Fractals, ZZ & the Pissing Dog | Page 18 | 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

  • #341
  • Jan 27, 2014 4:20pm Jan 27, 2014 4:20pm
@ gg53
Very good,
to correct "errors of prediction", always use "Monty Hall"?
I have a problem in a zzleg.
The market gives me fractal pattern [0 1], I try to correct with Monty Hall so that my remaining combinations are:
[1 0] = 204 Occurs
[1 1] = 278 Occurs
[0 0] = 278 Occurs
[0 1] = 224 Occurs.

As I look for a [1], I take the pattern [1 1] and continuing the calculation, but the pattern with length 7, it gives me a wrong signal in place, in the middle of the zzleg.

Attached Image


Many thanks
  • #342
  • Edited 3:01am Jan 28, 2014 2:38am | Edited 3:01am
Quoting Thoughts
{quote} Thanks for explaining GG..your statistical approach towards RR sounds similar (in principle) to something I read in one of Victor Sperandeo's books..but I think he was looking at much longer time frames. I'm curious though, why not just trade 'always in' with the indicator?
Thoughts,

The target is to try and maximize the profit that you can get out of every $$$ of your fund.
"Always in" means that you are restricted to only one currency pair, which might not be the most potentialy profitable right now and might be even "caught" in a ranging market situation.

G.

Edit: But this is the first live run of this combined algorithms - so we'll have to see how effective it is.
There is no historical leverages and it should learn the numbers as it goes.
Let's see...
  • #343
  • Edited 3:48am Jan 28, 2014 3:35am | Edited 3:48am
On our way to "double within a month" (or not...)

First trade recommendation was short GBPAUD @ 1.88813 at ~London open.
The trade went Ok, BE was set at 8 pips.
BE was "eaten" and the trade went to ~+105 pips.
So - Good trade choice - bad trade management...
Result: +$34

G.
  • #344
  • Jan 28, 2014 4:02am Jan 28, 2014 4:02am
Quoting burnssss
@ gg53 Very good, to correct "errors of prediction", always use "Monty Hall"? I have a problem in a zzleg. The market gives me fractal pattern [0 1], I try to correct with Monty Hall so that my remaining combinations are: [1 0] = 204 Occurs [1 1] = 278 Occurs [0 0] = 278 Occurs [0 1] = 224 Occurs. As I look for a [1], I take the pattern [1 1] and continuing the calculation, but the pattern with length 7, it gives me a wrong signal in place, in the middle of the zzleg. {image} Many thanks
burnssss,

Two fractals does not qualify for "Monty Hall" / "3 Doors" implementation.

G.
  • #345
  • Jan 28, 2014 7:07am Jan 28, 2014 7:07am
Quoting gg53
{quote} Thoughts, The target is to try and maximize the profit that you can get out of every $$$ of your fund. "Always in" means that you are restricted to only one currency pair, which might not be the most potentialy profitable right now and might be even "caught" in a ranging market situation. G. Edit: But this is the first live run of this combined algorithms - so we'll have to see how effective it is. There is no historical leverages and it should learn the numbers as it goes. Let's see...
I see your point..I've got more of a traditional view on RR and MM, but with a slant I picked up from an old presumably Jewish trader. I'll definitely be watching the experiment.

Cheers.
  • #346
  • Jan 28, 2014 10:31am Jan 28, 2014 10:31am
Quoting gg53
On our way to "double within a month" (or not...) First trade recommendation was short GBPAUD @ 1.88813 at ~London open. The trade went Ok, BE was set at 8 pips. BE was "eaten" and the trade went to ~+105 pips. So - Good trade choice - bad trade management... Result: +$34 G.
On our way to "double within a month" (or not...) ,

Found a bug in this monster system:
It opened a BUY instead of BUY STOP...
That's an ouch!
Never mind, We'll catch up... (or so I hope...)

G.
  • #347
  • Jan 28, 2014 1:12pm Jan 28, 2014 1:12pm
Quoting gg53
{quote} On our way to "double within a month" (or not...) , Found a bug in this monster system: It opened a BUY instead of BUY STOP... That's an ouch! Never mind, We'll catch up... (or so I hope...) G.
Did you mean a sell stop?
  • #348
  • Edited 2:46pm Jan 28, 2014 2:31pm | Edited 2:46pm
@ gg53
I continued calculating the above example without changing the pattern [0 1] that the market gives me. But the signal appears in next bar of previous example.
How to evaluate these situations?. Some other statistical principle not yet seen?

Attached Image


Many thanks

EDIT:
A little reflection.
I remember in an old post, you said that we the length we had to look at (-1), current, (+1). It means that also employ Mnty Hall in length?.
If in predicting the pattern [0 1], the most likely predizco length is 6, but the market gives me a new bar, there is a problem of prediction in length. We have to solve this system too. But as Monty Hall implanted in length?

thank you very much
  • #349
  • Jan 28, 2014 2:48pm Jan 28, 2014 2:48pm
Quoting Thoughts
{quote} Did you mean a sell stop?
Thoughts,


No, a BUY instead of a BUY STOP on the GBPAUD. A stupid bug.

G.
  • #350
  • Jan 28, 2014 3:14pm Jan 28, 2014 3:14pm
Hi G,

Why trade so defensively (small BE pips) when the algo always gives profitable signals?

In your first trade (M5 TF), the entry wasn't the start of a down leg? You aren't using the algo's signals as the final arbiter?
Jan 28, 2014 10:08
Sell 1.88813

Thanks
Stinky
  • #351
  • Jan 28, 2014 3:17pm Jan 28, 2014 3:17pm
Quoting burnssss
@ gg53 Very good, to correct "errors of prediction", always use "Monty Hall"? I have a problem in a zzleg. The market gives me fractal pattern [0 1], I try to correct with Monty Hall so that my remaining combinations are: [1 0] = 204 Occurs [1 1] = 278 Occurs [0 0] = 278 Occurs [0 1] = 224 Occurs. As I look for a [1], I take the pattern [1 1] and continuing the calculation, but the pattern with length 7, it gives me a wrong signal in place, in the middle of the zzleg. {image} Many thanks
burnssss,

If everything is calculated properly there is nothing wrong with an "error" sometimes.
Continue calculating from that signal and if you don't get another "swing" signal - re-calculate from the actual ZZ once it is confirmed.
I don't have your code or database to confirm your results - but the signals (red arrows) up to that point looks like winners to me.
All 3 of previous signals are better trading entries than the fractals.

G.
  • #352
  • Jan 28, 2014 3:25pm Jan 28, 2014 3:25pm
Quoting gg53
{quote} burnssss, If everything is calculated properly there is nothing wrong with an "error" sometimes. Continue calculating from that signal and if you don't get another "swing" signal - re-calculate from the actual ZZ once it is confirmed. I don't have your code or database to confirm your results - but the signals (red arrows) up to that point looks like winners to me. All 3 of previous signals are better trading entries than the fractals. G.
@ gg53
ever had that kind of false signals?.
I use Monty Hall from 2 fractals, look long and if everything is correct because "next bar" prediction errors but I'm a little lost. If you've got signals how are you then accept the model. If you have this type of signals, yet still calculating means something wrong.
I would also like to know if you use Monty Hall in calculating length.
It would be very important because it is so maybe you could implement, see what happens.
thank you very much
  • #353
  • Jan 28, 2014 5:49pm Jan 28, 2014 5:49pm
Quoting burnssss
{quote} @ gg53 ever had that kind of false signals?. I use Monty Hall from 2 fractals, look long and if everything is correct because "next bar" prediction errors but I'm a little lost. If you've got signals how are you then accept the model. If you have this type of signals, yet still calculating means something wrong. I would also like to know if you use Monty Hall in calculating length. It would be very important because it is so maybe you could implement, see what happens. thank you very much
burnssss,

There is no logic in using "Monty Hall" or "3 doors" based on just 2 fractals. I'm using it from 3 or more.
I accept the model and continue calculating until I get a ZZ confirmation.
I don't use "Monty Hall / 3 doors" for length - for that I have the tolerance.

If you are using it from 2 fractals then your database is very different than mine.

G.
  • #354
  • Jan 30, 2014 3:39am Jan 30, 2014 3:39am
Hi everyone,

My name is Michael Lanner. I am new to forexfactory but I am following this and previous thread (Ultimate truth) with great interest for quite some time.

I did not finish this indicator yet but I'm close (I think).

@ GG53

Are you/Did you build this indicator for renko bars in the same way as time based bars ?

Thank you.

Kind regards,
Michael
  • #355
  • Jan 30, 2014 4:03am Jan 30, 2014 4:03am
Quoting Michael_Lnr
Hi everyone, My name is Michael Lanner. I am new to forexfactory but I am following this and previous thread (Ultimate truth) with great interest for quite some time. I did not finish this indicator yet but I'm close (I think). @ GG53 Are you/Did you build this indicator for renko bars in the same way as time based bars ? Thank you. Kind regards, Michael
Michael_Lnr,

Yep.

G.
  • #356
  • Jan 30, 2014 4:08am Jan 30, 2014 4:08am
On our way to "double within a month" (or not...) ,

Ok, now that we overcome the stupid bug (BUY instead of BUY STOP) let's see how this monster (?) behaves.
We wasted almost two days overcoming this - and it should catch up...

G.
  • #357
  • Jan 30, 2014 4:56am Jan 30, 2014 4:56am
Quoting gg53
On our way to "double within a month" (or not...) , Ok, now that we overcome the stupid bug (BUY instead of BUY STOP) let's see how this monster (?) behaves. We wasted almost two days overcoming this - and it should catch up... G.
Hi GG53,

Didn't you tell us something about bugs and long term relationship with customers.


BR.
Huub
It is not important how often you fall down, but how often you get up again
  • #358
  • Jan 30, 2014 5:28am Jan 30, 2014 5:28am
Hi,

I think my algorythm is performing pretty much the same as GG53's.

Here are the results for current m5 eur/usd.

Still having doubts about renko. Working on that.

Kind regards,
Michael
Attached Image (click to enlarge)
Click to Enlarge Name: EUR-USD-M5.PNG Size: 25 KB
  • #359
  • Jan 30, 2014 5:50am Jan 30, 2014 5:50am
Only problem I have is that I get 2 consecutive buy signals on today's chart. This happen to me more often. Need some more tweaking.

Kind regards,
Michael
  • #360
  • Jan 30, 2014 6:11am Jan 30, 2014 6:11am
Quoting Michael_Lnr
Only problem I have is that I get 2 consecutive buy signals on today's chart. This happen to me more often. Need some more tweaking. Kind regards, Michael
Hi Michael_Lnr,

great result. your algo is based on crodzilla probability calculation pdf or different? Are u handleing the error predicitions in your algo?

regards
0 traders viewing now