Machine learning in trading: theory, models, practice and algo-trading - page 3656
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
Does anyone know what the r-bloggers.com site is doing? I've had it down for days.
It only opens with a VPN.
Maybe google blocked the blogger engine because of sanctions.
No, blogger.com opens. Maybe the site is on the wrong engine, although it looks like it :)
Now it opens without VPN at all... I refuse to understand how it works )
Can I advertise chat rooms here?
Thought to create a chat"Geniuses of Thought", if you want to poftotopit, you can write in it on any topic. Fun for :)
Zero moderation. Well, it is clear that except for politics and other obscene things. For such a ban.
Which flatness or trendiness test can be used to classify trading instruments (currency pairs, for example)? To show meaningful differences between instruments.
Is Dickey-Fuller's test suitable?
Which flatness or trendiness test can be used to classify trading instruments (currency pairs, for example)? To show meaningful differences between instruments.
Is Dickey-Fuller's test suitable?
By its nature, this test does not compare two series with each other - it compares one series with several theoretical models.
In principle, it is probably possible to compare two series according to the degree of their similarity to a given theoretical model, but this is not particularly welcome in matstat. For example, there are two versions of the Kolmogorov-Smirnov test - one-sample and two-sample. In the first version, the sample is compared with the theoretical distribution, and in the second version, two samples are compared with each other.
In other words, you need some kind of test in which you need to put two series rather than one.
And, perhaps, the main thing is that each test only looks for some specific difference - between averages, for example. There is no test for differences "in general".
By its nature, this test does not compare two series to each other - it compares one series to several theoretical models.
In principle, it is probably possible to compare two series by the degree of their similarity to a given theoretical model, but this is not particularly welcome in matstat. For example, there are two versions of the Kolmogorov-Smirnov test - one-sample and two-sample. In the first version, the sample is compared to the theoretical distribution, and in the second version, two samples are compared to each other.
In other words, we need some kind of test in which we need to put two series rather than one.
And, perhaps, the main thing is that each test only looks for some specific difference - between averages, for example. There is no test for differences "in general".
Why is it always so complicated?)
Why is it always so complicated )
Which flatness or trendiness test can be used to classify trading instruments (currency pairs, for example)? To show meaningful differences between instruments.
Is Dickey-Fuller's test suitable?
First of all, you need to come up with a definition of flat/trend for yourself
We take the quote increments and look at the median. If the absolute value is less than the spread, then flat, if it is more, then the corresponding trend. As it should be such a flat-trend depends on TF and window size.
And no dickey-fullers and kolmogorovs are needed.
We take the quote increments and look at the median. If the absolute value is less than the spread, then flat, if it is more, then the corresponding trend. as it should be, such a flat trend depends on the TF and window size.
This works for a symmetrical distribution. In general, any non-zero expectation at zero median is possible.
This works for a symmetric distribution. In general, any non-zero expectation at zero median is possible.
It doesn't matter.
We are interested in the probability of direction on the traded TF and window. But as we know, all this is nonsense, as there is no median, and at the same time there is no average. At any moment there can be a change of trend, or an outburst of more than 100 pips, or all together, including disconnection of the communication channel with the broker.
In general, it is gloomy with lightning flashes.