Not Mashka's business! - page 4
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
to Neutron
It's really not that simple, the picture shows one of the best plots, statistically it's not so good, and therefore it's very risky to use this approach. To get the forecast MA (no matter what method) is not enough for trading, we should estimate the price behavior around this MA. If we rebuild the price values directly, it will cause significant errors. I've tried to calculate stable "restored price" levels but I wasn't completely satisfied with the result - though there was some profit, I won't make any definite statement about its stability.
to Neutron
Seryoga, maybe I don't understand you very well. Here is an arbitrary section (only H and L are shown):
In other words, only the constraints of the process development are shown. Yes, there is also an "input" and an "output" for each datum. Seryoga, can you tell where "sits the real", maximally close to some nebulous truth process?
OK, let's move on and explain. For this process, it's taken
Calculated ACF for them:
And ACF for the first differences:
Another question: What are the differences? What empty correlation are you talking about? What 10%???? Maybe I'm doing something wrong?
to Neutron
It's really not that simple, the picture shows one of the best plots, statistically it's not so good, and therefore it's very risky to use this approach. To get the forecast MA (no matter what method) is not enough for trading, we should estimate the price behavior around this MA. If we rebuild the price values directly, it will cause significant errors. I tried to calculate stable "restored price" levels but I wasn't satisfied with the result, though I had some profit.
I have a strong feeling that Seryoga was predicting ACF, and is going to recover BP by it. But I could be wrong.
As far as I understand it, the variant of Neutron's approach described actually predicts not the price level, but the time of the maximum.
I have a strong feeling that Seryoga was predicting ACF, and is going to recover BP by it. But I could be wrong.
To get the price level in this variant, you need to summarise several increments. If an error is not normal for them (increments), it is possible to get quite far away (from the target).
P.S. Yes and if normal too, the statistics will not have time to play on a relatively short interval.
it was a thought out loud and I think it was my thought :o)))
PS:
Прогноз АКФ - это что-то новенькое в трейдинге...
I admit, not only did I stall, but I also overreacted. It happens :o)
I still don't fully understand it:
This is an attempt to predict the first derivative of an ideal МА (the one without lag and whose first derivative gives absolutely accurate entry/exit signals, shown in the figure with a solid red line). I did it this way: I stepped back on the timescale from the current value of the derivative by N bars and found the weighted sum of n common MA derivatives (those with lag), equated it to the value of the red line and solved a system of linear equations for weights. Then knowing these weights I multiply the current values of the lagging MAs by them and obtain the predicted value of an ideal MA "for now".
In fig. I have made a forecast for six bars ahead. We can see that the forecast does not lag behind an ideal MA though somewhere it differs considerably from the last one in terms of amplitude, i.e. the idea turned out to be not absurd, though an attempt to extend the forecast horizon leads to "scattering" of the forecast series. I've attached a animation below showing the stability dynamics of the forecast while trying to increase horizon from 0 bars to 10 (one frame - one bar+).
Is it the first derivative on 200 samples that changes like that? And what is the quote? Seryoga, please explain.
to Neutron
Seryoga, maybe I don't understand you very well. Here is an arbitrary section (only H and L are shown):
In other words, only the constraints of the process development are shown. Yes, there is also an "input" and an "output" for each datum. Seryoga, can you tell where "sits the real", maximally close to some vague truth process?
OK, let's move on and explain. For this process we take
Calculated ACF for them:
And ACF for the first differences:
Another question: What is the difference between them? What empty correlation are you talking about?
I don't get it! My way of looking at it is this:
The autocorrelation coefficient was plotted for the first difference series High, Low and (High+Low)/2.
I screwed up about ten percent positive autocorrelation, but nevertheless the effect is noticeable.
Is it the first derivative in 200 samples that changes? What is this quote? Seryoga, please explain.
Oh, how interesting!
Here's what would happen if the adder inputs dipped not on the lagging masks, but on itself :-) i.e. on the first derivative of the ideal MA.
This is already a forecast for the width of the window, i.e. for 100 bars!
Maybe there 's an error in the code... It cannot be.
Below is an avid with the dynamics of horizon increase from 0 to 100 bars.
to Neutron
That was a long time ago, can't remember where it came from, some book, but it's also one way of estimating autocorrelation:
although... I'm not much of a mathematician :o)))
Ну, да, первая производная от двойного прогона ФНЧ (вперёд-назад) с окном усреднения 100 баров. Окно слегка условно, поскольку для анализа я использовал МЕМА, но это не принципиально. Использовался ряд минуток EUR/JPY.
I've already admitted that I'm a bit over the top. I couldn't hide this medical fact for long :o)