Machine learning in trading: theory, models, practice and algo-trading - page 2294
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
Alexei, how can you relate the theory of martingale to martingale in forex, to look from a scientific perspective on MO + martingale? :)
In bourgeois it's the same word - they write that science emerged as an attempt to explain the futility of this way of betting)
For now I can only say that equity should not be a martingale, but a submartingale (or supermartingale - I confuse them) - expectation should grow) I have no more sensible ideas yet).
In bourgeois it is the same word - they write that science appeared as an attempt to explain the uselessness of such a way of betting).
For now I can only say that equity should not be a martingale, but a submartingale (or supermartingale - I confuse them) - expectation should grow) No more sensible ideas yet)
The expectation (or rather the probability of winning) grows with each doubling, because you get a 2nd, 3rd, etc. chance. But that's as long as you have enough money to double.
But once in 100-100000 trades you don't have enough... With black swans for example. They come to martins more often)))
In bourgeois it is the same word - they write that science appeared as an attempt to explain the uselessness of such a way of betting).
For now I can only say that equity should not be a martingale, but a submartingale (or supermartingale - I confuse them) - expectation should grow) No more sensible ideas yet)
In this case we'll do as usual, in the petty bourgeois way.)
The expectation (or rather probability of winning) grows with each doubling, because you get a 2nd, 3rd, etc. chance. But that's as long as you have enough money for doubling.
But once in 100-100000 trades you don't have enough... With black swans for example. To martins, they come more often))))
At SB the expectation for any system is always zero (without taking into account the spread). With respect to real prices it is difficult to say something for expectation is abstract and exists only within the mathematical model and no exact price model has been developed yet. The only thing to say is that if prices give an opportunity to earn, then in case of an error in choosing a deal direction this opportunity becomes an opportunity to get a margin call (especially when using a Martin).
I don't understand what you're looking for in this way. You need to find a statistically significant deviation from the mean over specific periods, or autocorrelation
it might make sense to go through low-frequency filters rather than just MAHs
In econometrics, a deviation indicates that there is an unaccounted for component in the residuals. But for example in pair trading I've never seen such small values.
In econometrics, the deviation indicates that there is an unaccounted for component in the residuals. But in pair trading, for example, I have never seen such small values.
This is not the deviation.
There is one more obvious possibility for potentially useful application of the forces of tsosniks. We are talking about randomness tests in general and NIST tests, for example, in particular. It has everything that they love so much - cycles, Fourier, patterns, etc. It is true that only binary sequences are studied there, but it is possible to limit oneself to study ranco graphs first.
These tests are not full-fledged, I saw an article that the randomness test was able to pass music. It is interesting to use gan network for this, but probably will not be able to cope.
Then we'll do the usual bourgeois thing)
Count probability, expectation and dependence on the coefficient of martin. By the way the lot risk 0.5 is almost a martin and holds for 2 years, but the result is comparable to the 0.2 risk).
Thanks. By the way thought about how to remake the metric, came to mind so - the example if the classes are 0 (30%) and 1 (70%), then for the right value on the reverse pass to give not 0 and 1, but 0 and 0.7 ( or 0.85?) as an option.
But I can not immediately think about whether this is the right way to do it, or such numbers, or make some magic with the activation function, etc., and immediately did not find such examples.
There are some ready-made metrics, look them up. Or give 2.33 for the 0th grade and 1 for the 1st.
To count the probability, expectation, and dependence on the coefficient of martin. By the way, on the owl from the market the risk of a lot 0.5, it is almost a martin and 2 years holds, but the result is comparable to the risk of 0.2).
For some reason everyone is used to calling martin averaging on positions, mostly not meaningful or over-optimized
it is possible to leave the lot fixed, but still use a grid. Trade sets (entry and exit points) will change, so will their representation in the feature space for MO. This is the interesting point.
I don't know if the stability effect will appear on the new data. I do not have such a mathematical formula. Check it empirically.