Fractal dimension Trend Following System

In this post I will share more on the Fractal Dimension Trend Following System.

You can see a shot of this system on the 1 h time frame. In fact I do not recommend to go lower.

You need to make a proper optimization of the system: the JJMA is set to 14: It is recommended to go from 6 to 14.

This is the simplest form of use of those indicators.

Any consistent modern system trading need to have at least three sub-systems: It is important to have different systems from different market conditions: Trend following system: Range system and Break - out system.

But the heart is to have an instrument to differentiate between those conditions. This is not easy and this is called experience. What I try is to use a mathematical formula giving us more insight of the problem. So this is not a traditional trend following system in fact it is an upgrade to a traditional system.

Signals:

It is as simple as it can be.

Blue dot: buy signal: the signal is valid only at the end of the bar

Magenta dot: sell signal: the signal is valid only at the end of the bar

Price action reading:

Yes this system has a price action reading:

Blue Bar: uptrend bar

Red Bar: downtrend bar

Green Bar: range bar

-It is a good idea to have a confirmation between the price action and the signal generation.

-If a trend is established you can enter or add when the proper price action bar is identified.

Measure of the Cyclic component:

This is interpreted as regular oscillator. We have a mod version of the Polarized fractal efficiency indicator and a Normalized version of Singular Spectrum Analysis indicator.

This folks is the best I found for cycle extraction. Needless to say it is a very good idea to have a conformity from the cyclic component and the trend component. From time to time arrows will appear from the SSA indicator, treat them as alerts, this is not a trade signal.

Measure of the Trend component:

This is an equivalent of the ADX. But it is far more reactive. When the faster ADX goes to its maximum it reveals a correction.

Measure of the Fractal Dimension:

Here it is the real deal.

In fact all the mentioned available instruments are to measure the actual state of the market. Nothing more and nothing less. It is important to measure the actual conditions of the market.

The fractal dimension gives us a probability that what we have measure will continue in the future.

If the fractal dimension is in the blue zone: we have a greater probability that the movement will reverse.

If the fractal dimension is in the red zone: we have a greater probability that the movement will continue.

In fact the probabilities are not great but this is something concrete. So what is important if the probability increases or decreases.

I have found that the best use of this system is done on 1 h, or more. Below that level it is far more complicated.

3 rules:

When we have a FGDI in red both in the 15 m and 30 m time frame this is a special place. Use the signals as they are given follow the signals.

When we have the inversed conditions: do not trade a trade following system or do not trade at all.

When we have the 15, 30, or 1 h, the FGDI at the level of 1.5. That means we are at fifty fifty situation.

The Hurst difference. Measure a peak of the change of the fractal dimension of the price. When we have a peak this is a signal that something important is going to happaen. If that is observed in a conjunction with a transition of FGDI from the red to the blue zone this is an indication of a break - out market condition.

Copyright:

All rights are reserved for the technical indicators distributed.

The fact that the files from this system are freely distributed and I found them on internet does not mean that its components are free of copyright. All of the indicators in this system are freely available with their own copyrights. Most of them are a trial for reverse engineering of some very famous systems. For this reason I will upload the source code files wherein you will find the proper copyright information. This effort is done only to save you many hours on internet search.

Any commercial exploitation and distribution of this system is prohibited because it will violate the copyright of its owners.

The ideas of the Fractal Dimension Trend Following System can be realized with almost any technical indicator, even manually with the proper expertise. The fact that I have included those available indicators and systems is because they are very popular and easy to use.

Installation Note:

You can download here the necessary files:

http://www.mediafire.com/file/od11kos81vbhzt7/experts2.rar

Install all the files from the folder in the experts folder of Metatrader

Install the file from include folder in the include folder of Metatrader

Install the file from library folder in the library folder of Metatrader

Install the template file into the template folder of Metatrader

There are some files that are not included in the template. You can use them if you wish.

October

Unfortunately I did not take position this month. Otherwise I would write the position in the blog. However the purpose of this post is to show some of my trend following strategies.
It is funny everything on this screen shot is free as freely available on the net.

OK there are tho shots one is on the daily chart and another on the 4 h time chart.
How to interpret. It is fairly easy.

The dots are the entry points blue for buy and magenta for sell.

The line can be above or below the price. This is a stop loss level.

This is a modification of the Wilder volatility system (one of the best trend following systems ever, ask why it is not so popular). The modification consists in the smoothing by an innovative digital filters. The idea is freely available and elaborated by Russian hackers. For me this is one of the most advanced trend following strategies available outperforming even the famous able trend.

But this is not all. Every trend following strategy needs a trend filter. Nothing original. Yes but this is different. Traditionally the ADX is used do distinguish between trend and range environments. What we have here is a better version of the ADX. What is better?
This indicator is far more reactive than the ADX. Especially for the end of the trend, when it reverses after the peak this means that we have a correction. That is why I use two on the same window one is more reactive and one is less reactive. One measures the little bursts and trends and the other the bigger trend.

And that is not enough. We have the fractal dimension index. I can write a lot about it. But this indicator is also freely available. In a nutshell on this bigger times frames it can be analyzed as follows.

When we are in a red zone: we have a bigger probability than the next movement will be in the same direction (the right term is to say that the movement is persistent).

When we are in a blue zone: we have a bigger probability than the next movement will be in the opposite direction (the right term is to say that the movement is antipersistent).

This indicator estimates the fractal dimension of the time series. The times series in the market have a fractal dimension. I know folks this becomes a little complicated. So look at he price the more linear it is the more it is closer to the 1 (one dimension). The more jagged the more it is closer to 2 (tho dimension), in fact the price can never be 2 dimensional because it cannot close itself.

The medium line is 1.5. This is the fractal dimension of the Brownian motion. The probability is 50 % for the next movement to be in either up and down.

In fact it has been mathematically proven that the time series in the Forex market have the so called long term memory. That means that what happens today and what has happened on the past affects the future in a non linear way and complex way.

So a further enhancement of this indicator is the Hurst difference indicator. This indicator has been developed by a friend of mine and is also free (in the modern sens of face book of course).
I am particularly happy that I have contributed to the discussions that have led to the discovery of this indicator. So this is a very particular indicator. This is seismic meter of the market.
So when you see a peak you have to know that something important is happening or is about to happen that will affect the market. Often it is a confirmation of a break-out (in fact a valid break out is accompanied by a transition from the blue dimension to the red dimension: I call it a fractal break-out and this has nothing to do with the fractals of Bill Williams). This pattern is a new and has been documented for the first time some years ago in the forum TSD (where you can find most of the indicators).

Finally we have an advanced oscillator. In fact I use two indicators in one window:
One is a modified polarized fractal efficiency PFE. The other is really the best oscillator ever (in fact not a true oscillator) this is a SSA (singular spectrum analyses tuned to extract the oscillatory component of the time series). And it is interpreted as a regular oscillator in the trend following strategies. We buy or sell in the direction of the trend only.

We have a Bollinger band but they are not really necessary in the big time frames. And finally there is a very fast digital filter changing color. This one particularly is not free its name is gauss filter and it is available in the Elite section of the forum TSD. But there are some other indicators that do the same job sometimes even better.

End of month

The long position opened from the beginning of the month 1.2883
is closed at 1.3540.

It makes 657 pips.
We used 0.01 lots: that makes 65.7 USD

The losses were 16.60 USD

The total gain for the month is 49 USD.

Not bad for the microhedge fund. We had 100 USD and now we have 149 USD. The idea is to
get out of the account the wins of the month.

So 49 USD in the pocket let's buy some beer.
Cheers.

First loss

The short position has been closed at 1.2883.

The incurred loss is 16.60 $.

The position was not managed well because it was positive for a certain amount of time.

I open a long position from the same level.

Entry 1 Short EUR/USD

I do not have a conformation of this entry, from my indicators. But I will enter into this position now because I do no want to miss the movement to the basis of the correction. The stop loss will not be according to the daily graphics.

Entry: Sell short from 1.2702

Stop loss: 1.2968

Strategy: Discussion of entries and exits

We are going to use the 4 h time frames for the main entries.

So when we identify an entry on the daily we are going to ride the trend with 0.01 lots.

When we have a correction (in the 4 h time drame) we are going to close the daily position and open an entry in the direction of the correction.

We can also hedge the daily and open a new position in the direction of the correction, it is a matter of choice but I do not want to use more capital for the margin and pay extra swap costs for three positions.

Signal generation

The signal generation is fairly complex.

1. It is based basically on an upgrade of the volatility system of Wilder with digital filters.

We will use also some indicators based on the chaos theory.

2. We are going to use the fractal dimension index (and the exponent of HURST) in order to measure the probability of the trend continuation. And the variation of the exponent of HURST. The idea is that every algorythm for the estimations of this exponent cannot give a precise estimate at any local point. So we are going to be interested at his sudden change for a break - out. This is part of some day trading strategies that will not be covered here.

3. We are using the Lyapunov exponent in order to estimate how chaotic the time series are.
It may be curious but the latest experiments showed that the EUR/USD is not chaotic at the 30 m. time frame (-2 )in fact it has a negative Lyapunov exponent. And that is very curious, that means that we can make predictions and we have to. The system is predictable. But the system looses the equilibrium from time to time.

First Entry

The first entry is difficult. It is very difficult to time the entry for the trend following strategy. Sometimes it is better to not time it at all. I think today is not the day for an entry on the market, because we do not know if the trend is going to continue or to reverse.

The idea is to evaluate in a stress free environment. When you risk 150 $ you are stress free, but who knows. The idea is that very often when we enter into a trend following we face the following problem.

People tend to take their profits too early and to hold loosing positions for longtime.
In fact in order to succeed in our experiment we have to force ourselves against our natural drive. That means that when we see a winning position we are not going to close it when we see an averse price action, we will let it breathe. On the other hand when the position is negative if it exceeds the stop loss level we will close it without remorse.

We are doing that not from a day trader perspective. We are not going to check everyday our deals. The money has to work for us. Right!

About

This blog is created as an experiment. This is a micro hedge fund experiment.

Is it possible to have a personal micro hedge fund to generate profits with money far bellow the imaginable?

My answer is probably Yes. That is why I have created this experiment. So what is the situation?

We have as low as 150 USD of risk capital that we want to invest. You are laughing OK, I understand. And we want to invest them in the market. The idea is that we are going to make a time frame diversification (that can be an eye opener for the day traders)

We are going to apply a high risk trend following strategy with our capital.The purpose will not be to accumulate profits and more profits. The purpose will be to take every month half of the profits and bye some beer.

I will give some technical details.

We are going to use mainly daily and 4 hours time frames. We will use a moderate leverage.
We are going to take positions with 0.01 lots or 1000 $.

The entry into position will be generated on a 30 minutes time frame.

The stop loss will be dynamic and will evaluate with the trend to lock in the profits. The stop loss levels will be deep in order to avoid the high frequency algorithms effects as much as possible.

I will open positions only on EUR/USD, because it is a well trending market we want to exploit. The idea is fairly simple this market has good trending characteristics we want to exploit. I have been forward testing this strategy for six months, before starting the experiment.