Machine learning in trading: theory, models, practice and algo-trading - page 2576
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
Not really. If you look at the previous graphs, you can see that the actual "rolling" turns on after ~ one-third of the sample. on real data, if there is a history of such a problem will not be.
But still calman is probably better, but I still think that it is better to split it from the stove.
We've already gone through this with Rena and the tractor, with examples of their predictions at 1 bar)))) I'm laughing
In one case it will be ahead, in the other delayed, the total is 50/50
You should at least study elementary...
Indeed, only recurrent meshes always predict the past bar, no matter how you twist them (the more advanced case of Kalman)
you should at least learn to get into the information first, and then philosophize
I'm not saying that the approaches are bad, but that they are not a panacea.
You can come up with 1001 ways to trade on cointegrated rows, the same neuronics will be no worse
i got a good idea... i don't know if it's good or not...I'm not saying that the approaches are bad, but that they are not a panacea.
I'm not saying that these approaches are bad, but they are not a panacea.
i don't think so, i don't see any profit in it.I'm also thinking about it, how do I teach them to build the right spread?
And what is a right spread in principle?
1) The spread should be stable (it's easy to calculate).
2) it should be relevant, i.e. zero spread should mean a profit on the traded pair (this is not a very "clearing" issue)
3) It should not "widen" with time
I'm also thinking about it, how to teach them to build the right spread?
And what is a right spread in principle?
1) Spread should be stable (it's easy to calculate)
2) it should be relevant, i.e. zero spread should mean a profit on the traded pair (this is not a very "clearing" issue)
3) It should not "widen" with time
The same as through linear/polynomial regression, but take returnees of arbitrary orders and partition the deals arbitrarily, fit. By overshooting, without any academic calculations of what's stationary and what's not, so as not to drown in formalities. It's going to drift apart over time anyway, unless it's fundamentally related instruments.
The same as through linear/polynomial regression, but take returnees of arbitrary orders and partition the deals arbitrarily, fit. By overshooting, without any academic calculations of what's stationary and what's not, so as not to drown in formalities. It's going to drift apart over time anyway, unless it's fundamentally related instruments.
Kalman doesn't get separated, that's the point of it, if you do it on a grid, it's not worse.
Tomorrow was a long time coming, I had no desire to do anything, even though it took 15 minutes.
So, I made a 100 point spread in the window of sliding regression (100 points)
EURUSD GBPUSD pairs (out of luck)
did not pass the cointegration check
but still made a spread, checked the trade on the deviations from the zero spread (like in the book)
I'm losing...
The most normal option is to build a Bollinger spread and trade from the border of the Bollinger to the SMA, which is also drawn from the spread, or to the opposite border of the Bollinger (there is even more profit)
All settings are absolutely standard, I have not optimized anything in general
The commission is 1p for each pair
I earned 200 points within a month.
Unfortunately the balance curve does not stand up to criticism. I have no idea what I mean by smooth growth at 45 degrees.
Otherwise it looks like random....