Stochastic resonance - page 27

[Christo Tsvetanov]

Mathemat:

lna01 wrote (a): the potential relief is floating all the time

I'm afraid it's not floating, but jumping: roughly speaking, during the lifetime of the channel the system can be tried to be described as closed (with conservation laws), but as soon as some critical parameters of it change, the system suddenly jumps to a new state and everything starts all over again. But this model is not like SR at all, since the flattening of the channel is already a transient process (catastrophe) and not quasi-stationary.

I would add - the width of the channel changes all the time too... Haven't found a better market model so far :(

By the way - I remembered my last year experiments with linear regression - there it turned out that when the trend starts, suddenly the state of the system is ordered - Pearson's coefficient started to get close to 1.

[Victor Nikolaev]

Yurixx:

I do, of course, but it's called 'through the back door'. I hope it's a temporary measure. As soon as it's working properly, I'll put the pictures in where they belong.

The theory is interesting and similar to the solution to my problem too. But it's getting a bit hard to make diagrams by formulas in my head. Maybe there is a draft in a more convenient form. But the theory is quite interesting.

[Prival]

First impression, I'll get into more detail later. Look at the Rayleigh-Rice distribution. It seems to fit. Most likely, if we subtract the LLG, we get the Rayleigh distribution, then we are left with one sigma parameter to describe the distribution. I have to run off to work, I'll check it when I have time. Too bad you can't see the histogram. If you don't mind, you may see it by its appearance.

[Avals]

Yurixx:

And finally, why all this is needed.

While I was at it, I saw several opportunities to use it all.

IMHO, there is no value in a single distribution for price. At different points in time, different distributions. For example, the same channel is a temporarily stable distribution of a certain kind. At transitional points, it is impossible to determine which one it will turn out to be, there are always alternative scenarios. We enter just counting on some particular distribution with positive IM (or use it to get + IMO). Of course we can mix them all up for an arbitrary range and pick a similar distribution, and then use as you suggest: reduction to a universal standard, normalization... But the purpose of indicators and other tools is to detect the moments when a particular distribution may occur (or continue), which we can use, but not to predict some global market condition. For this purpose an indicator shouldn't mix several previous distributions, and it's not clear with what period: we took some from this distribution, some from the other one and some from the third one. IMHO, there is a need for separation, perhaps even post factum, but not mixing. And then either use the revealed distribution in the hope that it is still preserved, or wait for the new one at transient moments and use it. In the latter case, the previous (completed) distribution may also determine the potential for the future to be used.

[Neutron]

Yurixx писал (а):

И в заключение, зачем все это нужно.

Пока я занимался этим, я увидел несколько возможностей использования всего этого

Bravo, Yura! Good work.

I solved a similar problem a year and a half or two years ago. The need was also dictated by the problem of normalization of indicator readings at different timeframes. Unfortunately, I failed to find an elegant analytical solution of the problem and limited myself to empirical dependences, which turned out to be relatively simple. Now I can use the results of your work. Thank you!

to Grans

Sergei, please take a look at the noise power indicator. It detrends the initial time series step by step and then smoothes (FLFPeriod) the square of the amplitude. This indicator reacts well to changes of the market "mood", especially on the minutes.

[Rashid Umarov]

Yurixx:

I do, of course, but it's called 'through the back door'. I hope it's a temporary measure. As soon as it's working properly, I'll put the pictures in where they belong.

PS

Alas, even that doesn't catch on.

Moderators, HOOOOOOOOOOOOOOOOOOO!!! Fix the site, pls. No picture to attach, no file ...

Thanks for the message, we'll correct it.

[Prival]

Neutron:

Sergey, please take a look at the noise power indicator. It detrends the initial time series step by step and then smoothes (FLFPeriod - smoothing period) the square of the amplitude. This indicator reacts well to changes of the market "mood", especially on the minutes.

I looked at the indicator. It seems to trace the history only. I put it on the chart. There's nothing. The indicator value is 0 all the time. In visual testing mode, it's also 0. Until you stop it. As soon as you stop it, it'll show up. Any tips on how to use it?

[Neutron]

Well, it depends on what is required. I needed to discuss some noise issues with Grans, so I needed a noise power indicator like this. That's actually what it does (at best), but the rest is a bit of a think-about-it-all.

[Yurixx]

The pictures have been inserted.

Neutron:

Bravo, Yura! Good work.

I solved a similar problem a year and a half or two years ago. The need was dictated, as well, by the problem of normalization of indicator readings at different timeframes. Unfortunately, I failed to find an elegant analytical solution of the problem and limited myself to empirical dependences, which turned out to be relatively simple. Now I can use the results of your work. Thank you!

Thank you Sergei, the assessment by a statistician is particularly pleasing.

There, at the end of the first part, I had a question. Maybe you'll be able to answer it?

Vinin wrote (a):
The theory is interesting and similar to the solution of my problem too. But it's already getting a bit heavy on formulas to graph in my head. Maybe there is a draft in a more convenient form. But the theory is quite interesting.

I'm not quite sure what draft I'm talking about. Actually, I wrote it all in wordpress, because I noticed that the local editor allows you to insert text in the upper and lower indices. So I had to convert my Wordboard file so that it is a copy of what's posted.

There's something of interest to you in this method. Unfortunately, it slipped my mind what exactly it is, but reading your posts about EA for the Championship, I've paid attention to it. If I remember, I'll tell you. :-(

I remember ! If you pay attention, the distribution of the kind I used has a stable MO<SCO. I remember in some post you wrote that this detail interfered with something there. I don't know what the point was, but I assume that in my case this ratio will hold for any set of parameters. Nevertheless, perhaps the solution to your problem still exists, you just need to find an adequate method. Maybe simplicity and obviousness of properties of this distribution will help to do it on model first.

[Yurixx]

Rosh:

Yurixx:

I do, of course, but it's called 'through the back door'. I hope it's a temporary measure. As soon as it's working properly, I'll put the pictures in where they belong.

PS

Alas, even that doesn't catch on.

Moderators, HOOOOOOOOOOOOOOOOOOO!!! Fix the site, pls. No picture to attach, no file ...

Thanks for the message, we'll correct it.

Thank you for your promptness.

There are, unfortunately, still problems. They may not be yours. I have IE6 with the 2nd service pack. Here's what's going on. in the post editor.:

The Delete key doesn't delete a character, although Backspace backspace does without issue. Using the arrow + shift, to select the block, not always works, the cursor just does not move in any direction.

Copy-paste from another window has stopped working for some reason. In order to reply to several people in one post I tried to copy and paste a quote from another window. To no avail. Moreover, it doesn't even paste a single word.

The quote works crookedly in the Styles combo box. Select text and then changing the style, make it a quote it manages 1 time out of 3-4.

And in general, it would be good to improve the editor. In this forum it is the weakest point.

Thanks for the collaboration.