Machine learning in trading: theory, models, practice and algo-trading - page 2826
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
Depends on the instruments and the TS, I guess. For a deposit in the Soviet Sberbank - deterministic, for example)
Traditionally, one uses a trended SB model (without reinvestment) and a geometric SB model (with reinvestment). These models are implicitly assumed when using Sharpe for valuation.
idea : what if we use the forecast of let's say arima on the balance of trade curve as a measure of the quality of the model's learning.... Like if the forecast is up and the confidence interval is narrow then there is a chance that the capitol curve will go up, this is something similar to montecarlo simulation
idea : what if we use Arima's forecast of the trade balance curve as a measure of the quality of the model's learning... Like if the forecast is up and the confidence interval is narrow, there is a chance that the capitals curve will go up, this is something like a montecarlo simulation.
As always, there are two quantities to be optimised - trend velocity and volatility (probably related to the width of the confidence interval). It is necessary either to build one of them (Sharpe, for example), or to build an optimal curve (as in Markowitz portfolio theory, for example) and then choose a point on it.
As always, there are two quantities to be optimised - trend velocity and volatility (probably related to the width of the confidence interval). It is necessary either to build one of them (Sharpe, for example), or to build an optimal curve (as in Markowitz portfolio theory, for example) and then choose a point on it.
I don't get it)
Well, optimising for the type of capital curve is common. Usually one tries to increase the trend rate and reduce the range of fluctuations around the trend line - so there are two criteria for optimisation. This two-criterion optimisation problem can either be reduced to conventional optimisation (making one of the two criteria, e.g. Sharpe), or one can search for a Pareto-optimal line on the plane and then choose a point on it from some considerations (e.g. Markowitz portfolio theory).
There are a lot of variations on this topic and a large number of articles.
Well, optimising for the type of capital curve is commonplace. ....
Ah, I see, but that's not what I mean, I'm just saying that it's interesting to optimise not the capital curve itself, but the forecast that the capital curve will grow in the future.
like optimising the capital curve so that the forecast is as good as possible(like a red channel).
The curve itself can be falling all the time...
I tried it with autoarima, pretty good results, but it takes a long time to train. 7-8 out of 10 guesses capital growth
Ahh, I see, but that's not what I mean, I'm just saying that it's not the capital curve itself that's interesting to optimise, but the forecast that the capital curve will grow in the future.
like optimising the capital curve so that the forecast is as good as possible(like a red channel).
The curve itself can be falling all the time...
I tried it with autoarima, pretty good results, but it takes a long time to train. 7-8 out of 10 guesses the capital growth
What I wrote before is quite consistent with this, because there is always a final capital curve that we optimise. If we imagine that we are not using ARIMA, but some black box like a neural network, the only difference is the addition of features corresponding to the capital curve of the original system.
I am prejudiced against the idea of building a TS as a metasystem on the basis of another TS. If this technique works, it is necessary to understand the reason for it and use it directly and explicitly.
The most difficult thing I have done in this direction is to stop trading at drawdown and resume it at growth.
The most difficult thing to do in this direction is to stop trading at drawdown and resume it at growth.
And how will the depo start growing if trading is stopped? )))
And how will the depo start to grow if trading has stopped? )))
Real trading stops, but virtual trading continues) It is clear that real equity would be different from virtual equity, but this is neglected.