Fractals, ZZ & the Pissing Dog | Page 11 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

ECN && STP && Scalping && Hedging - Broker Review - Oct 2009 24 replies

Alert on Fractals (MT) 2 replies

Zigzag & Fractals 10 replies

EA wanted for sound alert depending on fractals condition 2 replies

Fractals Edge Fx 3 replies

Attachments: Fractals, ZZ & the Pissing Dog
Exit Attachments

Fractals, ZZ & the Pissing Dog

  • #201
  • Jan 12, 2014 4:09am Jan 12, 2014 4:09am
@ gg53
very good, remember my post 188. Many mensajesen last hours and may have been overlooked.

Note:

"focus on the algorithm. carry 3-4 months and know how the system works and its good signs."
Many thanks
  • #202
  • Jan 12, 2014 7:04am Jan 12, 2014 7:04am
Hi All,

I only have a few things to say and I'll be as brief as possible.

First, I want to give a shout of thanks for the wonderful ideas being developed here. Being new to Forex and MT4, the "Factory" has been a real boon to achieving my trading goals, but this particular forum strikes me as the place where the real nugget (at least for me) will be found. I absolutely love the notion of developing algos to calculate probability. Rarely do I get this excited!

Second, I want to encourage those who differ with each other to learn to "reset" their orientation when the occasional misunderstanding or slip of the ego occurs. It's inevitable with online communication and after awhile, it won't matter who's actually "right" in the end. By then the issue would have been so convoluted and twisted the gains achieved won't be worth the pain and angst expended. Those in the gallery won't benefit much either.

Unfortunately it's really hard to know just when to cut it off and begin again. It's like taking a losing trade and being seduced into hanging on. How much better off would you have been to just cut it and begin again, even if it did eventually go your way? A little discipline and a willingness to inflict a very brief, momentary sting to the ego would have been a personal triumph had that loss been cut early. Instead, it went on and on and if it didn't ruin you completely and did go your way, you were willing to take but a tiny profit for all the pain endured. It's but the same problem with many faces.

This is an environment with very intelligent, Type A people from a wide variety of cultural backgrounds who are working on a very complex and powerful concept - a concept that could well change their lives. Couple those elements with the ease of miscommunication from online interaction and, well, there you have the 'perfect storm' for potential conflict. I want to encourage people to accept this difficulty and to realize it's unfair to expect a perfect, conflict-free environment with the kind of communication limits we face. People weren't designed to communicate in this way. We need facial expressions, vocal tone, body language and other subtleties to fully absorb what other people say. On top of that, the need for brevity is a recipe for frustration. It's naturally irritating and hence, all-too-easily it comes out in the printed word. It especially exacerbates the participation of the ego when played out in public.

Please, let's be a little more understanding and forgiving of our limitations. Let's accept them the best we can and see them for what they are. Maybe we can even privately handle our differences and return to the forum as friends. In my own life, the people I hated and who irritated me the most turned out to be my best friends and greatest benefactors but I never would have known that unless I took that first step.

Best to everyone,

Almuten
  • #203
  • Edited 8:14am Jan 12, 2014 7:06am | Edited 8:14am
About Total Length calculation:

Total length probability calculation is the next step after Fractals pattern match.
This step involves tolerance (+3/-3) and many PM'ed me about this - including Skype messages - so let me clarify.

Not all bars/candles are created equal - in some cases the bars can be very long, twice or more the average ones.
This leads to a ZZ end point appearing sooner than expected (less bars total length). The opposite situation is also true.
That's why we introduced the "tolerance" in calculating probabilities if current length is the "correct" length or should it be longer, and we must wait for another bar.

Example:
So we got our Fractal pattern, with total bar length of 10.
Now we want to know if this length is "correct" or should we wait for next bar.
We need to check the probability of length 7,8,9,10,11,12,13 against probability of length 14 (which means wait for another bar).

We are doing that by finding the maximum probability of 7-13 against probability of 14.
How to calculate probabilities is on crodzilla (http://www.forexfactory.com/crodzilla) Doc's - Post #1 (http://www.forexfactory.com/showthread.php?t=464948).

p7 = probability of length 7, p8 = probability of length 8, etc.

MAX(p7, p8, p9, p10, p11, p12, p13) > (p14).

If the above is true - our current total length (or bar count) is "correct".

G.
  • #204
  • Jan 12, 2014 7:44am Jan 12, 2014 7:44am
Quoting gg53
About Total Length calculation: Total length probability calculation is the next step after Fractals pattern match. This step involves tolerance (+3/-3) and many PM'ed me about this - including Skype messages - so let me clarify. Not all bars/candles are created equal - in some cases the bars can be very long, twice or more tha average ones. This leads to a ZZ end point appearing sooner than expected (less bars total length). The opposite situation is also true. That's why we introduced the "tolerance" in calculating probabilities if current length...
Thank you very much G.
Could you please give example on fractal probability calculation, if current zigzag still does not have any fractal ?

t168
  • #205
  • Jan 12, 2014 8:05am Jan 12, 2014 8:05am
we have the pattern [1 0] with length 10;

length---------------occurs
7----------------------10
8----------------------7
9----------------------6
10--------------------5
11--------------------5
12--------------------4
13--------------------4
14--------------------3

maximum P (7,8,9,10,11,12,13) = 10
P (7,8,9,10,11,12,13)> P (14)

Probability in length accepted

Is this correct??

Many thanks gg53
  • #206
  • Jan 12, 2014 8:10am Jan 12, 2014 8:10am
Quoting burnssss
we have the pattern [1 0] with length 10; length---------------occurs 7----------------------10 8----------------------7 9----------------------6 10--------------------5 11--------------------5 12--------------------4 13--------------------4 14--------------------3 maximum P (7,8,9,10,11,12,13) = 10 P (7,8,9,10,11,12,13)> P (14) Probability in length accepted Is this correct?? Many thanks gg53
burnssss,

The longest answer I can think of is: Yes.

G.
  • #207
  • Jan 12, 2014 8:14am Jan 12, 2014 8:14am
Another question,
in probability of next bar, you're using:
100% next bar
100% actuall bar

to get signal?

Thank you very much,
  • #208
  • Jan 12, 2014 9:33am Jan 12, 2014 9:33am
Am I the only one who considers the new length calculation formula in post #216 quite different to the "quasi-Gaussian" method first introduced in detail by Kilian, and then reaffirmed by Carl in the probability sheet?

Or are there 2 steps involved when length is evaluated (quasi-Gaussian distribution and the most probable total length)?

C.
  • #209
  • Jan 12, 2014 1:51pm Jan 12, 2014 1:51pm
Quoting candero
Am I the only one who considers the new length calculation formula in post #216 quite different to the "quasi-Gaussian" method first introduced in detail by Kilian, and then reaffirmed by Carl in the probability sheet? Or are there 2 steps involved when length is evaluated (quasi-Gaussian distribution and the most probable total length)? C.
Yeah, it seems different. I'm wondering if other things are different too.

In the example, do we only compare the -3 - +3 probability against length 14? Lengths 15, 16, 17, etc don't matter?
  • #210
  • Jan 12, 2014 2:03pm Jan 12, 2014 2:03pm
I don't think its different. Just put another way, I think.
Although the length (14) by itself is not what I understand (means longer?).

We want to know if the exact pattern match (with length tolerance - what I call my "in-length") has a higher probability of occurring than the longer fractal pattern (adding "0" and "1" - with "in-length").

And what we've said is, that if the longer fractal pattern occurs more often than the exact pattern match, but the length of the longer fractal pattern is shorter, then the longer fractal pattern is invalid... use the exact match with its in-length.

Carl
  • #211
  • Jan 12, 2014 5:22pm Jan 12, 2014 5:22pm
My understanding:

The length (14) is the next bar, the 4th bar from current. Previous +3 are within the tolerance window, that's why the "truly" next bar is the 4th.

Stinky, it is important to get the highest occurrence of length from the tolerance window (-3 +3), and then compare that to the occurence of length+4.

Well, to sum up - next indi versions in development, let the testing continue
  • #212
  • Jan 12, 2014 5:56pm Jan 12, 2014 5:56pm
I downloaded it and this indicators are repaiting... ;/
  • #213
  • Jan 12, 2014 6:37pm Jan 12, 2014 6:37pm
Quoting crodzilla
I don't think its different. Just put another way, I think. Although the length (14) by itself is not what I understand (means longer?). We want to know if the exact pattern match (with length tolerance - what I call my "in-length") has a higher probability of occurring than the longer fractal pattern (adding "0" and "1" - with "in-length"). And what we've said is, that if the longer fractal pattern occurs more often than the exact pattern match, but the length of the longer fractal pattern is shorter, then the longer fractal pattern is invalid......
Thanks. I guess they are similar, although not exactly the same. Maybe G implemented both, at different times, in different versions of the indicator.

By "we" do you mean both you and G?


@candero. Cheers, makes sense.
  • #214
  • Jan 12, 2014 9:49pm Jan 12, 2014 9:49pm
Quoting Stinky
{quote} Thanks. I guess they are similar, although not exactly the same. Maybe G implemented both, at different times, in different versions of the indicator. By "we" do you mean both you and G? @candero. Cheers, makes sense.
No, please understand... when I say "we"... it means us.
"we" = all of us going in the same direction toward the same goal.
G does not tell me any more than you already know.

I would like all of us to understand the correct way as a team.
I hope that all of us can have the same understanding how the algo should work.

Carl
  • #215
  • Jan 13, 2014 12:02am Jan 13, 2014 12:02am
Quoting crodzilla
{quote} No, please understand... when I say "we"... it means us. "we" = all of us going in the same direction toward the same goal. G does not tell me any more than you already know. I would like all of us to understand the correct way as a team. I hope that all of us can have the same understanding how the algo should work. Carl

exactly and i think that the reason why gg opened this thread if we keep it that as one then will be able to sail well.and thank you for the pdf.
  • #216
  • Jan 13, 2014 1:04am Jan 13, 2014 1:04am
Quoting go4it
GG53 - If you are able to can you please stop Magix making any further posts on your topic. It seams that he is only here to disrespect the thread and its members. Critical, probing and challenging discussion is fine, but puerile disrespectful rudeness and arrogance has no place anywhere on Forex Factory.
It is easy for Thread Starters to do this - just place a Ban on him !
  • #217
  • Jan 13, 2014 5:39am Jan 13, 2014 5:39am
Quoting vilas
{quote} It is easy for Thread Starters to do this - just place a Ban on him !
vilas,

Easy? Yes.
Not my style and I won't ban anyone from posting here. If it gets too ugly - I'll be the one who leaves.

G.
  • #218
  • Jan 13, 2014 5:42am Jan 13, 2014 5:42am
Quoting gg53
{quote}If it gets too ugly - I'll be the one who leaves. G.
Please don't let the very few haters win. We are all here supporting you (and your great team!) ... Keep up the astonishing work.
MT4/MT5 EAs/Indicators/Alerts coder
  • #219
  • Jan 13, 2014 5:54am Jan 13, 2014 5:54am
Quoting FerruFx
{quote} Please don't let the very few haters win. We are all here supporting you (and your great team!) ... Keep up the astonishing work.
I am agree. We want to keep learning.

thank you very much
  • #220
  • Jan 13, 2014 6:03am Jan 13, 2014 6:03am
@ gg53,
Focusing on the algorithm:

When using probability of "next bar", what percentage are using and how you use it?. According to the pdf shows a false signal, and not so well understood.

thank you very much
0 traders viewing now