Fractals, ZZ & the Pissing Dog | Page 34 | 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

  • #661
  • Edited 7:09pm May 4, 2014 6:41pm | Edited 7:09pm
Quoting gg53
{quote} giovanni4000, As I said, building the database is based on the ZZ. I'm not moving it an inch. I count the bars, register their OHLC, count the fractals in between and their spaces, and add one fractal at the top (for Long move) or bottom (for Short move). That's it. I'm not moving anything or changing positions. So, to answer your question based on your chart: the ZZ's are ABDEG. AB - one fractal (top) BD - 7 fractals DE - 5 fractals EG - 3 fractals G.
Hi G,

Great, thank you very much. Can you explain this :

Quoting gg53
count the fractals in between and their spaces
We need to put not only the fractals pattern and the zigzag length (bars counting), but also the "spaces" ....
does it mean, we need to put the candle up/down information also beside the fractal it self ?
Please see the image below. Does it mean, from A to B, we must put to database the fractals + spaces like this ( but without brackets "{}" ) :
{0} 0 {1} 1 {00} 1 {00} 0

Another question:
As we want to do a "fractal match" calculation, and at that time there is no fractal exist, so we must compare the current
bars with the "spaces" (exactly "1" / "0" bars) that exist in front of the first fractal (compare to zz leg that has only exactly 1 fractal in the zz leg), from the database ?
Example:
*) We now at B-C zigzag leg.
At the first time, there is no fractal (bar number 1), and it must be a fractal at the end of current zigzag leg. Because we are at "up" going zigzag leg, so we are waiting for a "1" fractal at the end of B-C zigzag leg.
*) We are now at second bar, and we have bars : 0 in current zigzag leg, so we compare to database with {0} bar in zz leg that has only 1 fractal. But there is no such pattern in database, we are waiting.
*) Now we are at 3rd bar, and have bars : 01 , then we compare to database with {01} bars in the zz leg that has only 1 fractal. But from database, we get {010} in zz leg with only 1 fractal has more probability. So, we wait again.
*) And then we get bars : 011 , and still have no fractals. So, we compare to database again, looking for bars {011} in the zz leg that has only 1 fractal, so the bars pattern {010} is a wrong prediction. And from database, we get the more probability is {0110}. We are waiting again ...

Is it like that ? Or .... how do you calculate the "fractal match" probability if there is no fractal at all in current zigzag leg ?

*) And the same thing, if we get 1 fractal already, we compare to database with zz leg that has 1 or 2 fractals in it. In other words: comparing current fractal to "+1" / "+0". Example: we have "0" fractal now. We must compare to database, looking for more probability of zz leg that has "0" with "00" and "01" fractals.

Thanks in advance.
Gio
Attached Image
  • #662
  • May 5, 2014 3:36am May 5, 2014 3:36am
When he says count the spaces and fractals he's referring to the ZigZag length.
  • #663
  • May 5, 2014 11:05am May 5, 2014 11:05am
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,

The pips of bars you take from the HIGH - LOW of the each bar ?

Tq
Similarity can change your life !
  • #664
  • May 6, 2014 6:23am May 6, 2014 6:23am
Quoting Mr_You
When he says count the spaces and fractals he's referring to the ZigZag length.
Thank you mr_You. So, we do not need to store the candle bull/bear information , but just need to count it ...

-Gio
  • #665
  • May 6, 2014 12:39pm May 6, 2014 12:39pm
Quoting jackprobe
{quote} GG53, The pips of bars you take from the HIGH - LOW of the each bar ? Tq
jackprobe,

Since the system is fractals-based, and fractals are based on high/low, I follow the same logic.
When comparing "large bar" to avg previous ones - I use the high-low measurement of the bars.

G.
  • #666
  • May 6, 2014 7:17pm May 6, 2014 7:17pm
Quoting gg53
{quote} jackprobe, Since the system is fractals-based, and fractals are based on high/low, I follow the same logic. When comparing "large bar" to avg previous ones - I use the high-low measurement of the bars. G.
Thank you GG53,

May i know how many previous bars do you take to calculate the "average" ?
Do you take the whole candles from 2011 until now, get the "height" (high-low) of each bar, and calculate the average ?
Example: the amount of candles from 2011 to now are 250.000 candles. So we do calculation of the average like this:

Inserted Code
for (int i=0; i<=250000; i++) {
    totalHeight += (  ihigh(xxxxxx,i) - ilow(xxxxx,i)  );
}
average = totalHeight / 250000;

Or we just take the candles from previous (last) zigzag leg only ?

Tq
Similarity can change your life !
  • #667
  • May 6, 2014 11:08pm May 6, 2014 11:08pm
Quoting jackprobe
{quote} Thank you GG53, May i know how many previous bars do you take to calculate the "average" ? Do you take the whole candles from 2011 until now, get the "height" (high-low) of each bar, and calculate the average ? Example: the amount of candles from 2011 to now are 250.000 candles. So we do calculation of the average like this: for (int i=0; i<=250000; i++) { totalHeight += ( ihigh(xxxxxx,i) - ilow(xxxxx,i) ); } average = totalHeight / 250000; Or we just take the candles from previous (last) zigzag leg only ? Tq
jackprobe,

Currently Just the avg of 5 previous bars.

G.
  • #668
  • May 8, 2014 2:54am May 8, 2014 2:54am
Quoting gg53
{quote} jackprobe, Currently Just the avg of 5 previous bars. G.
Thank you very much GG53 ....

jack
Similarity can change your life !
  • #669
  • Edited 3:27am May 8, 2014 3:15am | Edited 3:27am
Quoting gg53
{quote} burnssss, Monty Hall or "3 doors" question does not apply to bars - just FRACTALS. The idea is to find a matching sequences that ends with "1" (as the market point). You start searching by eliminating the first item in the string sequence (I explained elsewhere the logic behind it). if that's doesn't fit, return the first item and eliminate the second one - and so on... I never found a sequence that doesn't fit, using this method - but I wrote a procedure to continue "as is" if it doesn't find a matching new sequence. G.
In reference to the older post above you mentioned "I explained elsewhere the logic behind it". I've been unable to find this post or information. Just curious if anyone can point me there? Thanks.

EDIT: While I'm at it... I'm still having issues confirming the current ZigZag direction. There are occasions where a ZigZag end point is plotted and a fractal in the opposite direction occurs (which you use for confirmation?) but then the ZZ continues in the opposite/original direction thus making that previous ZZ end point invalid. I've incorporated checking the price to the latest ZZ end point price, but its not ideal.
  • #670
  • May 8, 2014 4:55pm May 8, 2014 4:55pm
Per some requests and PM's - here is todays spikes on M5 & M15:
Attached Image (click to enlarge)
Click to Enlarge Name: eurusdm5_1.png Size: 30 KB

Attached Image (click to enlarge)
Click to Enlarge Name: eurusdm15_1.png Size: 26 KB


My algo didn't trade because of big spread opening limits.

G.
  • #671
  • May 8, 2014 5:10pm May 8, 2014 5:10pm
Quoting Mr_You
{quote} In reference to the older post above you mentioned "I explained elsewhere the logic behind it". I've been unable to find this post or information. Just curious if anyone can point me there? Thanks. EDIT: While I'm at it... I'm still having issues confirming the current ZigZag direction. There are occasions where a ZigZag end point is plotted and a fractal in the opposite direction occurs (which you use for confirmation?) but then the ZZ continues in the opposite/original direction thus making that previous ZZ end point invalid. I've incorporated...
Mr_You,

My algo does not relies on ZZ end-point at all.
Only case that I look for ZZ and its "confirming" fractal are in cases where swing predicted mark (arrow) is much different from "real" ZZ mark AND the current leg is not finished its predicted next mark. In those case the mark (arrow) stays where it was but predictions starts over from actual ZZ end-point.
If the "error" prediction occurs after the new leg is already marked, the algo simply continues as is.

G.
  • #672
  • Edited May 9, 2014 5:31pm May 8, 2014 5:41pm | Edited May 9, 2014 5:31pm
Quoting gg53
{quote} Mr_You, My algo does not relies on ZZ end-point at all. Only case that I look for ZZ and its "confirming" fractal are in cases where swing predicted mark (arrow) is much different from "real" ZZ mark AND the current leg is not finished its predicted next mark. In those case the mark (arrow) stays where it was but predictions starts over from actual ZZ end-point. If the "error" prediction occurs after the new leg is already marked, the algo simply continues as is. G.
Ok. I was confusing this with the "expect bullish/bearish" comments. I'll rethink this. This might help a lot.

Thanks.

EDIT: Disregard previous question.
  • #673
  • May 9, 2014 4:11pm May 9, 2014 4:11pm
Okay guys new information.
The creation of set 1 and the next bar direction is depended on zzLegs.
If we create our set1 we only look at the rightmost 13 bars for the creation in every leg.
If the current zzLeg has only 8 bars we only look at exactly THOSE 8 bars for out next bar direction.

e.g.

| Signal
Up[0] Down[1] |Down[2]| Up[3] up[4] Down[5] Up[6] Up[7] Down[8] |Up[9]|
Our generated sequences for set 1 would look like this.

[6][7][8][9]
[5][6][7][8][9]
[4][5][6][7][8][9]
[3][4][5][6][7][8][9]

[3][4][5][6]
[4][5][6][7]
[5][6][7][8]

[4][5][6][7][8]
[3][4][5][6][7]
[3][4][5][6][7][8]
  • #674
  • May 9, 2014 5:02pm May 9, 2014 5:02pm
Quoting Kilian19
Okay guys new information. The creation of set 1 and the next bar direction is depended on zzLegs. If we create our set1 we only look at the rightmost 13 bars for the creation in every leg. If the current zzLeg has only 8 bars we only look at exactly THOSE 8 bars for out next bar direction. e.g. | Signal Up[0] Down[1] |Down[2]| Up[3] up[4] Down[5] Up[6] Up[7] Down[8] |Up[9]| Our generated sequences for set 1 would look like this. [6][7][8][9] [5][6][7][8][9] [4][5][6][7][8][9] [3][4][5][6][7][8][9] [3][4][5][6] [4][5][6][7] [5][6][7][8] [4][5][6][7][8]...
Much thanks for this. And what an extremely odd coincidence that I had this exact same thought a few seconds before reading this post.
  • #675
  • May 10, 2014 2:04am May 10, 2014 2:04am
Quoting Kilian19
Okay guys new information. The creation of set 1 and the next bar direction is depended on zzLegs. If we create our set1 we only look at the rightmost 13 bars for the creation in every leg. If the current zzLeg has only 8 bars we only look at exactly THOSE 8 bars for out next bar direction. e.g. | Signal Up[0] Down[1] |Down[2]| Up[3] up[4] Down[5] Up[6] Up[7] Down[8] |Up[9]| Our generated sequences for set 1 would look like this. [6][7][8][9] [5][6][7][8][9] [4][5][6][7][8][9] [3][4][5][6][7][8][9] [3][4][5][6] [4][5][6][7] [5][6][7][8] [4][5][6][7][8]...
Hi Kilian19,

At your example, we have 10 candles (form [0] to [9]). And we take 7 last candles ( [3][4][5][6][7][8][9] ) just once ?
We do not need to take : [2][3][4][5][6][7][8] , [1][2][3][4][5][6][7], [0][1][2][3][4][5][6] ?
And the same rule for 6 candles, we do not need to take : [2][3][4][5][6][7] . [1][2][3][4][5][6], [0][1][2][3][4][5] ?

And how about 8 last candles and the 9 candles it self ?

Regarding GG's post here : http://www.forexfactory.com/showthread.php?p=7189530#post7189530
i think it is different.

Can you explain it ?

jack
Similarity can change your life !
  • #676
  • May 11, 2014 10:42am May 11, 2014 10:42am
Quoting jackprobe
{quote} Hi Kilian19, At your example, we have 10 candles (form [0] to [9]). And we take 7 last candles ( [3][4][5][6][7][8][9] ) just once ? We do not need to take : [2][3][4][5][6][7][8] , [1][2][3][4][5][6][7], [0][1][2][3][4][5][6] ? And the same rule for 6 candles, we do not need to take : [2][3][4][5][6][7] . [1][2][3][4][5][6], [0][1][2][3][4][5] ? And how about 8 last candles and the 9 candles it self ? Regarding GG's post here : http://www.forexfactory.com/showthread.php?p=7189530#post7189530 i think it is different. Can you explain...

I marked bar 2 as the signal so bar 0 and 1 belong to the zig zag before that leg, which we do not look at if we build the sequences for this particular leg. To create set 1 just use the exact same algorithm you use to create the sequences for the fractals.
  • #677
  • May 12, 2014 2:49pm May 12, 2014 2:49pm
Here is a source of 1 minute data from FXDD. Get it before it is gone. Some goes back to 2005 so lots of date ranges to play with.

http://www.fxdd.com/us/en/forex-resources/forex-trading-tools/metatrader-1-minute-data/
  • #678
  • May 18, 2014 10:31am May 18, 2014 10:31am
Quoting Kilian19
{quote} I marked bar 2 as the signal so bar 0 and 1 belong to the zig zag before that leg, which we do not look at if we build the sequences for this particular leg. To create set 1 just use the exact same algorithm you use to create the sequences for the fractals.
Thank You Killian19

-jack
Similarity can change your life !
  • #679
  • May 18, 2014 12:46pm May 18, 2014 12:46pm
Quoting gg53
Important note: Since we are using PATTERNS to predict next swing point – it might not be exactly as the ZZ point – because we are predicting by PATTERNS - and the ZZ by actual PIPS.
Hi GG53,

I just re-read your post #7 : http://www.forexfactory.com/showthread.php?p=7188904#post7188904
Does "the ZZ by actual PIPS" means the Zigzag height ??

Thanks
Similarity can change your life !
  • #680
  • May 18, 2014 1:13pm May 18, 2014 1:13pm
Quoting gg53
{quote} tashi168, At the very begining, when we have no fractals, we check SET2 only for fractals pattern - ithe probabilities of it being longer. No need to check total length until this condition is met. Probabilities calculations description is on its way... and will be posted soon. G.
Hi GG53,

Could you explain this ... At the very beginning, we have no fractal. Lets say, we are at up wards zig zag leg (from A to B), so it must be a "1" fractal at the end.
And the zig zag leg with only one fractal is valid.

a) Must we wait for that "1" fractal appear first (this happened at point "B"), then we go to the next step (total length calculation) ?
b) As we want to begin our new fractal prediction calculation (our new zigzag leg) - lets say we are at point "A", must we wait for some candles first (lets say we must wait for minimum 3-4 candles), before we do the fractal calculation ? Or we do the fractal
calculation at the first candle, after we get our signal ?

Thanks
Attached Image
Similarity can change your life !
0 traders viewing now