Fractals, ZZ & the Pissing Dog | Page 31 | 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

  • #601
  • Apr 13, 2014 1:34pm Apr 13, 2014 1:34pm
7 , Now the ea applying some Frontline .
  • #602
  • Apr 16, 2014 9:28pm Apr 16, 2014 9:28pm
Quoting gg53
{quote} Mr_You, I really don't know what you are missing... Maybe the part about the "length" or "count" (total number of bars in zz leg) is not explained well enough and is missing from the docs. Lately it was brought to my attention. Sorry about that - I'm so deep in it that things seems obvious... Longer bars (pip wise) should be counted as 2-3 bars in the total zz leg bars count. The way I do it classify the bars by groups: Group1: bar size 1-5 pips = 1 bar Group2: bar size 6-10 pips = 2 bars Group3: bar size 11-15 pips = 3 bars etc... The numbers...

gg53,

I think I've found a potential discrepancy.

In the reply above you're referring to ZZ length in bars, but in your ZZ length prediction (http://www.forexfactory.com/showthread.php?p=7190111#post7190111) post you state the length is based on SET2 (the fractal set) which tells me the ZZ length is based on fractals.

Are we checking ZZ lengths against both bars and fractals? If so is there a particular order in the logic for when these occur? Could you clarify or correct this please?

Thank you very much!
  • #603
  • Apr 17, 2014 3:00am Apr 17, 2014 3:00am
Quoting Mr_You
{quote} gg53, I think I've found a potential discrepancy. In the reply above you're referring to ZZ length in bars, but in your ZZ length prediction (http://www.forexfactory.com/showthread.php?p=7190111#post7190111) post you state the length is based on SET2 (the fractal set) which tells me the ZZ length is based on fractals. Are we checking ZZ lengths against both bars and fractals? If so is there a particular order in the logic for when these occur? Could you clarify or correct this please? Thank you very much!
Mine (and I also think his) Set 2 isn't a fractal set but a set which holds all information of a ZigZag leg.

My set 2 looks like that

e.g.
Fractal Sequence 010101010 LengthValue(BarCount) 25 LegStartDatetime xxxxx LegEndDatetime yyyyy Direction 0
  • #604
  • Apr 17, 2014 7:18am Apr 17, 2014 7:18am
Quoting gg53
Creating SET2: From the same historical data on M5 (2011-2012) we now create our ZZ and Fractals dataset. we apply 2 standard MT4 indicators, Fractals and ZigZag (13,8,5) on that historical data. An MT4 script that writes the results to CSV will be posted later on. For EACH ZZ leg we store its fractal sequence and the total length (bar count) of that ZZ leg. UP fractal is marked as "1". Down fractal is marked as "0". So a record of each ZZ leg should look something like this: {10011} - fractal sequence within that ZZ leg 12 - total length (bar count) of the ZZ leg. This is also stored in a database, and this dataset is called from now on as SET2. G.
In dataset2 do we specify the direction of the zz leg with a "0" or "1"? As I don't recall gg stating that.
  • #605
  • Apr 17, 2014 7:36am Apr 17, 2014 7:36am
Quoting Madmoney
{quote} In dataset2 do we specify the direction of the zz leg with a "0" or "1"? As I don't recall gg stating that.
we don't but I believed it to be useful for further research.
  • #606
  • Apr 17, 2014 9:57am Apr 17, 2014 9:57am
Quoting Kilian19
{quote} Mine (and I also think his) Set 2 isn't a fractal set but a set which holds all information of a ZigZag leg. My set 2 looks like that e.g. Fractal Sequence 010101010 LengthValue(BarCount) 25 LegStartDatetime xxxxx LegEndDatetime yyyyy Direction 0
Ahh yeah I see now, that is what I'm doing too (without the time). Late nights are taking their toll. Doh!
  • #607
  • Edited 12:17pm Apr 17, 2014 10:28am | Edited 12:17pm
didnt want to hijack the thread so ...
nothing to see here :-)
  • #608
  • Apr 17, 2014 12:56pm Apr 17, 2014 12:56pm
Quoting Madmoney
{quote} In dataset2 do we specify the direction of the zz leg with a "0" or "1"? As I don't recall gg stating that.
What you use is completely up to you. For example I personally use:

0 = no fractal
1 = up bar/fractal
2 = down bar/fractal
3 = end of up ZZ / beginning of down ZZ
4 = end of down ZZ / beginning of up ZZ
  • #609
  • Apr 17, 2014 10:48pm Apr 17, 2014 10:48pm
Quoting Kilian19
{quote} Lovely said . For me the most tricky part which hasn't been spoken off is from where to start the prediction and how to indentifiy a false prediction. If my last prediction was not correct I might be missing a few fractals of the next legs etc ...
I'm still working on it, but I think I just had a big realization. Please correct me if I'm wrong.

Run a prediction check when you get a ZigZag high/low signal (each ZigZag draw). This will also tell you the current ZigZag direction and the direction you're trying to predict. As gg53 mentioned, if you're counting the fractal at the end of the ZigZag you will need to wait two bars for the fractal to occur. Take this information into consideration when "error checking" (checking for false prediction).
  • #610
  • Apr 18, 2014 9:31pm Apr 18, 2014 9:31pm
Hallo masta how are you , are youdoing well ? miss you here .



K
  • #611
  • Edited 6:47am Apr 19, 2014 5:33am | Edited 6:47am
Fractals, ZZ and the Renko bars (and don't forget the Pissing Dog...)

The Renko version of the ForexGT_Arrow EA is finished (incl. Currency-Strength scanner and other supporting filters) .

A major problem was the Backtesting of the Renko chart - which is always a problematic issue on almost all trading platforms.
I think that we solved that issue too, and soon I'll post backtests results with at least 90% accuracy.
Solving the Renko backtesting issue is a major cause of happiness here.

Currently, the results of ForexGT_Arrow (Renko Ver.), EURUSD (Renko block size 10 pips), for the last 2 month backtest showing:
Parameters:
2% Risk ($10,000 Acct.) - 1Lot, 20 pips SL, 20 pips Trailing, Always in.

Results:
Net Profit - $4833
Expected Payoff - 24.79
Maximal Drawdown - 13.61%
Profit Trades - 43.59% (!!!)
Avg. Profit - $191
Avg. Loss - $103

Outcome:
48.33% Profit, 1.85/1 R:R
In 2 Month...

Of course it's a very short period to consider as indicative, but it's a promising direction.
The major thing to consider now is the LOW profit trades. We need to analyse the loosing trades and reduce them.
More work...

G.
  • #612
  • Apr 19, 2014 11:33pm Apr 19, 2014 11:33pm
Quoting gg53
{quote}The ZZ end-point is NOT my target. My target is the most probable trading SWING point. It's true that many times they are same but often it's not - because ZZ end-point is the HIGH or LOW of the bar and not its CLOSE. So looking for the "next bar in opposite direction" will give you different result than ZZ end-point - and it's not a "miss" or "error" as long as the trading entry point is better or equal or at least very near the ZZ end-point. G.
I apologize if this has already been answered. Just point me there.

This seems to be my issue. Do you have any advice or are you able to describe how you go about looking for matches around the ZZ end-point?

Or I guess another way to put it. If I have a ZZ match is there a number of bars previous and after that I should be looking for a SET3 match?

Thanks!
  • #613
  • Apr 19, 2014 11:52pm Apr 19, 2014 11:52pm
Quoting gg53
Fractals, ZZ and the Renko bars (and don't forget the Pissing Dog...) The Renko version of the ForexGT_Arrow EA is finished (incl. Currency-Strength scanner and other supporting filters) . A major problem was the Backtesting of the Renko chart - which is always a problematic issue on almost all trading platforms. I think that we solved that issue too, and soon I'll post backtests results with at least 90% accuracy. Solving the Renko backtesting issue is a major cause of happiness here. Currently, the results of ForexGT_Arrow (Renko Ver.), EURUSD...
Hi G, If you strip out money management (appears to be 10x leverage), how many pips did it make during these 2 months? About 480? Is this better than the average pip production on minute based charts?
  • #614
  • Apr 20, 2014 4:11am Apr 20, 2014 4:11am
Quoting Mr_You
{quote} I apologize if this has already been answered. Just point me there. This seems to be my issue. Do you have any advice or are you able to describe how you go about looking for matches around the ZZ end-point? Or I guess another way to put it. If I have a ZZ match is there a number of bars previous and after that I should be looking for a SET3 match? Thanks!
Mr_You,

I'm not looking for ZZ match.
Just fractals pattern & bars count - to match in probability calculations.

G.
  • #615
  • Apr 20, 2014 4:20am Apr 20, 2014 4:20am
Quoting FXEZ
{quote} Hi G, If you strip out money management (appears to be 10x leverage), how many pips did it make during these 2 months? About 480? Is this better than the average pip production on minute based charts?
FXEZ,

Leverage on the acct is 1:100 (Std.)
EURUSD=$10/pip
Right, it's 483 pips.
From start date till end date - the pair made 128 pips.
From start date to highest close during that period - the pair made 252 pips.

G.
  • #616
  • Apr 20, 2014 4:33am Apr 20, 2014 4:33am
@gg53

In an ideal world the direction for nextbar prediction is determined from last arrow drawn.
But what about the first time you attach the indicator on your chart. No arrow to "check".
Can you please tell how you determine the direction for this cause.

Thanks in advance
  • #617
  • Apr 20, 2014 4:58am Apr 20, 2014 4:58am
Quoting Etecy
@gg53 In an ideal world the direction for nextbar prediction is determined from last arrow drawn. But what about the first time you attach the indicator on your chart. No arrow to "check". Can you please tell how you determine the direction for this cause. Thanks in advance
start somewhere 2 legs in the past at an zz end point and go in the direction of the swing
  • #618
  • Apr 20, 2014 5:27am Apr 20, 2014 5:27am
Okay thanks Kilian19.

next question for @ gg53 :-)

Upper half of next screenshot shows the predicted arrows.
Second half shows ZZ end points that are not repainted.

Attached Image (click to enlarge)
Click to Enlarge Name: Untitled-1.jpg Size: 417 KB


Question 1:
The first 2 aqua arrows are "errors" because they are more than 3 bars after the MT4 ZZ?

Question 2:
If so we count the fractals and bars from last red arrow or green (latest confirmed ZZ high or latest confirmed ZZ low)?

Thanks in advance
  • #619
  • Apr 20, 2014 6:18am Apr 20, 2014 6:18am
Quoting Etecy
@gg53 In an ideal world the direction for nextbar prediction is determined from last arrow drawn. But what about the first time you attach the indicator on your chart. No arrow to "check". Can you please tell how you determine the direction for this cause. Thanks in advance
Etecy,

The algo starts from last confirmed ZZ leg end-point (Std. ZZ indie).

G.
  • #620
  • Apr 20, 2014 6:25am Apr 20, 2014 6:25am
Quoting Etecy
Okay thanks Kilian19. next question for @ gg53 :-) Upper half of next screenshot shows the predicted arrows. Second half shows ZZ end points that are not repainted. {image} Question 1: The first 2 aqua arrows are "errors" because they are more than 3 bars after the MT4 ZZ? Question 2: If so we count the fractals and bars from last red arrow or green (latest confirmed ZZ high or latest confirmed ZZ low)? Thanks in advance
Etecy,

1. No, because there was no confirmed ZZ near by, so the algo continues as usual.
Once second signal is activated, there will be no "error" checking - because it will cause repainting.

2. We always count from the arrow - except in cases of "error" and the leg is not finished yet by another signal.

G.
1 trader viewing now