Mladen Rakic
216706
Mladen Rakic  

Maybe the question should have been is there a similarity between the two.

Fractal dimension index (originally developed by Carlos Sevcik A procedure to Estimate the Fractal Dimension of Waveforms) and corrected by Alex Matulich (if you take a source in basic from Sevcik's site, you are going to discover that it almost never passes 1.5 value, correcting that error and making faster code was done by Matulich) have a lot to do with Hurst coefficient ((2 - Fractal dimension index) = (Hurst coefficient)) but nothing to do with cfb

There was one early work from Mark Jurik called Fractal dimension (posted here since it is published on tradestation open source site (opened for anyone Knowledgebase)) and even that does not have anything to do with that (you can easily compare the math used)

So my answer is obvious : those are completely different things.

See the picture for comparison

Upper CFB, then Fractal dimension and the lowest is Fractal dimension index - Matulich version

regards

mladen

biddick:
So what is the difference between Eric Long's Fractal Dimension İndex Traders Tips - July 2003 and the Mark Jurik's Composite Fractal Behaviour ?Many thanks.
Files:
Linuxser
5921
Linuxser  
mladen:
Maybe the question should have been is there a similarity between the two.

Fractal dimension index (originally developed by Carlos Sevcik A procedure to Estimate the Fractal Dimension of Waveforms) and corrected by Alex Matulich (if you take a source in basic from Sevcik's site, you are going to discover that it almost never passes 1.5 value, correcting that error and making faster code was done by Matulich) have a lot to do with Hurst coefficient ((2 - Fractal dimension index) = (Hurst coefficient)) but nothing to do with cfb

There was one early work from Mark Jurik called Fractal dimension (posted here since it is published on tradestation open source site (opened for anyone Knowledgebase)) and even that does not have anything to do with that (you can easily compare the math used)

So my answer is obvious : those are completely different things.

See the picture for comparison

Upper CFB, then Fractal dimension and the lowest is Fractal dimension index - Matulich version

regards

mladen

Matulich already play us a good joke with T3. Remember our talks mladen?

Mladen Rakic
216706
Mladen Rakic  

:):)

In this case he did a good job

Even with T3 they (Fulks and Matulich) did write what they did and why, but somehow that got "lost in translation" and suddenly it became the "only" version of T3 for metatrader.

Well, we cleaned it up so,:D

Linuxser:
Matulich already play us a good joke with T3. Remember our talks mladen?
biddick
339
biddick  
mladen:
Me culpa, mea maxima culpa

There is one deviation in "my" version of cfb.

Since I do not like the "lack of smoothness" (cfb indicator does have what it calls a smoothing factor, but even with that it tends to be "bumpy").

And since we are looking for rising (trend is on) or falling (trend not confirmed, or reversing) values, I figured that some additional smoothing won't hurt.

So with a low lag additional smoothing it turned out to be rather useful.

As of using it : why don't you use it for rsx or similar indicators (as Jurik himself proposes - might be helpful) I am not sure if JMAs way of adapting would be better if cfb as additional "adaptor" (not the same though) was applied and with rsx it would make sense (faster signal).

PS: the "additionally non smoothed" version on the pic.

Jurik is suggesting, adjusting CFB to his indicators(RSX,DMI etc..) I did not read any suggestion of adjusting CFB to JMA just like you said .But in Ehlers FRAMA example ,Ehlers is using FDI to make AMA more responsive and faster.Since JMA is derived from AMA,so is this idea apllicable to JMA with FDI or CFB?Thanks.

krzysiaczek99
298
krzysiaczek99  

adaptive strategies

Here is a screenshot of 5 adaptive strategies using Jurik and Ehlers techniques

1) Ehler's Median Avg adaptive filter

2)JMA with CFB as a trend adaptor (SPAN=192)

3)JMA with HP as a trend adaptor

4)MAMA

5)MAMA & FAMA

all strategies were done on 5m EURUSD with optimization time 2 weeks and 1 day out of sample (last friday)

Buys/Sell signals were generated by crossing the MA with itself lagged by 1-4 bars except strategy 5 where cross between MAMA/FAMA was used. Lag period for strategies 1-4 were chosen by Genetic optimizer

Have a look to the results and trades, it is very clear that Jurik based strategies have tendency to overfitt, results during optimization are much better than during out of sample period so unwanted adaptation to noise occurs there.

Seems Ehlers strategies were much better

Comments are welcome how to overcome this.

Krzysztof

krzysiaczek99
298
krzysiaczek99  

screenschoot

Files:
adaptive_eu5.jpg  211 kb
Tamone
88
Tamone  

Jurik Files - Post 79

Perhaps someone could lend some assistance.... When I import the Jurik files...then Compile ... I get an error saying that

"'3c_BB_Osc.mqh' - cannot open the program file C:\Documents and Settings\Trader 1\Local Settings\Temp\wz9ccc\2 JJMASeries\JJMASeries\indicators\3c_JTRIX.mq4 (94, 1)"

I get the same error for all the files in MT4

Does anyone know why ... or how to correct this ????

[Deleted]  
montealex
5
montealex  

Color Coded Jurik Type Moving Average

Hi All,

I've been using a great Jurik type moving average to build various stochastic and MACD type indicators.

However, I am trying to color code the moving average so that when it trends up the indicator is colored white and when it trends down its colored red.

The guys over at Paint Bar Factory have a very similar indicator they call Fast2MA. I have attached an image of their indicator below so you all know what it is I'm trying to do. Also, Jurik's website has an image of his color coded JMA.

JMA demonstration

The one I am using I got off a board awhile ago. It's called JMA Starlight. Props to the author. I love it but I want to color code it so I can spot trend reversals easier. I think it could be a great help for everyone.

Can anybody help?

Thanks.

I have attached a copy of the indicator I want to color code. And several images for reference.

Files:
fast2mas.jpg  108 kb
fxbs
3203
fxbs  
monte.alex:
Hi All,

I've been using a great Jurik type moving average to build various stochastic and MACD type indicators.

However, I am trying to color code the moving average so that when it trends up the indicator is colored white and when it trends down its colored red.

The one I am using I got off a board awhile ago. It's called JMA Starlight. Props to the author. I love it but I want to color code it so I can spot trend reversals easier. I think it could be a great help for everyone.

I have attached a copy of the indicator I want to color code. And several images for reference.

how bad it repaint? (the one you've got, using, love and attached)