a trading strategy based on Elliott Wave Theory - page 222
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
My bad, I didn't understand your form of writing. As the result I took a hyphen as a minus sign. That's why I did not understand the reason of repeating what was done for the difference between the indicator and the price change, and what the point was. But to draw it on coordinate planes (BL,dA) and (BR,dA) is an understandable and reasonable idea.
You are also right about the spread. I corrected for it in this case for two reasons. 1) Programmatically it was easy to do. 2) The principle possibility of obtaining statistical superiority under the conditions of my experiment had already been obtained by me. So I found it more interesting to cite data adjusted for the spread.
I'm fine with the link to GAIN Capital and I'll download the data from there, but if your data is from another broker, I'd be very grateful for it. I suppose the system has to be completely robust to the data flow. There is no doubt that they differ from broker to broker and on our real accounts we do not get a pure image of the market, but more or less a similar representation of it. Since in Forex there is no pure market as such, this problem is fundamentally unsolvable. That is why there is only one way out - to make stable systems, i.e. systems capable of filtering out the main thing from the details. And the easiest and most reliable way to check this is to test them on data from different sources.
I personally spent 3 months collecting data from my broker. That was enough for my research last year. However, now it is no longer enough. Collecting is long, it's easier to find someone who has already done it. :-)))
You may be right. The point, however, is that Hearst is relied upon to identify the nature of the market - trending or counter-trending. It does not make sense to calculate it on all available data, as the market behavior changes in time, and only due to this it is possible to profit from it. That is why the variant of calculating the local Hearst's value is of interest. Using it in such a context, which I have seen, deserves attention. But if you bind Hearst calculation to FAC - integral by definition characteristic of time series, then you won't get any localization. IMHO.
In your branch on fxclub ForAxel gave pictures of sets that not only can be considered sufficiently separable, but which also have pronounced localization centres. I don't know just they reflect some real data or it's just a searching program for now.
Sergey, maybe you can share how to process large volumes in Matcadet, if you have such experience?
Everything is laid out separately for each currency pair, by year and by month.
As a result, the entire database consists of a pile of monthly tick archives. Download only what you need.
So, it's quite possible to try ticks for one month only. That's about 30-40 thousand points.
Sergei, try it on that many, it might work.
It's all laid out separately, for each currency pair, by year and by month.
As a result, the whole database consists of a bunch of monthly tick archives. Download only what you need.
So, it's quite possible to try ticks for one month only. That's about 30-40 thousand points.
Sergei, try it on that many, it might work.
It works at this number. I have 40 000 records now. But I need much more for the experiment
"MQL4: A picture for the metaquotes forum",
I found out that the maximum file size is 1 Mb :-(.
I've got 60 Mb of EURUSD data for a year... And 5 Mb if I cram it all in.
Give me some advice.
to Yurixx
The way I proposed to calculate the Hurst index uses integration over history as well as the others. Really, we need to find the value of volatility of an instrument beforehand and it is the sum over history. So, the problem will remain. Besides, we will have to find volatility in two points on different TFs and then take the logarithm of the ratio and this procedure will only increase noise.
The figure shows how the FAC reacts to the behaviour of the market. From EURUSD 1m 2005 a renk-row with a discreteness of 17 points has been synthesised (in the figure in red). The resulting series was analyzed for "trendiness" and "reversal" using FAC with a sliding window of 100 bars. Recall that the FAC has a range of values from -1 to 1, the negative value suggests that after a profitable trade should immediately open in the opposite direction, and positive - to open in the direction of price movement.
I've found out that the maximum file size is 1 Mb :-(.
I've got 60 Mb a year on EURUSD... And 5 Mb if I cram it all in.
Give me some advice.
1. You can divide your 5Mb into 1Mb parts using WinRAR, change the extension from rar to zip, because forum does not accept RAR files and then put them into www.mql4.com. You can see an example on page 26 of "MQL4: a picture for metaquotes forum".
When you download the files back, you will not need to change the extension from zip to rar, since WinRAR will unzip the files with zip extension as well.
2. Upload the archive to www.filefactory.com for temporary download
When downloading the files back, change the extension from zip to rar is not required, as WinRAR will unpack the files with zip extension as well.
2. Upload archive to www.filefactory.com for temporary download
Thanks!
Here, put it on http://www.filefactory.com/file/aef4cf/
to Grans.
Sergey, pay attention to the picture from my previous post. The moving window size is 100 Renko-bars, it means that phase delay due to the averaging procedure on historical data does not exceed half of this value, i.e. 50 bars. The characteristic period of market volatility (see fig.), is about 300-400 bars! Thus, we can state the fact of REAL identification of the trend (deterministic) on the time series with the use of renko-construction! This has never been possible with a classical FX time series and has never been reliably positive on all FQFs.
Here you go.
Downloaded fine. Thank you!
It's a pity the data recording format changed during the accumulation process:
2006.04 0.1 3 1144639388 1.2105 1.2108
......
EURUSD 2006.08.10 15:59 1155225540 1.2829 1.2831
......
2006.10.02 14:00 1159797628 1.2691 1.2692
Well that's basically quite correctable.
It's a pity that in the process of accumulation the format of data recording changed:
Well, in principle this is quite correctable.
It is not a matter of principle.
Cross-cutting time in seconds: A(i,3).
Bid prices: A(i,4).
Ask prices: A(i,5).
This is true for ALL the file!