Obtaining a stationary BP from a price BP - page 25
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
You will be second, after Reshetov. ;о)
Oh, you... Scrounger!
FOXXXi wrote(a) >>
WMT share,traded in all kitchens.AKF minus 10-20% with stat. signifikant,either on days,or on hours,nemnu.Go and try to ruin though one DC.
In the task of "profitability" of the instrument besides the internal, obvious and not very obvious relations between results in RPP, there is such parameter as commission of brokerage company. For sure, the WMT commission is exorbitant and completely "eats up" all your profit of 10-20%.
>> About 99,9 % I did not understand. Is it the correlation coefficient between neighbouring readings in the price series? If so, this figure is for fools who confuse meaningful relationships between bars with the constant component present in each bar.
Peter, finish your humpbacked excuses and hurry up and join us in this feast of life - get to work!)
))) I'm trying!
(I'm trying!) OK. In all seriousness, I've made my point on this thread. Last time I didn't mean it seriously - sorry, it just seemed really weird.
Then maybe go from goal setting? After all, everyone's goal is the same - to buy cheaper/sell more expensive. Or in reverse order. And more than once and with a stable result.
I understand your desire to provide you with a theoretical basis - of course in order to do something you must understand what you are actually doing - otherwise it's pure alchemy and shamanism. I agree.
Sergei! Can I have a decile of specifics? For example, no one disputes that you can make money on the trend. But you (let's keep it on a first-come-first-served basis, OK? I understand that you don't mind))) initially consider an absolutely impractical model of this useful phenomenon. Hence (and not only) my passage about a honeymoon with a spherical horse in a vacuum hotel.
Let me take away some K of your neural transactions in your light mainframes and give a practical example of working with a local model.
For an example, I give this definition of a trend: a trend is a unidirectional movement of extrema defined within the current volatility. // for up-trends - troughs, for down-trends - peaks.
Other definitions based on other parameters can be given but I will confine myself to this one in my demonstration.
So, the keyword and the clue to stationarity is the overall volatility. This allows us to recognise, with a high degree of probability, areas which are suitable for trend operations and which satisfy the other conditions of the definition.
Let us proceed to capitalise on this model. But first, let us discuss the method of identifying extrema. It can be any adequate, but, not to get lost, I have chosen B. Bolli (not to be confused with Johnny D.). (not to be confused with Johnny D.!)), and I define the peak as the maximal value of price between the crosses of the price from the lower gr. bar. The trough is mirrored. I don't think you, fellow scientists, need to explain the meaning of this method.
From the definition of a trend, it is clear that we consider a corrective model of movement and therefore the tactics of opening/closing/turning positions is appropriate (for an up-trend):
Opening/re-opening - at the outset of correction to the wave on a detrended by definition trend, which coincides with the definition of the peak by the method described above - i.e., at the price's touch of the lower border of the B-band.
closing/fixing in a long context - again, there are many ways, but let it be by an entry beyond the level of 1.618 from the difference of extrema (the width of the horizon channel).
closing on the cancellation of the long context - also on the peak detection, but when the sequence of rising troughs is violated. May (but not necessarily!) coincide with a position reversal.
(Shit - it looks like an article - "Tactics of trading along the horizontal channels")))
Anyway, I drew up a quick indicator for demonstration purposes. Here is a picture of a trend segment. For convenience - price extremums are displayed as horizontal channels and Bollinger Bands - as oscillator in order not to clutter up the price chart. Blue triangles are for closing a position; red and green are for opening.
And here is a picture of an exit by loss of context. The first rectangle marks an exit after a long trend by breaking the sequence of rising troughs. The second rectangle is sort of a false entry with an exit for the same reason - no confirmation of the trend with a higher trough.
The moment of reversal:
Why am I being so detailed? 1) So as not to be unsubstantiated. 2) Just to show that it makes sense to model a PARTICULAR model, not a market at all. All IMHO, of course.
=========
I realise that my post is a complete offtopic, but I had to explain for the goat... ...ugh! A horse? ))))
In short, I don't think your approach is practical. It seems to me that trying to predict the market using its model is not interesting from a practical point of view. Working on a profitable plots model is another matter. Or, as I call it, contextual trading.
... I realise that my post is a complete offtopic, but I had to explain it for a goat... >> a horse... >> to explain myself? ))))
In short, I don't think your approach is practical. It seems to me that trying to predict the market using its model is uninteresting from a practical point of view. Another thing is to work on a profitable plots model. Or as I call it, trading by context.
What's the point? "Profitable" plots are a shitty fit. There is no need to look for them. You just need to select a small area and run optimization in the tester.
But brokers will not pay for "profit" on these very plots, i.e. for yesterday. They will only pay or charge losses for extrapolation, not for approximation.
I.e. when the accruals are for tinkering, then we'll listen carefully to your thoughts. In the meantime, suck it up granny.
What's the use? "Profitable" plots are a shitty fit. There is no need to look for them. All you have to do is to choose a small area and run optimization in the tester.
No shit. It's not a fit. It is an exploitation of a described trend model relying on a quasi-stationary process - volatility. All the parameters are logically defined.
But brokers will not pay for "profit" on these very plots, i.e. for yesterday. They will pay or charge losses only for extrapolation, not for approximation.
??? I mean, how is it yesterday? You misunderstand - it is not the price movement that is predicted but the context. In the above example - the trend. When there is no trend, there is no trade. There are other models to work with. Do you understand what I mean? I'm not talking about the fit or not, I'm talking about the approach to trading.
I.e. when the accruals are for the fit, then we will listen carefully to your thoughts. In the meantime, bummer grandma.
Well, well. And a dime a dozen for you. Granny, i.e. me, is screwed long and steady by these approaches.
===
I wonder if you guys have ever seen more than three rubles from the trade. I'm starting to have doubts.
!!! I don't want to offend anyone - that's just the impression I get.
Oh, you... >> you punk!
Come on, guys, stop burying the subject. I assure you, everybody, this subject is not as simple as it seems.
1) The problem of "profitability" of the instrument, in addition to the internal, obvious and not so obvious, relations between the readings in RPM, includes such parameter as the commission of the brokerage company. Sure, the WMT commission is exorbitant and completely "eats up" all your profit of 10-20%.
2) I do not understand the 99.9%. Is it a correlation coefficient between adjacent readings in the price series? If yes, then this figure is for fools who confuse the meaningful relationship between bars with the constant component present in each bar.
1) Judging by your knowledge of commissions, you have nothing else but currency pairs.
2)It's a correlation between random walks, you could have guessed that.
??? I mean, how is it yesterday? You don't understand - it's not the price movement that is predicted, it's the context. In the above example - the trend. When there is no trend, there is no trade.
No, we all understand perfectly well.
For this purpose it is necessary and sufficient to predict not the skin of the unkilled bear, i.e. the profit on the trend, but its potential possibility of this very trend on the next segment. I.e. will there be a trend or a sideways trend?
Otherwise it turns out that one can predict anything, for example, winning a bet on red in the casino. Like? All you have to do is smoke bamboo when black comes up.
It's only a matter of time.
Svinozavr wrote :>>
Grandma, i.e. me, is screwed long and steady by these approaches.
===
I wonder if you guys have ever seen more than three rubles from a trade. I'm starting to have doubts.
!!! I don't want to offend anyone - it's just the impression I get.
That's where you should have started, not with 500 useless tips on taking profits in "profitable" areas.
Those who have seen more than three roubles on a real balance sheet know how to trade on those very plots. Give me one of those plots, and I can trade more than three rubles on it like two fingers on the pavement.
What I meant to say is not necessarily previous prices or their increments are the cause of subsequent movements (although it happens, and appropriate TA methods use it, e.g. momentum trading). Sometimes price movements may be used to identify some market phase or its finish, which is exploited by pattern trading. It means that the prices or their increments only serve as a trace of continuation or termination of some objective market processes.
In both cases we have a dependence on the price series.
I still want to understand what it is in your case, is it a disguised pattern trade, or is it timing, and it exploits the property of the SB.
You do not understand - it is not the price movement that is predicted, but the context. In the example given - the trend.
I want to support it. However, with reservations.
Predicting the price movement has so many problems that it's easier to kill yourself. :-)
But predicting the context is another matter. There are only two contexts - a trend and a flat. It's enough to choose a "good" definition for each of these contexts and you can develop your own TS. There is only one problem here. These very "good" definitions. That's why the most frequently asked questions on this forum - what is a trend? What is a flat? In this example, the definition of the trend is quite correct, i.e. it has all the details for the software implementation. Therefore, it is quite easy to determine whether it works - program entries/exits using this definition and run it on a large block of history. The ratio of successful/unsuccessful virtual trades will show its effectiveness.
What is the "but"? In its private nature. Such definitions can be thought of many and very many. And each needs to be programmed, and then investigated for effectiveness. And even if it is positive, it does not mean that it will be sufficiently positive. And one more point. This particular definition relies on phenomenology, i.e. the "seen" effect of a TA tool. The survival of such a phenomenon in the future is highly doubtful. As we know, the biggest problem of working TA - the short-term nature of its profitability - is connected, imho, with its phenomenological nature.
In connection with what has been said, I will allow myself to disagree with the following:
But it seems to me that you are looking in the wrong place and for the wrong thing. You are looking for a market model IN GENERAL, while I am deeply and justifiably convinced by successful trading that you should be looking for models of potentially profitable PIECES. Someone already pointed out that you are trying to measure the average hospital temperature. I would add from myself that sometimes in a rectal way.
No way. This is not a fitting. It is an exploitation of a described trend model relying on a quasi-stationary process - volatility. All parametres are logically defined.
No, it's not a fitting, it's a "counter-trend strategy". We set the bollinger with a 20-period, for example, when the price deviates from the extreme Bollinger Bands we go in, and then it turned out that we were at the point of no return and the price was so high that it didn't look too bad (the loss exceeded the previous profit).