Stochastic resonance

 
Interesting article http://elementy.ru/lib/164581

In a nutshell with a quote

In 1981, two groups of physicists - one in Rome led by R. Benzi, the other in Brussels, led by C. Nicolis - independently of each other - proposed to focus on the general features of the behaviour of the climate under the simultaneous influence of weak periodic and strong chaotic forcing. By constructing a simple mathematical model and studying it, they discovered an absolutely striking - and at first sight even unnatural - phenomenon. It turns out that noise of a certain intensity not only does not disturb but even helps a weak disturbance to manifest itself in the system's response. The phenomenon is called stochastic resonance. The word "resonance" here means an unexpectedly strong response of the system, and "stochastic" reflects the fact that the cause of this effect is chaotic action, noise.

Description ...
and conclusion

Thus, the Earth's climate is a kind of system which, under the simultaneous action of strong chaotic and weak periodic forces, regularly "switches" between two relatively stable states. Now we can make a standard transition in theoretical physics: forget about concrete situation (Earth, climate and glaciers) and focus on the most general features of the phenomenon. In the language of theoretical physics, the model constructed is called a stochastic bistable system with a forcing force.

Very similar to price movement, there is some periodicity (however periodicity exists), there is noise (volatility ?), two stable states (TRENDS ?) and there is a theory trying to explain THIS. Couldn't we try, as NEO from the spider elegantly put it, using this theory to drag plushies out of the bazaar ?
Of course, the distance from the popular science article to Forex practice is ... But I always want some candy.
Maybe the community will have some thoughts? Eh, and maybe it's not news, then excuse me. Good luck.
 

Very curious. I came across something similar to resonance, when I was designing exotic MAs with sign-variable coefficients. Here is a picture:

I can't remember the formula to calculate it now, it was long ago. Here is an example of how to calculate such a МА with period 7:

Alt_sign_LWMA(i,7) = ( 7*p(i) - 6*p(i-1) + 5*p(i-2) - 4*p(i-3) + 3*p(i-4) - 2*6*p(i-5) + 1*p(i-6) ) / 4

Here is a period of 55 on the chart to make it clearer, I don't know how to interpret it. Black shows the candlesticks on the daily chart, blue shows the MA.

 
Mathemat:

Very curious. I came across something similar to resonance, when I was designing exotic MAs with sign-variable coefficients. Here is a picture:

I can't remember the formula to calculate it now, it was long ago. Here is an example of how to calculate such a МА with period 7:

Alt_sign_LWMA(i,7) = ( 7*p(i) - 6*p(i-1) + 5*p(i-2) - 4*p(i-3) + 3*p(i-4) - 2*6*p(i-5) + 1*p(i-6) ) / 4

Here is a period of 55 on the chart to make it clearer, I don't know how to interpret it. Black shows the candlesticks on the daily chart, blue shows the MA.

Do you have a similar machine for MT4? It's a funny picture, I'd like to see it on a longer interval...
 

It's easy enough to write it, Figar0. The coefficient at the bottom is MathCeil( Period/2 ), where Period is the period of the MA. I'm just not really interested in it at the moment, but the picture is really funny...

 
Mathemat:

It's easy enough to write it, Figar0. The coefficient at the bottom is MathCeil( Period/2 ), where Period is the period of the MA. I'm just not really interested in it at the moment, but the picture is really funny...


I've got it a bit wrong for some reason.

 
Mathemat:

Here on the chart is a period of 55 to show more clearly what is coming out here. How to interpret it - I don't know yet. Black shows the candlesticks on the daily chart, blue shows the MA.

Nice picture. Where the resonance is wobbling, it's flat, where it's clear, it's trending.
 
Vinin писал (а): I got it a little wrong for some reason.

Yeah, I can see that. There's a lot of delay in yours. Mine has almost no lag (and it should be). And I forgot, that we must take a module from the entire expression, because an even period may cause a negative value. Can you post the code? I was making it for Trading Solutions and it's quite bad with the code there, it's inconvenient because all functions are written in prefix form. Well, let's say, a+b - as Sum(a,b). You have to sort through all those parentheses there.

P.S. I see that I have posted it. Thank you.

 
timbo:
Mathemat:

Here on the chart is a period of 55 to show more clearly what is coming out here. How to interpret it - I don't know yet. Black shows the candlesticks on the daily chart, blue shows the MA.

Nice picture. Where the resonance is wobbling, it's flat, where it's clear, it's trending.


Corrected the error in the indicator, the indicator is attached.
Files:
 
timbo писал (а): Nice picture. Where the resonance is wobbling, it's flat, where it's clear, it's trending.
It seems to me that it only seems like that. Probably, to confirm trend/flat condition we need to take several of them with different periods.
 
AAB:
Interesting article http://elementy.ru/lib/164581
Read the article, very interesting. There's a lot to think about.

So, what do we have? There is noise - quite strong: volatility. There is a weak regular signal (hardly periodic, but it is definitely there). Weakness of the regular signal is confirmed by a very low return value in comparison to the volatility itself even on strong trends. I already gave these values somewhere by the example: EUR goes ascending trend on daily data for 6 years that is about 1600 daily bars. During this time period EUR has gone through 6000 points. So, mathematical expectation is less than 4 pips (regular low impact). At the same time the volatility on the daily bars is about tens of points (noise).

The steady states are flat states at the tops during reversals or corrections. Trends are unstable states of transition from one flat to the next. Before a trend, a regular signal is amplified by flat noise and appears as sharp, often momentary jumps from level to level.

How can we learn something practical from this?

P.S. For example, how can we extract only the random component (pure noise) from volatility to get a regular signal? Volatility is known to be an antipersistent process. Simply subtracting a constant from it will not work, as the signal is getting stronger during the trend. Detrend? And what, I wonder, is the amplification coefficient equal to?
 

It feels like it somehow resonates with the potential models, or rather my view of where and how to use them :).