To follow up - page 22
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
Can the resulting contexts of an algorithm be different? In other words, is the context identical to the algorithm that extracts it? It seems that in your interpretation, yes, it is. In mine it is not. But we can try to classify them by identifying a limited number of context types. In this paradigm it is easy to find counterarguments to your arguments.
In your interpretation, this choice is a matter of taste. In mine, it should be dictated by the type of context.
If you have a different algorithm for each context, then the task is set incorrectly. Above I gave my definition of context. It follows that there is only one algorithm that forms the entire phase space, which in turn is determined by market parametrization. This algorithm is not a consequence of my or your idea of context, but an indication (if you like - direct) of all points of history, in which trading decisions should be made. That is, it is the principle of the largest (from the TS creator's point of view) profit that rules here. These points are displayed in the phase space. If they are grouped, then we obtain context areas - types of context. It is not known in advance how many of them there will be. So your impression, as far as I understand it, is a bit off.
The state parameter is what determines the type of context (or rather it should be determined by the set of state parameter values). And zig-zag partitioning results in both breakdown and breakdown contexts.
See, you came up with your own contexts based on your notions of trading strategy. And I came up with other long and short. So what? Where's the robot's guide to action here? In Karaganda. And if you pointed out the points of change of its direction on ZZ, and then the values of the set of your parameters were displayed in the phase space, you would immediately see from the distribution of these points whether the pair works (points of decision-making - parameterization of market conditions) or not. That is, whether it highlights the sought context areas or whether these points are evenly dispersed in the phase space.
Here's the conclusion about fitting the WP parameter is completely unclear, where does the need for fitting in my logic come from? After identifying the type of context, I simply choose the appropriate strategy.
From this paragraph right here.
So you're going to build up a historical base of contexts based on one of the state parameters? But in this case you sacrifice it, it will have to be excluded from the analysis. And the choice will be more or less oriented not to the final goal (maximum benefit with minimum risk), but to the optimization of this defining context parameter.
Since I had neither "to build up a historical base of contexts" nor "optimize this setting context parameter" in my mind, I assumed it was inspired by your own ideas of what is supposed to be done. If wrong, that's fine. I hope I was clear enough above so that such interpretations don't arise.
ZZ vertices can't be signals
A signal in my definition is a moment in time to make a decision. That is why the moment of fixation of a top in a GZ can be a signal, but the top itself cannot be, because it is always determined with a lag.
Clarification: not those groups. The market is not driven by elliots or breakers.
Так может определится, что стоит.
Ведь без определения системы, все рассуждения о фазовых пространствах - "решетовщина"...
;)
!!!))) The honing of terminology continues? Buggahaha!!!)) Poor Reshetov... // let's not get personal, ok?
===
From what I've reviewed/read/understood (listed in descending order of scope)))). Later, when I figure it out. Right now what's on the surface:
A breakdown by context would generally be by actual trading patterns. But here's the point. I have arrived at a breakdown by micro-context, which should be a) a common basis, a brick for identifying/analysing/forecasting everything else, and hence b) contain sufficient market information to do so, and c) be based on a relatively well-established (quasi-stationary) process.
Volatility is most suitable for these purposes - a) the brick is generic enough (nothing is lost) and shallow enough not to fly off into eternity; b) there is enough information, incl. the rate of price change... c) the process is well-established for liquid instruments and quite suitable for binding.
I use my calibrator (variant of MasterSlave in Kodobase) to mark the peak values of volatility for micro-samples (approx. 5 conventional t-f bars). They are the separators of new frames/bars. I will try to show it all graphically, in the form of candlesticks to make it clearer. (I haven't done it for myself, as well as merging the series into a file, but it will be necessary for the discussion). Of course, you can find illogicality in the fact that volatility is analyzed on a fixed sample, but a) something must be taken as a resume, and b) maybe we should make the sample floating. The threshold is floating anyway. In short, initial inconsistencies can be found. (In anything.) But you have to, I repeat, take something as a basis...
===
Zigzag, on the other hand... Well... The zigzag itself is just a markup. It could be based on anything. You can mark it up with what I'm using. The classical threshold in head-on doesn't suit well at all - the threshold is rigid and it must be tied to the same volatility, but at a higher level, of course. By the way, here is the threshold with hard threshold and floating ATR. The threshold itself (without the multiplier) is in the subwindow for information. In principle, you can use this variant as well. It adequately enough reflects fluctuations in the current context. But normalised volatility is still better.)))
Clarification: not those groups. The market is not moved by elliots or breakers.
Yes, they do. It really depends on which market we are talking about and how much they move and in what time horizon. Obviously, it's silly to say that the breakers rule over the long term trends. But micro-trends and micro-contexts may well be created even on FX. It's useless to argue about who rules where - you have to try to formalise and test.
Не ожидал подмены понятия. 33 - дает не сигнал, а место сидения наблюдателя. Иногда ветхий заборчик. Иногда цитадель.
И если равнодушный наблюдатель замечает ветхость забора, он тот час переходит на следующее, более выгодное - по его оценкам, место.
I have not substituted anything. You will only be able to choose your observer's seat when it becomes stable, i.e. when it is registered. The price will have already run off by just the amount you would like to have in your pocket.
I am not against such a scheme at all, but it cannot work without an effective prediction of the next seating position.
Speaking of crowds.
Here's a picture to think about:
It would be good to sign what is on abscises and ordinates. Otherwise it turns out to be a Malevich Square.
The abscissa axis shows the numbers of the intervals in which the values fell. On the axis of ordinates - values of frequencies of hits divided on quantity of elements of samples.
Я ничего не подменял. Ваше место сидения наблюдателя Вы сможете выбрать только тогда, когда оно станет устойчивым, т.е. в момент его регистрации. Цена за это время уже убежит как раз на ту величину, которую Вы хотели бы уже иметь в кармане.
Я совсем не против такой схемы действий, но она не сможет работать без эффективного прогноза следующей точки сидения.
Agreed. But I would then determine the size of that unreceived price in my pocket. Could it be relatively small? And it's worth neglecting.
That's also what a merchant with a big horizon thinks. Isn't that what we're talking about?
------
And if we have taught the observer not to turn his head - the next point in the distance is where he is looking.
Another context can be introduced - from the point of view of the Shy Dog. This context, if set up correctly, will give an estimate of the likely turn.
Then our "non-buy-in" context is formed by the inconsistency of these "observers".
Как я понимаю, чем выше количество степеней свободы (~групп агентов с независимым поведением) - тем стабильнее ведёт себя цена. Соответственно, цена возвращается к некоторому среднему и мы можем использовать откаты для получения прибыли. // вторая гистограмма из моего предыдущего поста
Когда количество степеней свободы невелико, цена имеет тенденцию к продолжению движения в какую-либо сторону, однако это движение скорее всего будет неустойчивым, т.к. изменение поведения одной группы может оказать очень сильное влияние на цену и попросту развернуть её.
In terms of sustainability, yes, I agree. But predictability is also important for us, and here it is exactly the opposite, the more degrees of freedom, the more the behavior of the system resembles a random one. Hence the assumption that "our" optimum is closer to the case of small m.
There are many options
.It all depends on whose sanction we're tracking.
...
It is a curious approach. The purely speculative interpretation of what takes place in the market is a bit dubious. However, like any formalism it does not have to literally correspond to the reality, the main thing is the result. Also, I have not yet familiarised myself with the discussion on the link spider.
.
.
.I don't like ZZ, because the moments of signal registration on it are quite far from "ideal inputs
".That's not entirely true, already the junior-senior ZZ scheme makes possible (among others) inputs that are very close to ideal. However, I haven't said anywhere in this thread yet that I've decided on a partitioning scheme.
Your dream of Ultimate Partitioning I fully share :) . And have you not tried to actually build this very conjunction of N mashes? It would be extremely curious to see a picture.
If you have your own algorithm for each context, then your task is not correct.
...
If you decided that you may identify WP reversal only at "top fixing" moment, thus you have crossed out the whole idea of context
..
.
.Are you going to dispute that ?
It's pretty clear in general terms. I won't answer in detail any more - there's a risk to fill the whole branch with our discussion :). I have only left two obvious inaccuracies in my opinion to answer:
1. Exactly the opposite - one algorithm for selecting different contexts. Only I'm talking about extraction in history. And in real time, I think it's only possible to recognize the current context.
2. No. The context should not be reduced to a single transaction, there can be several within the same context. Well and above in this post I've already reminded about the senior-senior ZZ scheme.
3. no I won't, you can't say "yes" or "no" here, you have to agree on concepts long and hard :)
волатильность как рассчитывали? без учета времени/периода в барах?
And in relation to what? Ward average, K-period waving, or M-period?
averaged on Volum, or on n bars...?
---- a million questions.
Where is the answer?
How do you measure it - cicatly to find out.