a trading strategy based on Elliott Wave Theory - page 257
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
I don't get it, we take away channels with minimal energy and then it leaks somewhere else. If this is not the case, energy between the channels should equalize and thus replenish energy of the channels taken (to make them less stable). Maybe I have confused functions and functions again :) .
Basically yes, it's a function, which I stubbornly called a function. But it seems to have been sorted out. Concluding the story about dynamic analysis, it's worth noting that this simple method allows us to track the life cycle of channels. Conceptually, it's similar to your "real-time" channels "looking" into the future. By observing the indicator, you watch the development of the channel until it ceases to exist and a new one is formed. This method allows you to "read" the entire process in history.
With proper identification, which is not difficult, you can identify very important additional characteristics such as "age", "growth/ageing rate" and many others for each channel. The trajectory of local extrema in the window plane looks quite simple and can be well approximated by a parabola, which in turn can be conventionally (!) represented by a life cycle. We can stop here or we can go further, connecting, for example, the theory of catastrophes.
The dynamic model can be applied not only to PE, I hope that the general (and very simple) approach stated earlier will be useful for someone, maybe for Solandr, (the channels can be not only linear) I will be only glad. :o)))
Only the method of calculating PE is left out of the picture, but this is correct. Otherwise the discussion of some aspects of the model would have turned into an elementary publication. And such a goal is hopefully not worthwhile.
Exactly right, I am not intended to do that and I'm not a good "publisher". I do not intend to post "my" Hearst, upgraded MSP, potential energy and so on. In this connection I am grateful to Vladislav for not publishing anything except his idea. :o)))
So I wonder what results such a multi-component and mathematically rich system would lead to.
There's an old joke "the best is the enemy of the good". My point is that at the moment I have two competing models and as a consequence work on the model is ongoing. Now I am pleased that I have abandoned trading in favor of deep research and digging into history, although I broke my word a few times, but I have benefited from it.
The current results are more than satisfying. In comparison with them even 2-7 hours of calculations (depending on the price series) seem to be just a trifle, after all, solved with a "scrap", say, a dual-processor computer. I'm not even intimidated by picking history, because in fact all good theories emerged from its analysis, including Elliott Wave Theory.
Soberly evaluating the time of transferring the system to real (believe me, I like beer, but not 24 hours, 7 days a week and 365 days a year, so I am very often sober :o) I realized that it is a very, very long process, and given its boundary conditions, the transfer to MT becomes almost impossible to achieve in a reasonable time. It has nothing to do with MQL4 capabilities, they're good enough. I'm not enough for this work. And taking 5-6 months leave is somehow too much of a problem.
I am writing a requirements specification for dll development for MT, by the way, here is a short excerpt from it:
Technical architecture
(1) Server for calculations and database (presumably a very powerful computer, since the duration of calculations is significant)
(2) Client DC (presumably a laptop)
The MT client is supposed to have two network cards. One for the Internet and one for the internal network. There might be some problems with ODBC.
System components
(1) Database
(2) Expert
(3) MT
Database and Expert are located on a separate calculation server. Computer with name "client" hosts only MT, antivirus and firewall. This is what it looks like:
Database
Two versions of the database are allowed, Oracle or MS Access. The second option is preferable as it requires less administration and is not as demanding on system resources. In addition, it already supports a considerable amount of data for tables. The database itself consists of two subsystems:
(a) Quotes Warehouse
(b) Order Management
Plus procedures for checking data quality: whether there are gaps in quotes history, whether there is a delay in data from DCs relative to the current system time, synchronization of system time, etc.
MT
On MT all the time the windows of quotes are open, on which the trading is running with one indicator running on each of them on the minute charts. The main functions are as follows
(a) sending minute quotes from brokerage company client to the server
(b) reading instructions for order management from the database
(c) synchronizing real order management with the database
I suppose that with a correctly written dll there is no need to use an EA. In the code of the indicator the call of the dll is executed, it accesses the database, reads the instruction code, sees what it means and executes and synchronizes it with the database. Although, we may have to implement some minimal order management actions in case of server failure. We also need to figure out how the client should control that the Expert Advisor works. After all, pinging computer does not mean that the program is working.
Expert Advisor
MathCAD is a wonderful system for research, but I cannot use it for real for various reasons. First full version of Expert will be implemented in MathLAB. At the moment the porting has already started. This is much easier than transferring all code to MT. In the future I may completely switch to C/C++ implementation, respectively, as I myself am not capable of such a feat any more.
So, there are enough problems and tasks for now. I will finish TOR, and will look for a programmer to implement the necessary dll and procedures for the database.
to Alien
Thank you grasn, you have once again convinced me that I am on the right track. It is true that my potential energy distribution function turned out to be multidimensional, but on any of the orthogonal planes of the functional space it looks approximately the same. For my analysis I use a complex of neural networks, architecturally connected with each other.
The grid has to take into account the "flow" of energy from channel to channel. and define the reversal zones separately for each of the channels. After that I make a parabolic approximation and synthesize a function graph (price graph) using the wave function. The problem of strong distortions due to edge effects remains unsolved. I am working on it now - I have some ideas. The results are not impressive yet. Calculations take 5 minutes. It takes me 24 hours to train the grid.
P.S. All coincidences with your system, please consider casual. Since I can not copy because your system is not seen:).
Very interesting! I'd be curious to see it, and more detail about the overall scheme, as much as possible. Maybe you could prepare material for review, within reason of course. We'll think it over together.
By the way, I have long ago abandoned the use of neural networks of any kind. All my researches in this field did not bring any positive results.
To Jhonny
If we take away channels with minimal energy and then it still leaves somewhere. There is some misalignment, then it should be the other way round, energy between channels should be equalized and thus replenish energy of selected channels (to make them less stable). Maybe I have confused functions and functions again :).
There is no misunderstanding whatsoever. It was about the potential energy of the channel, the criteria for selecting stable channels and stability issues (I was particularly concerned about cyclists and motorcyclists :o), not about the total energy. If I calculate potential energy, it means I know how to calculate total energy as well, which I haven't told you about yet. In general, potential energy may also change in a system, especially a complex one with many degrees of freedom.
But whether it should equalize or not is a question of the chosen model. The main thing is that the theory corresponds to practice. As such a model, I chose the MSP, though I had to revise it and develop it in the right direction. :о)))
You have a very correct perception. A stone will always fall to the ground, tending to take a position with a minimum value of potential energy. If we drop it on Pluto, it will fall on Pluto, not fly to Earth. But there is another factor that is most often neglected. The stone will fall to the ground, but only if no one is interested in it. Let me explain with a simple example. A ball is thrown onto the field. It will, of course, flop, and will lie there, if not 30 players who care about it. They will be running around the pitch and kicking it all the time. The law of gravity will apply all the time, all the time the ball will have potential energy, but it is unlikely to help the ball stay in an equilibrium position until half-time, and with a minimum value of potential energy as well. And if they're not playing on Earth, but on, say, an asteroid rotating on a tricky trajectory, ...and if they're still, ...though, the guys have had enough. :о)
And there's no contradiction with the channels, the only difference being that it's a much more complex system. At intervals of an hour I "measure" the potential energy of the system, limited to a finite number of counts. Of the whole set, I am interested in the channels that have the minimum PI value. But that's for now, at this point in time they are considered the most stable if, figuratively speaking, nothing happens. But within an hour, the price has been "kicked". After an hour has passed, I measure the PE again. The channels remain approximately stable - plus or minus a few counts (you can see it on the graphs). Although there may be bifurcation points, when a channel, for example, starts to split into several ones. You also take into account the fact that the relative length of the channels has changed (the System itself is changing, i.e. it is rebuilding). Then I analyze what happened, why the channel loses its stability or vice versa, it increases. Who "took away" the energy. This is a bit (in the literary sense) like a freeze frame, in which several football players hit the same ball at the same time.
From the example given, the same channel (24 count), loses length and increases the PI level (took energy from someone) - it becomes less stable, it does not contradict anything you have said. Channel 88, on the other hand, has become more stable (reduced PI level) - it has practically won, by carrying a greater flow of energy. And channel 117 also won! It lived quite a long time, considering its original length. I have described everything in detail. Well almost everything, I should leave something for myself... You need a model, just taking the EP level directly is not enough, channels have their own length, age and so on. To summarise, they are different.
A channel that has taken more energy may burst out of greed altogether!!! But that's a working hypothesis for now :o))))
A ball lying on the ground is a "dead system", although it depends from which reference point to measure PE. If you measure it from the centre of the Earth? It's got tons of energy!!!! Can it be realised by the ball? And then there's kinetic energy. The system can give or acquire it without changing the level of PE. So, a football player walked up and kicked the ball lightly. The PI level didn't change. Are you sure it hasn't changed? What if there's a ditch in front, what if there's a bump? Unless, of course, you turn over a textbook and read the problem all the way through.
That's exactly why I wrote that you need a good model. I am not distinguished by the ability to clearly explain the first time, but believe me, there is no contradiction to the fundamental laws and your perception is excellent. And channels do not go in formation and are not peer-to-peer. If you go out in the street and see a ball flying upwards at a 45-degree angle, the thought that the laws of physics have ceased to operate will not occur to you. Chaos, after all, is difficult to predict...
My channels are not destroyed (because I can draw LR or PR on any sample) but they are not optimal. One of the optimality characteristic is PE, so as soon as it was increased it means that the channel is not optimised any more and it means that somewhere the channel got PE (or not for other channels it is another question).
Here's another thing, channel destruction is also a relative thing, my channels are not destroyed (as on any sample you can draw LR or PR), but they stop being optimal. And one of optimality charateristics is PE, so as soon as it was increased it means that the channel ceased to be optimum and it means that somewhere the channel got PE (or not for other channels it is another question).
Oh, I see, you just have a slightly different perspective on it. But I do not use LR to calculate PE. Only when I got a count, which symbolizes a channel, do I build LR. The terms, of course, are very relative and make sense for a particular logic - consider a channel "dead" or less than optimal.
Alien.
Paternoster is too broad a concept not to deal with. It includes channels, adverse tactics, level work and many other things. Basically any calculation can be reduced to paternals. Patterns are rather peculiarities of the calculation process or can be peculiarities of the language in which the trading system algorithm is formulated.
The basis of a trading system should be a trading idea, which should bring profit, never mind how and in what language the trading system should be formulated. So I am looking for it, but paternals or regressions, potential energies are not that important.
Sergey, what you propose is grandiose! And terms of reference (ToR) look mature. Impressive.
You, of course, before developing your ToR, estimated the average rate of return per transaction for the proposed trading strategy. Tell me, what is the expected profitability excluding brokerage commissions and for which currency pairs? Does it cover the spread?
Let me remind you that the sample should be representative.
Hi Sergei ! It's nice that you are still looking in here.
On the last page IronBird gave a link to a post where UP on forexclub forum publishes his mathematical proof that profitable trading is possible on forex. And (!) not as a result of breaking the Markov process, but precisely based on the assumption that it is a completely random, i.e. Markov process.
I have now read this post and have mixed feelings. On the one hand, I see places where the author uses in my opinion incorrect assertions. On the other hand the way of thinking seems to me interesting. Probably, these discrepancies are of technical, computational character and do not cancel the correctness of the general scheme. Besides, alas, there is something that is not clear to me at all.
I wonder what you would say about it.
I also wonder what the Northern Wind would say, if it blows here sometimes.
The link is http://forum.fxclub.org/showthread.php?t=22097&page=3
For information. I have already implemented the computational scheme you suggest. However, my goals were slightly different, namely to stay in the familiar software environment, but the philosophy is the same: divide and conquer.
I can share with you an Expert Advisor that will implement the exchange, with the purpose of joint testing and so on. Truth I haven't made dlls and made file swaps - a more universal thing, as it seems to me.
However, he sells his EAs but he is a good programmer and I don't need to work out anything. I don't really like it when you use files.
Since you have decided to pay anyway, it may be interesting.
What about the database? A database is good for smart selections and simple price series runs with no finesse.
Good luck.