Machine learning in trading: theory, models, practice and algo-trading - page 1789
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
For what you want but within reasonable limits, but please just the price and balance, nothing too much, so as not to torture my laptop)
No way - let's define the terms, because I trade on the minutes, the charts show that the trade can go well for a whole quarter, and it is already thousands of 30 bars.
I think you have to look over 3 years :)No way - let's define the terms, because I'm trading on the minutes, the charts show that the trade can go well for a whole quarter, and that's already thousands of 30 bars.
I think you need to look for 3 years :)Mammy ((((
that's how many counts?
....
send me 500K
What can be a snapshot of the BP state? What parameters to consider in the first approximation, averages, descriptive behavior, how to gradate, current, how to relate the scale from month to minute and to tick.
Obviously, it depends on the mathematical model used. The choice of model is determined by what aspects of the real phenomenon it should model (many different models can be used for the same BP). For a given model, we take sample estimates of its parameters - expectation is estimated by the mean, probability by the frequency, etc.
Obviously, it depends on the mathematical model used. The choice of the model is determined by what aspects of the real phenomenon it should model (many different models can be used for the same BP). For a given model, sample estimates of its parameters are taken - expectation is estimated by the mean, probability by the frequency, etc. etc.
Significant parameters do not depend on the model and on the researchers, they are either there or not, or they are not constant.
Aspects are the same, to determine the efficiency of various EA logics.
This approach yields limited results so far, apparently.))
Our BP initially has 3 parameters. Then there are a lot of theories. But I haven't found any work on the characteristics of BP with the scaling from month to tick. And there is an understanding that accounting is needed at all scales.
In general, taking into account the development of opportunities) I am inclined to the idea of historical traces of states / parameters, there can be a lot of them, there should be less of all BP in what do not know the number of times and they should be significant. And then the task is simplified to tracking the emergence of new states.
Meaningful parameters do not depend on the model and on the researchers, They are either there or not, or they are not constant.
Aspects are the same, to determine the efficiency of various EA logics.
This approach gives limited results so far, apparently)))
Our BP initially has 3 parameters. Then there are a lot of theories. But I haven't found any work on the characteristics of BP with the scaling from month to tick. And there is an understanding that accounting is needed at all scales.
In general, taking into account the development of opportunities) I am inclined to the idea of historical traces of states / parameters, there can be many of them, there should be less of all BP in what do not know the number of times and they should be significant. And then the task will be simplified to tracking the emergence of new states. Like this.
Disagree. Examples:
1) SB: p(n)=p(n-1)+d*e(n) - One parameter is d
2) SB with a linear trend: p(n)=p(n-1)+c+d*e(n) - Two parameters - c and d
3) Ornstein-Uhlenbeck: p(n)=a*p(n-1)+c+d*e(n) - Three parameters - a, c and d
4) Ornstein-Uhlenbeck with parameter change (decay) at time n=n1:p(n)=a1*p(n-1)+c1+d1*e(n) at 0<n<n1 and p(n)=a2*p(n-1)+c2+d2*e(n) at n>=n1 - Seven parameters - n1, a1, a2, c1, c2, d1 and d2
And so on and so forth
Disagree. Examples:
1) SB: p(n)=p(n-1)+d*e(n) - One parameter is d
2) SB with linear trend: p(n)=p(n-1)+c+d*e(n) - Two parameters - c and d
3) Ornstein-Uhlenbeck: p(n)=a*p(n-1)+c+d*e(n) - Three parameters - a, c and d
4) Ornstein-Uhlenbeck with parameter change (decay) at time n=n1:p(n)=a1*p(n-1)+c1+d1*e(n) at 0<n<n1 and p(n)=a2*p(n-1)+c2+d2*e(n) at n>=n1 - Seven parameters - n1, a1, a2, c1, c2, d1 and d2
And so on and so forth.
If from this point of view, yes. The question then is how to choose the correct model and then the significant parameters. The problem is from the reverse.
The first option can not correctly describe the second, and vice versa works. But too complex model will not always correctly find meaningful parameters for simple states.))
And SB (pseudo) doesn't really fit into my logic, the Wiener process with discrete time then. It looks more like Brownian motion).
mamamy ((((
how many counts is that?
....
Give me 500k.
What counts? :)
Leave the entry and exit date, or just the entry in the balance?
What counts? :)
Leave the date of entry and exit, or only entry in the balance?
well 500 minutes :)
Nah, let's try a simple one first, just the balance and prices
well 500k minutes :)
No, let's try a simple one first, just the balance and prices.
The date of entry and exit on the short run, too, for comparison would be good.