a trading strategy based on Elliott Wave Theory - page 221
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
This is retrospective. It's not like I'm running a strategy on a tester. I just have the same indicator whose signals
need to be filtered out. On history these signals are known, in real time it takes time to identify the signal
. If this lag time is taken into account, perhaps this advantage will disappear completely.
At best it will be significantly reduced.
I think I understand what you are trying to do.
What you have, judging by your actions, is a kind of simple neural network.
You input two parameters, output one...
What you have, judging by your actions, is a kind of simple neural network.
You input two parameters, output one...
So, you can also make a neural network into a function f(a,b)={return(a+b)}
Maybe it makes sense to draw points in plane BR-dA and BL-dA, where dA is amplitude of price movement? The points should also be coloured in two colours - depending on the result of the trade. Only, it seems to me, you should not take into account the commission from each trade at this stage - the result will be more transparent.
So, the function f(a,b)={return(a+b)} can also be subsumed under a neural network.
That's about right, in its essence.
You could, of course, put it that way. The problem, however, is that in order to build this "neural network" you need at least statistical separability of sets. And, judging by the picture, it turns out that unsuccessful inputs are almost uniformly distributed over the set of successful ones. I had hoped for something else. But still we need to look at the numbers.
Maybe it makes sense to draw points in the plane BR-dA and BL-dA, where dA is the amplitude of price movement? The points should also be coloured in two colours - depending on the result of the trade. Only, it seems to me, it is not necessary to take into account the commission from each transaction at this stage - the result will be more transparent.
I am not taking into account the commission at all for the time being. First we need to show the possibility in principle, and then the practical implementation.
I do not understand what sense the construction of points in coordinates BR-dA and BL-dA may have. Or these coordinates themselves. And what do you call the amplitude of motion. Explain.
And I can colour it. :-))
By the way, Sergey, you've recently mentioned that you have tick archives of several pairs. Could you share the EURUSD archive? And I would be also grateful to you if you could view the Hearst's indicator in its current form. In return, I can send you the indicator of fractal dimension, it is simple but corresponds to its idea. As you know, fractal dimension and Hurst are related by a simple relation. One could check to what extent it is fulfilled in this case.
You could, of course, say so. The problem, however, is that to build this "neural network" one needs at least statistical separability of the sets. And, judging by the picture, it turns out that unsuccessful inputs are almost uniformly distributed over the set of successful ones. I had hoped for something else. But we still need to look at the numbers.
I have not yet seen a clear division in such cases. Well, maybe except in one case. Mostly it's a 50/50 split, plus or minus 2-4%.
I am not taking into account the commission yet. First we need to show the possibility in principle, and then the practical implementation.
Please read your post on p.110:
Instrument - GBPUSD, M15
Amount of bars on the chart - 38856
Potential entry points - 4038
Blue points - successful entry, red - unsuccessful
Criterion - entry that has brought +6 points or more (correction for spread) was considered successful,
others, including positive ones - unsuccessful.
Total successful entries - 3482, unsuccessful - 556
From my point of view, the only thing I can say about this picture is that using phase plane
of BR,BL indicators does not allow to build a good entrance.
Assuming that spread and commission in this case are one and the same, then you do take into account the spread.
I don't understand what sense the construction of points in coordinates BR-dA and BL-dA can have. Or these coordinates themselves. And what do you call the amplitude of motion. Explain.
And I can colour it. :-))
Read your post on p.108.
The phase space of the system is larger than this plane. At the very least another dimension should be added here - the value of the price change that occurred over some (in each case different) period of time. The idea of BR and BL is that when BR is superior, one should sell, and when BL is superior, one should buy - this is clear. The check of this hypothesis should show that the price changes in the right direction when there is such an advantage.
Thus, there should be two points of 2 colours on the chart, red and blue. The red ones are where the hypothetical trade was successful, the blue ones - otherwise. It's good not to lose the values of price changes. Not every change will satisfy me, after all. Obviously it can be represented by the bar histogram, where the red bars are above the plane (positive direction), and the blue ones are below it.
If we assume that "amplitude of price movement" and "values of price changes" are the same in this case, then the meaning I put into it does not differ from yours, and judging by the context of your post, you know the reasonability of its addition to the analysis:-)
By the way, Sergey, you've recently mentioned that you have tick archives of several pairs. Could you share the EURUSD archive?
If for some reason you are not satisfied with the archive, the link to which was kindly provided by Northern Wind:http://ratedata.gaincapital.com/, I promise to immediately post mine. By the way, if it is important, it is from a real (I mean not demo) account
And I would be also grateful to you if you could view the Hearst's indicator in its current form. In return, I can send you the indicator of fractal dimension, it is simple but corresponds to its idea. As you know, fractal dimension and Hurst are related by a simple relation. One could check to what extent it is fulfilled in this case.
Let me remind you that the Hurst index(h), FAC and H-volatility(as it is defined in Pastukhov's thesis), are connected by the obvious correlations:
FAC=1-2/H=2h-1.
Recall that FAC can be defined as the difference between all co-directional price jumps (moves) and counter-directional moves, divided by the sum of all moves. I see the advantage of using FAC as a simple form of expression for estimating the average return of an instrument:
s=FAC*sigma, where sigma is the standard deviation.
Besides, the Hurst index h has a clear interpretation as an index of time in the expression that connects value of standard deviation(sigma) with TF:
sigma(t1)=sigma(t0)*(t1/t0)^h.
From this expression it is not difficult to obtain an estimate for the Hurst index for the selected TF, see fig:
It should be noted that according to the above algorithm, the correct estimation of the Hearst ratio should be done in a two-point scheme - to the left of the TF of interest and to the right. The desired value of h can be determined as an arithmetic mean. Generally speaking, it is not quite clear to me, why should I get involved in such a scheme, if the Hurst index can be easily expressed through FAC or H-volatility?
http://vtsystem.narod.ru/market.htm
http://vtsystem.narod.ru/market2.htm
http://vtsystem.narod.ru/market3.htm
http://vtsystem.narod.ru/market4.htm
http://vtsystem.narod.ru/strateg.htm
Thank you very much for the links! It was very interesting to read! The author of the website suggested representing prices in different coordinate systems, as well as comparing different methods of representation.
It would be cool if an article like this appeared on www.mql4.com. It would be interesting for both beginners and experienced traders who have not read this information yet. If somebody (for example komposter) could write a script (expert) that could display charts in different coordinates in MT4 (in offline mode of course), then such a script (expert) would be a hit of downloads in CodeBase section at once!!! :o)