This is the development community for the indicator that was discussed in great detail at:
http://www.forexfactory.com/showthread.php?t=464948
Please upload documents such as notes from this thread that you think are important and code samples that you have.
GitHub: https://github.com/PenguinTraders/FractalArrowPrediction
Instructions: http://penguintraders.com/groups/fractal-arrow-prediction/forum/topic/installation-getting-started-guide/
Group Admins
How to: Creation of Fractal Arrow Indi
This topic has been reported for inappropriate content
- This topic has 86 replies, 14 voices, and was last updated 9 years, 6 months ago by
.
- Posts
post is obsolete.
Hi all Posting two charts of my working/not working prediction. In the first Picture you can see two working signals as an example. In the second Picture you can see where my algo has problems. As soon as the “swing” or better the amount of bars between to fractals are longer as usually i get to early signals. The combination of pattern match and length match seems to be my problem. I think my pattern match is wrong or maybe even what i compare in my pattern match. Suggestions how to solve this? @GG53 @Saver0 @smallcat @Emonts @jimsterk @VlanFx
Sorry for late reply.
Maybe you should run your indicator for longer period to see if it is “profitable or not”. (not if it is “exact or not”)
What I realized was that sometimes my indicator gives good result (almost perfect) but not other days.
And then if I tweak the code to attempt to give perfect predictions for longer period, the indicator deteriorate.
Doing this over and over, now I have no idea what I can do.
Hope this will help you.
Thank you,@jimsterk
What do you mean with “the indicator deteriorate.”
Have you tried syncing it everytime with a confirmed ZZ?
What do you mean with “the indicator deteriorate.”
Sorry for my poor english. It means the indicator lose accuracy of prediction.
Have you tried syncing it everytime with a confirmed ZZ?
Not remember clearly because I have not been working this indicator for 6 months, but I think yes.
What do you mean with “the indicator deteriorate.”
Sorry for my poor english. It means the indicator lose accuracy of prediction.
Have you tried syncing it everytime with a confirmed ZZ?
Not remember clearly because I have not been working this indicator for 6 months, but I think yes.
What an “accurate” statement by someone that doesn’t have a working indicator… ;-}
Just an info on current ForexGT_Arrow indicator status:
Same old DB, same parameters, same accuracy…
My invitation for live screen-sharing via Skype is still open…
Now it is combined with Currency-Strength, Volume and Correlations into an EA, working on full-auto for almost two month. Some of you already saw the results.
G.
I am waiting for a reply from GG on Skype what to do with the ZigZags/Swings that miss the valid Fractal by one bar. In my opinion if they miss the valid Fractal even by one bar, the aren’t valid
A miss by one bar is perfectly Ok, and need no changes or re-calculation.
Re-read the FF thread on the subject. 2-3 bars is considered an error that requires DB modification.
G.
This wasn’t exactly what I meant but your answer reminds me to say that you have to be really carefull when creating the database. Even a tiny mistake can screw up everything.
With the help of VlanFX it turned out that the ZZ indicator I was using produced these misses.
Has anyone else had any success with this?
I stopped work a while back and most recently I guess I misunderstood @gg53 in another post because I thought he said this/”the pissing dog” strategy had failed. Maybe it’s worth looking at again?
Maybe it’s worth looking at it again?
What do you expect from this question? Why do you need someone else to tell you what to do?
Maybe it’s worth looking at it again? What do you expect from this question? Why do you need someone else to tell you what to do?
I think I explained that in my post: I was under the impression that this indicator/strategy had fallin apart, but @gg53 latest comment says otherwise. So I’m not asking someone to “tell me what to do”, I’m asking if anyone can clarify whether this is worth spending (again) weeks/months of work. Because up til now this topic has been dead.
But thanks for your helpful comment in regards to my original question.
EDIT: FYI, I have/had already started working on it again because I had some new insights I want to explore. Also up to this point, AFAIK, no one has come out publicly to say they have been able to duplicate @gg53 results.
@MrYou This is how far I am at the moment…….
@MrYou This is how far I am at the moment…….
Looks good and on the correct path !
Such progress deserves one “extra” tip:
Have you thought about “reversing” the function?
Example:
Let’s say that you have a function that draws MA according to PA, and the current direction of that MA is UP.
Can you “reverse” that function and tell at what FUTURE price that MA will start to point DOWN?
The answer is YES!
Take that thought to the Fractals ARROW algo…
G.
putting the huge thinking cap on……………..
@gg53 “reversing” the moving average is easy to understand. But reversing my algo……. I don’t even understand which goal this serves. Obviously it is to make my algo better, but how……?
My algo looks at every bar and checks this against the database. “Reversing” this would mean that I look into my algo before I get a bar and look where the first possible swing point could appear. Or reversing could mean that I look at my current signal and start a second calculation backwards to look if I get a signal at my previous signal/swingpoint. If not then my current signal is probably wrong.
@gg53 “reversing” the moving average is easy to understand. But reversing my algo……. I don’t even understand which goal this serves. Obviously it is to make my algo better, but how……? My algo looks at every bar and checks this against the database. “Reversing” this would mean that I look into my algo before I get a bar and look where the first possible swing point could appear. Or reversing could mean that I look at my current signal and start a second calculation backwards to look if I get a signal at my previous signal/swingpoint. If not then my current signal is probably wrong.
Leave the DB alone…. {-;
What if you add to your algo a Direction-Change “prediction”, based on the A/M example?
You already know at what price the direction is going to change – add this to the probabilities.
G.
@gg53 what do you mean with A/M?
OK…I feed my database with the input of the bars/fractals, do some “calculations” and I get a signal. But I should leave my database alone for the “reverse”. But everything I do with my algo needs the database.
Sorry, I still don’t have a clue how and why I should reverse the algo
@gg53 what do you mean with A/M? OK…I feed my database with the input of the bars/fractals, do some “calculations” and I get a signal. But I should leave my database alone for the “reverse”. But everything I do with my algo needs the database. Sorry, I still don’t have a clue how and why I should reverse the algo
A/M = Above Mentioned…
I didn’t say anything about changing the database, nor about “reversing” your algo.
Just add another parameter to your algo that influence your prediction based on direction change.
G.
That means I can use a small moving average in “reverse” to get the price where the moving average changes its direction and then i can limit my search in my database to every swing that has the “reversed” price. The only thing I need to add to my database for this is the price of the bar that comes after the end of my leg/zz
That means I can use a small moving average in “reverse” to get the price where the moving average changes its direction and then i can limit my search in my database to every swing that has the “reversed” price. The only thing I need to add to my database for this is the price of the bar that comes after the end of my leg/zz
Leave the Database as is…
This should be done in real-time, by the algo, otherwise it will be too late and on the bar CLOSE, while doing it in real-time it can be done immediatly.
G.
Ah now I think I got it. I calculate the “reversed” MA and when my algo gives me the signal but the price needs to fall to much on the next bar for the MA to change direction I don’t give a signal. Or the other way around. The price doesn’t need to fall much for the market to change the direction of the MA so we probably don’t have a swing point or the market is flat.
Ah now I think I got it. I calculate the “reversed” MA and when my algo gives me the signal but the price needs to fall to much on the next bar for the MA to change direction I don’t give a signal. Or the other way around. The price doesn’t need to fall much for the market to change the direction of the MA so we probably don’t have a swing point or the market is flat.
The MA was just an example, but you can “predict” many other indies behaviour if you “reverse” their function.
G.
@MrYou This is how far I am at the moment…….
Looks good and on the correct path ! Such progress deserves one “extra” tip: Have you thought about “reversing” the function? Example: Let’s say that you have a function that draws MA according to PA, and the current direction of that MA is UP. Can you “reverse” that function and tell at what FUTURE price that MA will start to point DOWN? The answer is YES! Take that thought to the Fractals ARROW algo… G.
this is interesting G. Could you please give example how to predict future (reverse) price based on MA? Thanks in advance G.
…if only someone would knew the solution to solve the problem of signals that are one or sometimes even two bars to early
@MrYou This is how far I am at the moment…….
Looks good and on the correct path ! Such progress deserves one “extra” tip: Have you thought about “reversing” the function? Example: Let’s say that you have a function that draws MA according to PA, and the current direction of that MA is UP. Can you “reverse” that function and tell at what FUTURE price that MA will start to point DOWN? The answer is YES! Take that thought to the Fractals ARROW algo… G.
this is interesting G. Could you please give example how to predict future (reverse) price based on MA? Thanks in advance G.
You can’t predict future price.
But you can predict at what future price the MA (just an example) will start to turn or will go flat.
Same goes to RSI, Stochastic, CCI, StdDeviation, etc.
G.
The MA was just an example, but you can “predict” many other indies behaviour if you “reverse” their function. G.
What a profound idea…
Nothing has ever motivated me more than this...
- You must be logged in to reply to this topic.