Fractals, ZZ & the Pissing Dog | Page 8 | 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

  • #141
  • Jan 10, 2014 10:06am Jan 10, 2014 10:06am
Quoting Magix
Playing the numbers... Off of the Fractal 5 candle pattern, TOG traded on 3 USD cross pairs this weeks with a Stochastics filter...M1 chart. See the results. LINK (http://www.myfxbook.com/members/Magix/sto-fra-swb/796008). Expectations? With the right qualifier and perhaps a better set of criteria, could actually improve the overall performance...maybe? {image}
Hi Magix, thanks for your posts in this thread (and your thread). One of the thing i look at in trading results is what i call margin (i know there are other terms people use). basically $ made per $ notional traded. It gives a good idea how well strategy will do in real trading where there is slippage involved. In these results it comes out to about 4-5 pips per trade margin (121 pnl on 2.5 lots traded) which means if there is a slippage of 1 pip (spread costs, exec costs) etc per side, the returns would cut in half.

i know this is on 1m chart so margin is going to be low. but usually i look for strategies to have 10+ pips in margin.

would be interested in your comments...
Think : Trade : Live Life
  • #142
  • Jan 10, 2014 10:39am Jan 10, 2014 10:39am
Quoting stt
{quote} Hi Magix, thanks for your posts in this thread (and your thread). One of the thing i look at in trading results is what i call margin (i know there are other terms people use). basically $ made per $ notional traded. It gives a good idea how well strategy will do in real trading where there is slippage involved. In these results it comes out to about 4-5 pips per trade margin (121 pnl on 2.5 lots traded) which means if there is a slippage of 1 pip (spread costs, exec costs) etc per side, the returns would cut in half. i know this is on 1m...

stt,

The nature of the thread and GGs studies is that the market is Fractal in nature. The fact that I am using the Fractal pattern is just an example of candle pattern recognition and applied to a base trade system, but again, still trying to keep relevant to this thread.

Given that this is an M1 chart, my expectation of a Net $1 per cost averaged trade is what I consider a reasonable expectation, so if we moved to the M5-15, it would be reasonable as well...just a matter of where to set the range and actually prove that the same criteria would fit into different time frames.

Appreciating that we have seen an average of 4-5 pips per trade on the M1, we may possibly be able to move to the M15 chart, adjust the ranges accordingly and bring up the overall average per trade into the 10 pip range that you are mentioning.

In theory, of course.

With this being said, there is always the other side of the coin. What are the chances for losses encurred by swap fees? Would the strain on Margin also increase proportionally?

Having studied automated traders with different variables for the last couple of years, I will still honestly say that I am still relatively fresh when comes down to these. About the one thing that I have found consistent is that working with the smaller ranges can produce a lot better results because of the expected range alone. On others that I have worked with, and although different criteria, they will do well on larger ranges, but when market goes into a big trend, can get crushed by the weight of the cost averaged trades.

There is definitely a delicate balance as applied to these.

Additionally...These too could be tested on a SLTP principal instead of the grid trades. Setting a range that is based by just triggers alone would allow the market to define profit and loss based upon the triggering criteria, not allow more than a single trade per direction and could still be fit into any time frame.

There is a lot of road to travel and a lot of weight to be carried.

I think that with all the efforts combined here and information gleaned from other sources, we can spread the load and narrow the path.

Just my dos centavos.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #143
  • Jan 10, 2014 10:46am Jan 10, 2014 10:46am
Quoting Magix
{quote} stt, The nature of the thread and GGs studies is that the market is Fractal in nature. The fact that I am using the Fractal pattern is just an example of candle pattern recognition and applied to a base trade system, but again, still trying to keep relevant to this thread. Given that this is an M1 chart, my expectation of a Net $1 per cost averaged trade is what I consider a reasonable expectation, so if we moved to the M5-15, it would be reasonable as well......
Thanks for the reply. Agree there is a lot of room here to experiment and find good balance. In general I avoid grid strategies but agree there is merit in those (as long as your underlying signal has some + expectancy).

Sorry GG if i derailed your thread a bit here. just wanted to make sure people understand the impact of slippage in real accounts (have seen strategies that look perfect in bt/demo but cant make anymoney because margin is <2 pips).
Think : Trade : Live Life
  • #144
  • Jan 10, 2014 11:53am Jan 10, 2014 11:53am
Quoting stt
{quote} Thanks for the reply. Agree there is a lot of room here to experiment and find good balance. In general I avoid grid strategies but agree there is merit in those (as long as your underlying signal has some + expectancy). Sorry GG if i derailed your thread a bit here. just wanted to make sure people understand the impact of slippage in real accounts (have seen strategies that look perfect in bt/demo but cant make anymoney because margin is <2 pips).

I will add to my previous reply...

If the idea is to study 4-5 candles on an M5 chart, expectation for probable forecast would be to either take all of the 4 bar candle patterns and predict the 5th with some sort of higher probability.

Inside of a single bar on the M5, what would be the expected range of this 5th bar? How much of this would be consumed by spread? What is a reasonable expectation?

Or would we look to this 5 bar candle period as a reversal pattern?

Taking only a few pieces of information in to consideration, still leaves the application of additional data as the, still needs to be done.

But really, each journey of a thousand miles, still needs the first step taken.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #145
  • Jan 10, 2014 1:22pm Jan 10, 2014 1:22pm
So, I accidentally had my algo indicator attached to the Mini-Dow yesterday and today.
To my surprise, great signals!!!

Please note, that the algo history/database is based on EURUSD 5-min bars.
Yet the zigzag (real-time), fractals, and bar information are coming from this 1-Hour chart of the DOW!!
Kind of shows you that all markets may be similar in fractal construction, no matter the timeframe.

@GG53, I remember you talked about a universal indicator... I think this confirms that it is possible.

Carl
Attached Image (click to enlarge)
Click to Enlarge Name: DOW-1h-dna.jpg Size: 259 KB
  • #146
  • Jan 10, 2014 1:29pm Jan 10, 2014 1:29pm
Quoting crodzilla
So, I accidentally had my algo indicator attached to the Mini-Dow yesterday and today. To my surprise, great signals!!! Please note, that the algo history/database is based on EURUSD 5-min bars. Yet the zigzag (real-time), fractals, and bar information are coming from this 1-Hour chart of the DOW!! Kind of shows you that all markets may be similar in fractal construction, no matter the timeframe. @GG53, I remember you talked about a universal indicator... I think this confirms that it is possible. Carl {image}

At a quick view, and forgive my ignorance here...but if were are going off of candlestick pattern recognition, why was there no signal where I have circled on your chart?

Thanks
Attached Image (click to enlarge)
Click to Enlarge Name: crod.png Size: 240 KB
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #147
  • Jan 10, 2014 1:50pm Jan 10, 2014 1:50pm
Quoting Magix
{quote} At a quick view, and forgive my ignorance here...but if were are going off of candlestick pattern recognition, why was there no signal where I have circled on your chart? Thanks {image}

We're not going off of candlestick pattern recognition. We're going off of a humongous database of completed zigzag swings, and determining the most likely end-of-swing pattern for the swing.

Its all in the thread (previous one), or in the pdfs and docs at the beginning of the thread.
  • #148
  • Jan 10, 2014 2:47pm Jan 10, 2014 2:47pm
Quoting Starry
{quote} We're not going off of candlestick pattern recognition. We're going off of a humongous database of completed zigzag swings, and determining the most likely end-of-swing pattern for the swing. Its all in the thread (previous one), or in the pdfs and docs at the beginning of the thread.

Yeah...and yet...

This same thing could have been accomplished with the use of the ZigZag indicator and zero pattern recognition or Algo...

So, let's save the chatter and please allow Croc to answer.

This would seemingly prove to be a correcting indicator as it is seemingly a look to the hindsight and correct in the fore.

I ain't trying to burst anybodies bubbles here, but 4 bars in a single direction and trigger on the first counter is seemingly the pattern.

Have a look at the chart I posted and try to figure in that until the continuation of trend, why wouldn't an arrow be placed there.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #149
  • Jan 10, 2014 3:41pm Jan 10, 2014 3:41pm
Hi Carl,

Are you applying the Monty Hall Problem part of the algo yet?

Thanks
  • #150
  • Jan 10, 2014 4:24pm Jan 10, 2014 4:24pm
Quoting Magix
{quote} Yeah...and yet... This same thing could have been accomplished with the use of the ZigZag indicator and zero pattern recognition or Algo... So, let's save the chatter and please allow Croc to answer. This would seemingly prove to be a correcting indicator as it is seemingly a look to the hindsight and correct in the fore. I ain't trying to burst anybodies bubbles here, but 4 bars in a single direction and trigger on the first counter is seemingly the pattern. Have a look at the chart I posted and try to figure in that until the continuation...
From that previous low zigzag up to that point, the fractal pattern is (01). Searching that pattern told me that the (010) pattern is actually more probable. So, we wait for it. The higher probability pattern ends with a LOW fractal. You can't have a short signal when the next fractal should be a LOW fractal.

Price correctly made that bearish bar into a "0" fractal. So now an EXACT search of (010) is the correct configuration of fractals to find the next TOP fractal for a short signal.

Carl
  • #151
  • Jan 10, 2014 4:24pm Jan 10, 2014 4:24pm
Quoting Stinky
Hi Carl, Are you applying the Monty Hall Problem part of the algo yet? Thanks
GG53 has not posted that yet, and no he has not given it to me apart from everyone else.
I will learn it when you do.

Carl
  • #152
  • Jan 10, 2014 4:27pm Jan 10, 2014 4:27pm
Quoting Magix
{quote} Yeah...and yet... This same thing could have been accomplished with the use of the ZigZag indicator and zero pattern recognition or Algo... So, let's save the chatter and please allow Croc to answer. This would seemingly prove to be a correcting indicator as it is seemingly a look to the hindsight and correct in the fore. I ain't trying to burst anybodies bubbles here, but 4 bars in a single direction and trigger on the first counter is seemingly the pattern. Have a look at the chart I posted and try to figure in that until the continuation...
How did your algorithm tell you there should have been a short there?

Carl
  • #153
  • Jan 10, 2014 4:49pm Jan 10, 2014 4:49pm
Quoting crodzilla
{quote} From that previous low zigzag up to that point, the fractal pattern is (01). Searching that pattern told me that the (010) pattern is actually more probable. So, we wait for it. The higher probability pattern ends with a LOW fractal. You can't have a short signal when the next fractal should be a LOW fractal. Price correctly made that bearish bar into a "0" fractal. So now an EXACT search of (010) is the correct configuration of fractals to find the next TOP fractal for a short signal. Carl
So, you are searching a 3 or 5 candle pattern?

Because if that is the case, than the following short from top of the chain doesn't make much sense.

After which candle would the indicator paint the arrow?
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #154
  • Jan 10, 2014 6:06pm Jan 10, 2014 6:06pm
Quoting burnssss
{quote} In the example I gave above, I think from the beginning should have introduced [1], but regarding gg53 response in post 129, it makes me wonder how the process
burnssss,

Your example is Ok, but when you compare it to database you have to note that it always contains at least one fractal in it.

G.
  • #155
  • Jan 10, 2014 6:11pm Jan 10, 2014 6:11pm
Quoting emonts
{quote} Hi, As I understood only when you startup your algo you have to find the first swing point, you start with a fresh MT4 leg and your algo will find a swing point. Myself I woud prefer a new leg with several fractals in it. Better probability. Every swing point found by your algo is the start of a new leg (one bar later) for your algo. So you leave the MT4 leg endpoints for what they are. You have to compare the algo swing points with the MT4 leg end points. If to much difference then you have to synchronise your algo again with a MT4 new...
emonts,

Right!

G.
  • #156
  • Jan 10, 2014 6:20pm Jan 10, 2014 6:20pm
Quoting crodzilla
So, I accidentally had my algo indicator attached to the Mini-Dow yesterday and today. To my surprise, great signals!!! Please note, that the algo history/database is based on EURUSD 5-min bars. Yet the zigzag (real-time), fractals, and bar information are coming from this 1-Hour chart of the DOW!! Kind of shows you that all markets may be similar in fractal construction, no matter the timeframe. @GG53, I remember you talked about a universal indicator... I think this confirms that it is possible. Carl {image}
crodzilla,

After initial surprise - It makes some sense - we are analyzing PATTERNS .

G.
  • #157
  • Jan 10, 2014 6:20pm Jan 10, 2014 6:20pm
Quoting Magix
{quote} Yeah...and yet... This same thing could have been accomplished with the use of the ZigZag indicator and zero pattern recognition or Algo... So, let's save the chatter and please allow Croc to answer. This would seemingly prove to be a correcting indicator as it is seemingly a look to the hindsight and correct in the fore. I ain't trying to burst anybodies bubbles here, but 4 bars in a single direction and trigger on the first counter is seemingly the pattern. Have a look at the chart I posted and try to figure in that until the continuation...

you clearly need to read the thread. stop thinking and start reading
  • #158
  • Edited 6:38pm Jan 10, 2014 6:21pm | Edited 6:38pm
I always subtract one then?,
Do not quite understand.

Referring to post 126
the first red arrow, you would not have signal Highlights calculations and instruction, but in the second zz, having no signal i would start to calculate from the previous zz confirmed, first introduced fractal at the end of the current zz. I thought I should put [0] and start counting until you have signal, but also can not be me because the signal appears in the middle of zz.

Go I'm confused!

Many thanks
  • #159
  • Jan 10, 2014 6:29pm Jan 10, 2014 6:29pm
Quoting Magix
{quote} Yeah...and yet... This same thing could have been accomplished with the use of the ZigZag indicator and zero pattern recognition or Algo... So, let's save the chatter and please allow Croc to answer. This would seemingly prove to be a correcting indicator as it is seemingly a look to the hindsight and correct in the fore. I ain't trying to burst anybodies bubbles here, but 4 bars in a single direction and trigger on the first counter is seemingly the pattern. Have a look at the chart I posted and try to figure in that until the continuation...
Magix,

Simple, because the probabilities calculations suggests otherwise.
We are analyzing fractals pattern within a ZZ leg, then total number of bars, and prediction of "next bar" in opposite direction - BEFORE it happens.

G.
  • #160
  • Jan 10, 2014 6:52pm Jan 10, 2014 6:52pm
Quoting gg53
{quote} Magix, Simple, because the probabilities calculations suggests otherwise. We are analyzing fractals pattern within a ZZ leg, then total number of bars, and prediction of "next bar" in opposite direction - BEFORE it happens. G.

I totally understand the concept and logic...was just a simple question.

If anybody had taken and used the ZZ or Fractal pattern and only traded them, their success would be limited if they experienced any at all with reasonable Money Management in mind.

I was asking a simple question based on what appeared to be a fairly generic pattern.

Having worked with several strategies and indicators, the ones that seem to pick absolute tops and absolute bottoms typically repaint and going by observation alone is why I asked my question. Look at semafor as an example. That bitch looks like it nuggets tops and bottoms all the time, but in reality, there is no more value to that looking forward than will provide profits looking at it in reverse. Even worse, it creates false hopes and market casualties.

I have my serious doubts about the validity of what was posted...but that doesn't mean I am right. It just means that I need more information and live running proof before I could actually lend any credence.

Really GG...If I didn't believe that there was something that could be accomplished here, I wouldn't have taken the time to post and throw out some of my own efforts.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
2 traders viewing now