Machine learning in trading: theory, models, practice and algo-trading - page 1589
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
No, just the dependence, and a constant dependence.)
You can try to detect it with the help of the MO
This is the right question, you don't need to do anything, wait until there is a barely statistically significant sample, any action in such a situation will be for luck, if there is no insider.
Stationarity is the INDEPENDENCE of MO, variance and distribution function from time.
And what methods of MO determine "market change"?
A sample appeared - built a model - built a TS - small profit in each trade - "market change" - large loss, overlapping all profits. What next?
P.S. Constant dependence - is it something from philology?
Colleagues, hello,
I'm sorry for the stupid question, but is OnBookEvent event working in MT5 tester? I'm trying to test it, but for some reason I don't get into a loop and feel that it is ignored. But in theory the quotes change in the market review. HMM...
No
In our case, we can meaningfully work only with non-stationarity, which in one way or another is reduced to stationarity. Piecewise stationarity, autoregressive models, hmm, etc.
The main reason is that only one realization of the process is always known. For example, if we take speech recognition, there any word we can say as many times as we want. The quotes for a certain instrument over a certain period of time are a single variant. By the way, this is probably the reason of fuzzy distinction of a random process from its realizations.
Knowing a mixture of distributions, it is possible to create several realizations through generative models
I don't know whether it will help, but it is possible to check TC additionally.
It is amusing to observe how people mock the good old statistical (non)stationarity, implying by it anything at all, but not the relative persistence in time of the distribution.
If strictness is needed, we can assume that we are talking about the absence of stationarity in the broad sense of the logarithms of returns, for example.
Stationarity is the INDEPENDENCE of the MO, variance, and distribution function from time.
No, open a textbook and refresh your memory, stationarity is the STABILITY of the distribution over time, that is, when the MO, variance, and other moments remain constant, do not change.
And what methods of MO determine "market change"?
Different ones, neural nets, wood, etc.
There is a sample - built a model - built a TS - small profit in each transaction - "market change" - large loss, overlapping all the profits. What next?
And if you determine that the market will change, you disable the trade and do not get a loss.
P.S. Constant dependence is something out of philology?
No, it's the same dependence as any other y = const
And if you determine that the market will change, you will disable the trade and you won't make a loss.
but you may already be in a trade, and not just in a trade, but in a drawdown, which, if the market were the same, would be replaced by a profit, but since that is no longer the case, there may be zero chance of getting out of a drawdown
No, open a textbook and refresh your memory, stationarity is STABILITY OF DISTRIBUTION IN TIME, that is, when the MO, variance and other moments remain constant, do not change.
Different, neural nets, woods, etc.
And if you determine that the market will change, you turn off the trade and don't get a loss.
No, it's the same dependence as any other y = const
A constant is a constant, that is, a constant value. And a dependence is a variable. Constant dependence is gibberish.
A change in the market is a change in the probability characteristics, most often the variance. Any trading on a stationary market is when a trade is opened on the border of a channel and closed when a certain level or the opposite border is reached. Market change is when a trader opens a transaction on the border of a channel and it goes in the red, expanding the channel to abnormal sizes.
The loss is more than the profit.
***
More than 1,500 pages discuss the mathematical learning involved in trying to turn a trough into a washing machine.
If strictness is needed, it is possible to consider that we are talking about absence of stationarity in the broad sense for logarithms of returns, for example.
I read this article about 5 years ago, it is interesting, but there is not much additional information, the author makes something with OHLC to get a more "convenient" volatility metric, it is not new in principle, in the classic Dacorogna "An introduction to high-frequency finance" back in the last century it is recommended to take average absolute values of returns, not RMS values, as a measure of volatility. The predictability of the volatility is also a well-known fact. It depends on two factors, seasonality and inertia, which accounts for 95% of its predictability. But even if we align (log)returns according to volatility, it won't give anything, we need a sign for trading, and it doesn't affect the distribution in any way.
For example, if you take a Gaussian noise, you obviously cannot predict the following ones using previous samples, regardless of stationarity, but if you sort this series, for example, that will not change the distribution, but will make it completely predictable, then you can play with dynamic volatility within very wide limits and make it nonstationary, but still easily predictable.
A constant is a constant, i.e., a constant value. Dependence is a variable value. Constant dependence is gibberish.
Don't tell mathematicians that))