Machine learning in trading: theory, models, practice and algo-trading - page 673
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
Hilarious))) Once again, the last!) - Takes the previous day on the lower tf. Constructs a redrawing curve (polynomial).
If it goes according to "plan", he carries out a trade, if not, he blows it off. That's all, it has nothing to do with the MO
It has nothing to do with MO and is solved by more simple means. It is unclear just what the hell he's doing there, what he smokes and to what music he taught MLR without a teacher)))
I don't know why he smokes nets and what music he was teaching МLR without a teacher. No, he used monte carlo and annealing and the NS does not learn anything, but he also uses the market and the deals' feed, because the market is accidental for the observer and it cannot be predicted, plus the difference between MA
I don't follow theory and practice anymore. I don't know who's looking for what.) The word non-entropy gives me an idiosyncrasy.) Usually, the more terms, the less sense.)
They didn't parallelize it, but limited the scope of NS application, thus simplifying the training and the NS itself.
Not an additional signal, but the only signal - entry into the trade.
As for your system, I can't say anything about the applicability of NS in your case. It seems to me that it's a little bit different. I don't even think that these systems may have a common methodology. But I could be wrong, I know only in a general way.
The only one??? That makes a difference. I will reread your posts again.
plus the difference between MA
With the rest, okay, at least the words are familiar.) But where did you get it from?
But, in essence, yes. You can lay out the Grail algorithm in its entirety, no one will pay attention.
People have lost faith. In the Grail. And without faith no work will go well, even the impossible. I'm taking an example from Yusuf.)
The people have lost faith. In the grail. And without faith no work will go well, even the impossible. I'm taking an example from Yusuf.)
Gentlemen! I'm going to cry... Well, here's today's result:
Profit +540 pips.
When will you cheer up?
Gentlemen! I'm going to cry... Well, here are today's results:
Profit +540 pips.
When will you cheer up?
Gentlemen! I'm going to cry... Well, here are today's results:
Profit +540 pips.
When are you going to cheer up?!
Oh yes Pushkin, oh yes son of a bitch! (с)
But, in general, one deal does not solve anything. At least 10-20 in a row, without withdrawals).
I don't. I test with my tester, in a visual environment (something like R). After testing you can calculate anything - any statistics, any graphs, even 3-dimensional.
Ticks, yes, no, at least 1 minute. But they are not really needed. The minutes are enough even for scalper strategies.
I'm not talking about self-written software. I'm talking about choosing from a ready-made one. The tester and optimizer is a pretty complicated software. How much time will it take to write it? And for catching all the bugs? And who will test it? The more people who test it, the more likely it is that all the bugs will be caught. And trade through R, too? I do not dig it by the way. Python is closer to me somehow. For some reason libraries for machine learning are written primarily for it. Although if there were enough of these libraries for C# I would not bother to learn Python. But again, to write everything for yourself is to kill a lot of time. And one tester is not enough for a trader.
No doubt that testing in R (Python) is much more convenient and diverse than in the terminal: there are a lot of features there that will never exist in the terminal (crossvalidation, bootstrapping, Monte Carlo....). And these chips allow to justify the FUTURE behavior of an EA.
But there is one, extremely important nuance that makes testing in the terminal very desirable: testing in the terminal is extremely similar to real trading, with meta-quotes consistently trying to reduce the existing differences.