Looking for patterns - page 38
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
The slips are lagging behind, which is not good
Let them fall behind!!! We are only interested in the muwings as a target function and nothing more. Who prevents us from taking the history and moving the chart back by the required number of bars, so much so that it repeats the price movement as much as possible. It is only the target function that will show reversal on history. We are not talking about using it in trading yet.
P.S. Here look at not real data. A red arrow appears - it is a possible reversal. Then the blue arrow appeared, but the first impulse upwards is big, the muving from the higher timeframe was not interrupted - as a result it went further. The red horizontal line is an intermediate target, it has already worked out. If the current conjuncture does not change, it should reach 1.2924 +- with a probability of 80%.
The whole problem in defining the pAtterns that we are all looking for is to create a target function that gives an unambiguous pivot signal.
In order to determine it correctly, there should be as many modules involved as possible. If we maximally combine interactive data from sources that influence the movement plus pattern recognition, it is almost a so-called grail. The problem with creating such a system is simply that users do not care about what is needed to create the most error-free system possible. They are interested in a product at an affordable price that will work. Developers, on the other hand, are interested in how to create a product that meets the user's needs with the least amount of effort. Or even to create the semblance of meeting these needs. Creation of a serious system requires both effort and deep analysis, as well as a lot of resources. Also, only those with significant resources and capabilities have the ability to create such a system. It is not a question of two steps and 64 Kb algorithm. That's why very few people will really be interested even in thinking about such things.
Vitaliy, don't be offended.
I wasn't even thinking about it. And I understand the mood of the majority of people on this forum. Not many people are interested in the direction we are discussing here. And this is evident even from the topic Romfil tried to develop. And it is also obvious that if you start to fully understand such issues, it is possible only through data, from which the majority will have a detachment of free radicals in their neuronets. That would be a global offtopic. But to display the very possibilities of such options, is appropriate. There will be those who will notice not only the water, but also the sources of progress in this.
A neural network is a small brain, weaker than a human, but it does not get tired. This raises one but the most important question: if you do not know what to expect from the market, how does it know?
The advantage of AI is in operational resources and more. And humans are also tied to linear logic. And AI has been taught volumetric processes and it is like breathing.
Good afternoon dear visitors to the branch!
One promising idea that will allow for a definite statistical advantage. As promised.
I should say at once the original idea is not mine, but I have developed it. Who is interested the beginning here.
The idea is based on the principle of different speed of price movements of two symbols - let it be GBPUSD and EURUSD. Although it may be unfortunate instruments because of the Brexit that caused the pound to fly for the last couple of years. But let it be them.
So. All of you probably know and have seen how some currencies "move" on the chart. And you've noticed more than once the correlation of symbols, which have a common currency - for example, sometimes charts of GBPUSD and EURUSD are very correlated, but they have different absolute values, but the form of the charts is similar. Well, no one will mind that these two symbols are dependent on each other. What dependence between them is unknown, but it exists! Through the common currency USD.
Now let's pull a little Gunn ... :) :) :) Those who have read Gunn's theory should know about his main discovery - price range quadrupling by time. What this discovery is about is "... when price is squared by time or, when price and time go together ...". I personally understand (or rather take it as an axiom) as follows - each instrument has its own unknown "golden ratio" of price movement relative to time - the so-called "price scale" - this is by how many points the price moves relative to the timeline. The price always returns to this "golden ratio". This is not the purpose of this book, but I will not change the opinion of those who do not agree with this statement - those who do not agree with this statement do not need to read further. So both instruments mentioned above(GBPUSD and EURUSD) have these values of golden ratio - it is impossible to calculate their absolute values, but it is possible to suggest their relative value.
What may ensure an upwards movement of GBPUSD price, for example, there are several variants: (1) GBP goes up, USD does not change; (2) GBP does not change, USD goes down; (3) GBP goes up, USD goes down; (4) GBP goes up, USD also goes up, but at a slower rate. There are only four possibilities! But their absolute value is impossible to predict. What if it is relative???
Here are the charts of movement of GBPUSD and EURUSD on M5 timeframe for 06.01.2020:
Why this particular day? Because the screenshot is taken from the previous article, which was written in the middle of January.
What do we see? Is there a correlation? Definitely there is a correlation!
Now I want to make a note of one regularity that the author of the original idea had detected - it seemed to him back in 2012, but he was only now able to somehow calculate it. The regularity is that the speed of change of GBPUSD and EURUSD is different, but they always return to their "golden ratio" - always!!! The author of the idea tied this feature to probability, but in my opinion - it is the feature of the return to this "golden ratio".
I will try to explain the meaning of the idea in more detail - how to make money on it (we will do it in a straight line):
1) We need to define the relative "golden ratio" in the present! How to do this - it is necessary to make an analysis of history. To what depth? I will not give an exact answer here - to each his own, as they say. The author of the idea takes one day, for M5 it is 288 counts backwards. And we will use this depth. These are the charts of changes of the mentioned pairs quotes at the beginning of the market opening on 06.01.2020 (the charts are drawn according to CLOSE).
2) As the author of the idea put it - by some "shamanistic" method we will obtain a certain "golden section". To which both graphs will aspire.
3) We then scale all three graphs in a certain way and end up with this:
What does this graph tell us? Nothing so far, but if we look at this graph in history - start from some date and shift this window to the right, we will see that prices of both pairs diverge and converge to this curve, which I personally call "golden ratio" (it is yellow on the graph, blue is euro, red is pound).
4) How do you use it? Very simply. All three lines will always (well, sooner or later) tend to one point, i.e. the speed of price change will return to this golden ratio. It is not known where this point will be, but it is possible to guess where the charts will move. I wrote an indicator that displays these three lines in real time. It has exactly three of these buffers.
The red line is EURUSD, the blue line is GBPUSD and the magenta is the golden ratio.
5) Determine to which of the lines EURUSD or GBPUSD is the cross section, let that line be called slow, and the second one - fast. For these two pairs the EUR is almost always slow, GBP is fast. So, at the moment when the section starts to turn towards the slow line, we open on EURGBP instrument! If at the moment of entering the EUR is higher than the pound, then we open down. If on the contrary, then upwards.
6) Exit from the position you can offer yourself.
Well, I think I have told you everything. I am ready to provide methodological assistance in clarifying, so to speak, the technical requirements for the adviser. The indicator is attached; it may be used through iCustom.
Regards, RomFil.
In order to be able to detect without errors, there should be as many modules involved as possible. If one combines as much interactive data as possible from the sources that influence the movements, plus pattern recognition, it is almost a so-called grail. The problem with creating such a system is simply that users do not care about what is needed to create the most error-free system possible. They are interested in a product at an affordable price that will work. Developers, on the other hand, are interested in how to create a product that meets the user's needs with the least amount of effort. Or even to create the semblance of meeting these needs. Creation of a serious system requires both effort and deep analysis, as well as a lot of resources. Also, only those with significant resources and capabilities have the ability to create such a system. It is not a question of two steps and 64 Kb algorithm. That's why very few people would really be interested even in thinking about such things.
I haven't even thought about it. And I understand the mood of the majority of people on this forum. Few people are interested in the field we're discussing here. And this is evident even from the thread Romfil tried to develop. And it is also obvious that if you start to fully understand such issues, it is possible only through data, from which the majority will have a detachment of free radicals in their neuronets. That would be a global offtopic. But to display the very possibilities of such options, is appropriate. There will be those who will see it as more than just water, but also as a source of progress.
A lot of it is also harmful. Only by experiment can it be determined. For 120 inputs and one output I've defined it "optimally" as a two-layer structure of a deep network: first layer 120 neurons, second layer 40 neurons, output 1 neuron.
While writing the "long" post, I did a little haggling. Probably won't go any further. Although it did ... :) So I got out early, but I've had enough. :)
Here it is, the advantage of "in person". It's the regular LWMA that gave the signal to enter. And the targets are the red lines. The red lines have been reached and that's it! Then there is no predictable movement ... :):):)
Now about LWMA in more detail as promised.
Put two scales on the chart: period 27, apply to CLOSE, LWMA method, one with shift 0 and one with shift 1.
We look only at the beginning of the zero bar. Analyze from the first bar. As soon as one bar crosses the other one by more than the current spread, this is a signal of a possible reversal. The signal workout is subjective and is about 70%. So that's all the wisdom.
In my opinion, these signals are the unique target function for further search of patterns. What do you think?
Do you want me to teach you how to search for them (regularities) by yourself? I used to do it before and I'm ready to share my knowledge ... :)
Regards, RomFil
3) We then scale all three graphs in a certain way and end up with this:
What does this graph tell us? Nothing so far, but if we look at this graph in history - start from some date and shift this window to the right, we will see that prices of both pairs diverge and converge to this curve, which I personally call "golden ratio" (it is yellow on the graph, blue is euro, red is pound).
4) How do you use it? Very simply. All three lines will always (well, sooner or later) tend to one point, i.e. the speed of price change will return to this golden ratio. It is not known where this point will be, but it is possible to guess where the charts will move. I wrote an indicator that displays these three lines in real time. It has exactly three of these buffers.
The red line is EURUSD, the blue line is GBPUSD and the magenta is the golden ratio.
5) Determine to which of the lines EURUSD or GBPUSD is the cross section, let that line be called slow, and the second one - fast. For these two pairs the EUR is almost always slow, GBP is fast. So, at the moment when the section starts to turn towards the slow line, we open on EURGBP instrument! If at the moment of entering the EUR is higher than the pound, then we open down. If on the contrary, then upwards.
6) Exit from the position you can offer yourself.
Well, I think I have told you everything. I am ready to provide methodological assistance in clarifying, so to speak, the technical requirements for the adviser. The indicator is attached; it may be used through iCustom.
Regards, RomFil.
Pair trading is not new. And you can find thousands of posts on it in the red forum, even earlier than 2012, 2009-2010, if I'm not mistaken. And it doesn't matter if we trade both majors or crosses. There are a lot of indicators on pair trading. And there are a lot of Expert Advisors. But there's no successful long-term use. I personally lost several thousand dollars using the strategy described to you five years ago. And I don't even want to go back to it. If you want to step on the same rake, you can continue your research. Although, you won't be able to invent anything new. So many bright minds have beaten on this strategy that it is impossible to count...