > 50.0) period = 50.0;

3. Function CountVelue doesn't work properly(try to test without it).

Other versions are repainting the past(ex. #MAMA).

krzysiaczek99
298
krzysiaczek99  

sinwave and others

Here is a screenshoot of test of different Ehlers indis with noised signal with

known content (2 cyclical components 20 and 100 bars+ gauss noise).

The results were not astonishing.

Cycle measurement was done with three methods.

Hilbert transform original Ehler code (white and blue)

Hilbert transform new Ehler code (red and yellow)

Corona chart methods

only corona shown proper period of 20 bars. Hilbert tranform method had an error measurement sometimes around 50% !!! even after smoothing with JMA.

Sinwave indicator was showing some cycles but timing with main signal was terrible

Swing position (corona charts) in some cases was showing proper swings and in other cases failing regardless of smoothing with JMA. It was showing the proper swings in area where cycle was measured properly i.e. 20 bars.

well I thought it will be better. Comments how to improve it are welcome.

Krzysztof

Files:
hb.jpg 231 kb
fxbs
3203
fxbs  
igorad:
Godzilla's version bugs:

1. Q2 = Q1 - jI; should be Q2 = Q1 + jI;

2. if (period < 6.00*period) period = 6.00;

if (period > 50.0*period) period = 50.0;

should be

if (period < 6.00) period = 6.00;

if (period > 50.0) period = 50.0;

3. Function CountVelue doesn't work properly(try to test without it).

Other versions are repainting the past(ex. #MAMA).

thank you again, and big thanks 4 saring your works and thoughts

fxbs
3203
fxbs  
hazelj80:
I am trying something similar to this in the m5 charts and it does work. I have custom setting that are from OANFX who was on FF but got booted. it uses this COG or whatever this indicator is called with custom settings, and psychological levels of whole/half numbers so if it's above the COG center line and COG is pointed down and breaks a whole/half number, go short and get out at the next whole/half number. stop at the last whole/half number. rues are opposite for sell. a few people are doing well. whole/half numbers work pretty good if you use them like this with custom COG settings

Scalping system #11 (The Center of Gravity) | Forex strategies revealed

Statement: 568858 - 100-1

trailing (tracking) sticked behind COG (polynormal, not JE); put on tester, enjoy the show

Igor Durkin
7832
Igor Durkin  

I've developed Instantaneous Trendline according to Ehlers article(in attachment), but instead of MESA I use Hilbert Transform.

Files:
krzysiaczek99
298
krzysiaczek99  

Hilbert transform

Yes, That's very good instant trend line but I think Hilbert transform is not accurate enough to drive it. I use this one but use Goertzel for Dominat Cycle measurement.

See screenshots whats happen with period measured with Hilbert in case of noise. I also found Ehler document about it which I believe he removed from his site....

Regarding screenshots. In all cases there is a signal 20 and 100 bars. Only in case of clean signal DC is recognised properly.

I think the best combination will be R-MESA from Richcap from digital filters thread to measure DC plus this instant trendline.

Krzysztof

Files:
h1.jpg 109 kb
h2.jpg 112 kb
h3.jpg 105 kb
krzysiaczek99
298
krzysiaczek99  

Here is how three instant trendlines with elliptic filter perform.

Green - driven by Goertzel

Black - driven by Hilbert period indicator

Blue - driven by Corona Period

Seems to be quite big differences between them. The most accurate theoretically is green one as cycle period is measured with the lowest error.

Krzysztof

Files:
3ofthem.jpg 166 kb
PhoenixPip
4
PhoenixPip  

Sinewave Support and Resistance

Hey guys does anyone know how to code a sinewave indicator for MT4 to plot support and resistance lines when the indicator crosses. Something similar to the Better Sinewave indicator available for Tradestation? If you google better sinewave e-mini I believe there is a video explaining what I mean. The indi looks rather useful so hopefully this could work on MT4 aswell. I have a couple of different sinewave indicators that I can post.