Tendential planimetry method - page 8

New comment
 
vaa20003:
It would be interesting to see a picture where the MA is built not only "from below" but also "from above". "Flip" the price and build the MA.
This is so fantasy............a suddenly....

It may be considered that ordinary MAs represent the state of inhibited bears, and antipode MAs represent the state of inhibited bulls. It is theoretically feasible to build, but the picture may be disappointing - 1/x is a highly non-linear transformation. It's still interesting to look at, of course.
 
lna01:
vaa20003:
It would be interesting to see a picture where the MA is built not only "from below" but also "from above". "Flip" the price and build the MA.
This is so fantasy............a suddenly....

It may be considered that ordinary MAs represent the state of inhibited bears, and antipode MAs represent the state of inhibited bulls. It is theoretically feasible to build, but the picture may be disappointing - 1/x is a highly non-linear transformation. It is of course still interesting to look at.

So it seems to be enough to do a normal waving shift-period.
 

to Gep

Привет.

This algorithm is good and doesn't seem half-hearted. If we write the gradient in arrays and use it to build envelopes, it will be more clear and practically applicable.

If estimate the algorithm for calculation of 50 bars, it is quite usable (100x50=5000) in 2 loops and arrays for analysis.

But still try to build envelopes on the gradient and take boundary bars as borders. I wonder what the picture will be.

Good luck with the trend and big profits.

Hi! :о))))))

This algorithm is the first thing that came to my mind and it surely doesn't pretend to be the best. I apologize for my technical illiteracy, but I did not understand anything here: "If you write a gradient into arrays and use it to build envelopes". About the speed, all right, but it is not the main problem, I'm trying to push Master Yoda to the idea that it does not need to do, but he stubbornly continues to attribute MA mystical properties, loudly beating a drum and howling a long song of all Jedi.

Likewise, good luck with trends and big profits... and good health to spend them :o).

to Mathemat

Eh, I have long suspected that all these smooth functions on prices - with their discontinuous graphs - will not lead to good. By manipulating smooth mashups, we are essentially trying to build a continuous model of a fundamentally discontinuous process. Reality says: "prices are discontinuous, tails are thick, and the market is non-linear", and we ignore this as we internally protest against a chaotic, discontinuous reality with its black swans and clearly non-linear disasters (trends): ...

How true, Mathemat, that's why I switched entirely to the theory of fractals and borrowed some ideas from the theory of catastrophes.

And you drink Pu-er-Cha. It's my favourite tea at the moment (after beer). It smells rotten when you get used to it, but you get used to it...

I guess it's the strong tea that gets you used to it fast. :o) Tell me, is it green or black?

to eugenk

Well... I'm not supposed to ask such things to Thunderbolt and the Death of all DCs (sorry, I promised not to joke like that :)))))))))). You know, all density estimates are rationed.

Breaking convention? Are you calling me names? Then I won't call you a litter cat, I'll call you a fat, lazy, house cat. :о))))

Normalisation is nothing of the sort. If you change step of MA, worms will start simply "walking" on price field, while increasing number of MA no algorithm will cope with reliable check for worms - you will have them virtually "everywhere", will merge into one big worm however.

I'm not going to argue point by point for now, but ask one simple but very important question, which I asked, but didn't get an answer, and so on:

About the main thing.

What is the mission statement? What are we looking for, in other words, WHY do we need these worms. If you just want worms, just say: "Seryoga, I want worms". There are plenty of good places where you can dig for them, buy them, uh ... well, get them.

About those last pictures. Sorry, I'm dumb, but I still don't know what you were filtering. Densities ??? Anyway, it looks a lot better than my pictures from the script. It's not cluttered with unnecessary lines, and there's something to look at. The only question is the veracity of the method itself... And here... Learn to learn and learn, as Grandpa Lenin bequeathed us :)))))

I'm also not at my best mentally, but clarify - what is your definition of density? Is it a term "by definition" or is there a clearer wording? And the algorithm itself is very simple and is designed to isolate these worms. You actually have a two-dimensional array of "X - time" and "Y - MA values" points - nice pictures. Well, for example, from your picture to point 1: on July 12 (we have fixed time), moving from the bottom edge of the price to the top edge, you find many points of different MA periods, all dated July 12. For this series I find the differences MA[n]-MA[n-1] (here n is the index of all vertical points with fixed X). For the obtained array I determine the standard deviation. Once again I return to the vertical array, whose x-coordinate is equal to 12 July and consistently check the condition (MA[n]-MA[n-1]<k*SCO)&&(MA[n]-MA[n+1]<k*SCO). If this condition is fulfilled, then I leave MA[n] corresponding to July 12 in the array, otherwise the point is removed (not zeroed). And you do this for every date, i.e. you iterate through X axis. You can try to check for n values not only for points above/below the current one, but also for adjacent points in time. Did I make myself clear?

 
Vinin:
lna01:
the picture can be disappointing - 1/x is a highly non-linear conversion. It's still interesting to look at of course.

So it seems to be enough of a normal wizard to do an offset-period.

But then there won't be any on the current bar? But I'm wrong about 1/x - you can't get into the looking glass like that
 
lna01:
Vinin:
lna01:
the picture can be disappointing - 1/x is a highly non-linear conversion. It's still interesting to look at, of course.

So it seems to be enough of a normal wizard to do the offset-period.

But then there won't be any on the current bar? But I'm wrong about the 1/x - you can't get into the looking glass like that.

In order to get it (windowpane) on the zero bar, we need to know the prices from the future. That's why the limitation is the waveform period, where it is already formed. You can take a partially formed one, but then it will always be equal to the current price on the current bpres.
 
Vinin:

For it to be on the zero bar we need to know prices from the future. That is why the limit is the period of the wave, where it has already been formed. You can take a partially formed one, but then it will always be equal to the current price on the current bpra.
That is why I do not want to consider a shifted by period contramote - it cannot tell us anything about the space :). Although technically it is a countermote. Maybe eugenk will come up with something with negative time.
 
grasn писал (а): It seems to be a strong tea and it's very addictive. Tell me, is it green or black?
I'll tell you a secret: it's a cool Chinese tea. The production technology comes from ancient times, from Tibetan monks, and is based on the rotting (= fermentation) of freshly picked tea in a closed room. The longer it rots, the steeper the product and its healing properties. Teas with a fermentation period of more than 10 years are considered elite. The cost of such, it seems, is from 1000 roubles per 100 grams. I've been drinking tea for 3 to 5 years. One of the best teas is pressed tea from "Mlesna". Nadin teas are also good. It's better not to buy cheap teas in ordinary shops.

There's practically no caffeine in it, so you can drink it at night as much as you want. Without sugar, of course. It's kind of black tea.
 
Mathemat:
Grasn wrote (a): It must be a strong tea and it's addictive. :o) Tell me, is it green or black?
Let me tell you a secret: it's a cool Chinese tea. The production technology comes from ancient times, from Tibetan monks, and is based on the decomposition (= fermentation) of freshly picked tea in a closed room. The longer it rots, the steeper the product and its healing properties. Teas with a fermentation period of more than 10 years are considered elite. The cost of such, it seems, is from 1000 roubles per 100 grams. I've been drinking tea for 3 to 5 years. One of the best teas is pressed tea from "Mlesna". Nadin teas are also good. It's better not to buy cheap teas in ordinary shops.

There's practically no caffeine in it, so you can drink it at night as much as you want. Without sugar, of course. It's kind of black tea.

Thanks for the tip. Hmmm, I think I used to drink that kind of tea, but it seems to have a real problem with the taste, I remember not liking it at all. But there's some conceptual differences between Tibetans and yogis and the rest of the world about decay. They seem to think that any products based on putrefaction should not be eaten and drunk. Anyway, the East is a very subtle matter. But I will try again the healing properties of this tea.

 
Well, to predict the future, you have to give up time reversals :)
I) Good question - "why do we need worms".
1 - I'm not saying unequivocally on these worms a bounce or resistance - I'm looking at the fact - on these worms it's highly likely that price will linger.
Not always - I agree.
2 - Why draw channels - (whether they are drawn correctly or not) - here the price itself draws them. And as I have already said, the same worms resist price movement on higher TFs as well.
(I left the picture at home, but my computer is weak at work).
That is, I imagine the use of the picture is not to assess possible levels and possible limits to price movement in the future. Or maybe it's equilibrium states (which is what everything was looking for in stochastic resonance).
The only problem is the computer load. But if I could make an indicator, which would draw worms like on the drawings on page 7 and on different TFs - I think it would be useful.

II) I looked at the "cut" of the compressed worms. It looks like a smoothed price series curve (what else do you want - there is no other information in the dummies)
 
Post vaa20003 included this idea (isn't that what Candid suggested?): probably some detailed information can be provided by analysing the function of the Mach position as a function of the period on the bar. It's not density or distribution, it's a function. Horizontally is the period of the waveform, vertically is its value on a fixed bar.

Approximately: If this function is monotonous, this is a trend with a high probability. If it has a lot of local extrema, it is a flat. Its other parameters can also be analyzed correlating them to parameters of similar functions on earlier bars. Otherwise I do not see much justification in piling up all this excessive magnificence on the chart and its extremely superficial analysis.
12345678910111213141516
New comment