Zigzag indicator and neural nets - page 3
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 still don't understand why everyone wants to zigzag the input. It makes more sense zigzag is an exit, something to be trained for.
Then you will find it difficult to do without the zigzag.
All this talk about Zigzag... Makes one think that the trading systems market will be a sure thing. (And of course, without Zigzag).
Alex-Bugalter
You are not paying much attention. Read what I wrote and think. And if you are going to build a system whose main tool is the NS, read the theory of recognition. NS is just a tool and in order to use it correctly you need to have knowledge in this area.
You should not build waves from the top down, but from the bottom up. And build on the structure of the waves. Thus moving from a small wave to a larger wave.
Where this structure should be - at the input or output of the NS - is another question. Together with Prival, I am inclined towards the output. But, frankly, I do not yet consider the NS as a tool for solving my problems.
Alex-Bugalter
You're not paying much attention. Read what I wrote and think. And if you are going to build a system whose main tool is the NS, read the theory of recognition. NS is just a tool, and you must have knowledge of it in order to use it correctly.
I'm afraid it's you who misunderstands me. When it comes to zigzags, wave theory is automatically implied.
Using a pure zigzag, as a reference in a system, is not very acceptable, though it is beautiful. It is already a practice. It's a pity, but there's not a line about it in the books. It would have saved me a lot of time.
Not to build, but to build.
The only difference is that there is no adherence to the dogmas of wave theory.
However the main principle is the same, the strict adherence to the wave structure, regardless of whether it is a 3-wave or 5, 7, 9, etc. The main thing is that the network should understand what it is supposed to do.
And what about the network output, well, a pure extremum as I wrote above is certainly beautiful, but we should like to see some confirmation that the trend has finished.
If we look at the zigzag vector as a single trend. Here it is better to look for confirmation on smaller periods, as you have noted earlier.
Regarding the grid inputs.
Prival, the remark is perfectly legitimate. IMHO zigzags in general are not suitable for trading in their pure form. In fact, a zigzag is overdrawn and therefore it does not reflect the current situation. But zigzags wrap history very well and compactly, they save its most interesting moments - its extremums. So, I am planning a heterogeneous grid, in the sense that it should contain signals from different sources. Namely, several zigzag extremums, several wavelengths and several last bars. In this case, it seems to me, we will obtain a perfect coverage of the history. Zigzags with more coarsening will show long term history. Bars will show the medium term. Bars will show the current. So the independent use of zigzags as grid inputs is out of the question in any case. It is simply considered as an excellent mechanism of history compression.
A little about my experiments. My zigzag is different from the ones included in ZUP. The algorithm is described in the source code zigcat.mq4. Unfortunately, there is no guarantee that the version for matlab is the same as for mql, but I've made some very interesting experiments with matlab. I've already written about it briefly here. I was interested in how many segments the zigzag would consist of, depending on the step, and at which step the number of segments increases drastically. These are some funny pictures:
The number of segments from a step:
Its derivative (change in the number of segments) from the step.
The step here is shifted by 10. In other words step 0 in the picture corresponds to step 10. We took 5 000 one-hour candlesticks on EURUSD. The first picture turned out to be quite expected. Indeed, long movements surely occur less frequently than short ones. But the second one was a bit surprising. Pay attention to the spade around 30. It's pretty consistent from sample to sample. Don't forget that the pitches are offset by 10 and in fact the pitch is 40. This suggests that on hourly EURUSD movements with a swing of 40 are more likely than movements with a swing of 30. Very funny result, if it's not a glitch related to a bug in the program. I will check and recheck this case.
And finally, for those romantics on the high road who want to play with all this a bit, I'm pasting the source code for Matlab and mql.
The zigzag will always be there. It's just that sometimes you see it and sometimes you don't. :)
There will be a zigzag, but it will be useless. After all, a zigzag is only a coarsening of initial data with a lag. There is nothing more.
You probably just don't know how or don't want to cook them:)
I could cook it, only it's inedible:)
Zigzag will always be there. It's just that sometimes you can see it and sometimes you can't. :)
You will see the zigzag, but it will be useless. In fact, a zigzag is just a coarsening of initial data with a lag. There is nothing more.
You probably just don't know how or don't want to cook them:)
I could cook it, only it's inedible:)
Not a coarsening, but an exclusion of noise. What is noise is for everyone to decide for themselves. Not all zigzags are lagged.
I disagree. MA can be used to exclude noise. A zigzag does not exclude noise. It just finds historical extremums. And it takes some time to tell if there was an extremum in the most recent history or not. Depending on the size of an extremum, the lag time can be smaller or longer, but it always exists. On the basis of the zigzag technique it is in principle impossible to make a forecast, we can only state the fact that there was an extremum of such and such size some time ago.
If the Zigzag is not delayed, it is no longer a Zigzag (or simply a matter of terminology).
SK, I agree with that 100%. And in general, I see that the zigzags are usually used by fans of Elliott Waves. And to this theory I concern, to put it very lightly, rather cautiously. However, the zigzag has no equal in data compression. I think you have to admit that too.