Using Neural Networks in Trading. - page 8
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
and the arguments? why do you think so?
Because it is measured in ticks and differs from site to site.
How about this argument in favour of continuity: in fact, we are able to make a trade even when there are no new ticks... i.e. even if there are no ticks, the price is there...
and, by the way, this is why many traders insist on eliminating time gaps between bars... (i.e. bar gaps)
How about this argument in favour of continuity: in fact, we are able to make a trade even when there are no new ticks... i.e. even if there are no ticks, there is a price...
and, incidentally, this is why many traders insist on eliminating time gaps between bars... (i.e. bar gaps).
In fact, the ticks are there, they just reflect the same price.
Yaroslav, you see, the whole argument is actually about whether we should assume that the price function is continuous on Ox and therefore has a derivative at any point... or if you can't do that...
Neutron, apparently, this is what he is getting at - that one cannot...
If there are always ticks, it must be assumed that the derivative can be calculated...
That's the essence of our problem:
Плотность распределения случайной величины ŋ = g (ξ) существует далеко не при любых функциях g. Так, если функция g кусочно-постоянна, то имеет дискретное распределение, и плотность её распределения не существует.
ZS. No, we seem to be confused (or I am the only one :))... colleagues, help me out... When we talk about probability density, we mean the derivative of a distribution function... but price as a function of time is not a distribution function... I seem to have got off on the wrong foot... :)
A tick is the unit of change of information. Therefore, ticks in MetaTrader4 are absolutely correct.
On other platforms, where there is market depth, each tick carries a new change in market depth for the corresponding trading instrument. There, indeed, the price can stand and the ticks come as volume changes are taken into account in addition to the price.
This is the crux of our problem:
The distribution density of a random variable ŋ = g (ξ) does not exist for all g functions. Thus, if a function g is piecewise constant, it has a discrete distribution, and its distribution density does not exist.And again to the point!
I haven't tested it in combat yet, but it looks like we'll have to use interpolation between discrete samples of the distribution function F(n). This is where the problem with constructing a rectangular distribution on real data arises when trying to act on the input data with this operator.
What is the input, Neutron? Returns bars or something else? Bulashev seems to have something about their exponential distribution (for UES?).
I did not accurately formulate the problem.
So, if I, having plotted the first difference series (Returns bars), can find a more or less exact approximation with an analytical function, then there is no problem - I integrate and get a uniform distribution. But it is not always possible and almost always not convenient. I want to solve the problem without finding a smooth curve. There seems to be no problem, we construct a histogram for probability density function, integrate it numerically and obtain a discretely specified distribution function (DP)... This is where the problems begin, because the PDF is not continuous.
I can't figure out exactly what Neutron's problem is either... I don't have a handle on it yet...
In principle, as far as I understand it, the price can be considered both a discrete and a continuous process... the question is, which is the right one?
to be honest, i haven't got to shiryaev yet... left it for later as "the best part"...
Victor (renegate), could you please briefly state what conclusions Shiryaev comes to (I mean what he's getting at - should we consider price as a discrete or continuous value)?
Shiryaev in these books has laid out facts first and then models (they are option strategies) and he uses continuous time there already. I can't understand the content so far (although I don't think there is anything complicated there, just all in formulas), and I am not interested in option strategies. I got more interested in Renko and Kagi after Shiryaev's lecture "Mathematical formalization of Japanese candlesticks". However, for me personally, it's not so easy to switch to these constructions, because MT4 supports time representation of price, not quantization by levels.
Shiryaev, in these books, has set out the facts first and then the models (which are option strategies) and he uses continuous time. I can't understand the contents yet (although I don't think there's anything complicated there, just all in formulas), and I'm not interested in option strategies. I got more interested in Renko and Kagi after Shiryaev's lecture "Mathematical formalization of Japanese candlesticks". However, for me personally it is not so easy to switch to these constructions because MT4 supports time representation of price instead of quantization by levels.
By the way StatBars, this is just to the point about the vertical (not temporal) way of breaking down the price series.