Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 42
![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
I am interested in understanding the issues.
Which ones? So far, I haven't heard any questions that you're really interested in figuring out.
The ninth page talks about several other people with the same results as yours, but with different approaches, some through spectrum analysis and others through a neural network.
Or here: "Is there a process, the analysis of one part of which prevents from predicting the next part?
Me too... "floating"... Maybe "someone else"? Would you mind voicing your "questions about Alexei's model? As a point of reference for moving forward.
Yes, I agree. Since the topic has been brought up, I will answer constructive questions. And by the way, I believe there are several people on the forum who understand better than I do what I've been doing in this thread.
If the question is about noise, I'll pass, I haven't dealt with the noise issue.
Paukas and TheExpert keep repeating that there is no noise. I listen to their opinion, but they themselves, I believe, have not created a model that perfectly describes the cotier and leaves no errors. They simply postulate the possibility of reaching such an ideal at some point. Maybe that's roughly how they trade by hand.
Noise is a mask, an allusion used when a predictive model works with errors - yes, I agree. We think a fair coin falls out at random, meaning we can't predict with a probability other than 50%, although theoretically we can predict with better accuracy.
It's just all naked theory, a kind of technical faith. Looking into the future, and in real reality, TA models are built that give errors.
...
And yes, I do use TA, which is accurate to a point (not always, not everywhere, noise can get in the way, I'm just getting to grips with the issue) in real life forecasts.
...
And maybe these very not everywhere and not always is the very spork in a seemingly great system (sorry, tactics) of trading for plus or minus half a point?
Why do we need to show the Statment?
So that opponents (like me) have no doubts about the robustness of the system (excuse me, tactics) as a TOTAL. And we have such masters here in words, it's amazing sometimes.
I'm not interested in my image, and I do not need the money.
I am interested in sorting out the issues.
Are you a saint? You have to admit, a statement like I don't need the money is alarming right off the bat.
You want to know if there is noise, how do you define it, and if there is, how do you separate it? Have you ever thought that macroeconomics cannot be measured with a ruler? Of course there are supports and resistances, trend lines and many other things that have a basis. Economic, mass-psychological or whatever. Does TA have a basis?
Smoked it, thought about it, and yes!
Let's just leave out the third parties, OK?
You, as I understand, are not a beginner, you have no problem with it. Show me, please, a pattern or whatever you had in mind that would work universally on any numerical sequence, is formalized enough to be turned into a code and has predictive value for the future (note, not the past).
I listed exactly what TA has.
Have you ever thought that macroeconomics cannot be measured with a ruler? Of course there are supports and resistances, trend lines and many other things that have a basis. Economic, mass-psychological or whatever. Does the TA have one?
Okay, I get it. You don't know anything about "..." or about TA.
We do not have to, by the way.
But over the years we have written and shown so much that I won't repeat myself here, so as not to clutter up the resource (and there are enough places where it gets dusty).
"Have you ever considered that macroeconomics cannot be measured with a ruler?"
Are you serious?! That's what TA was created for, too. And yes, measure it.
Google it if you need to. Do your research. No, no problem.
What a grail you're describing...
We'll talk again tomorrow, okay? I just have to go.
It's not certain that I'm going to show you something that fits the description, but I'm not jumping off the topic. Just force majeure.
I'll see you tomorrow.
Paukas and TheExpert keep repeating that there is no noise. I listen to their opinions, but they themselves, I believe, have not created a model that perfectly describes the quotient and leaves no errors. They simply postulate the possibility of reaching such an ideal at some point. Maybe that's about how they trade by hand.
Yes, I agree. Since the topic has been brought up, I will answer constructive questions. And by the way, I believe there are several people on the forum who understand better than I do what I've been doing in this thread.
If the question is about noise, I'll pass, I haven't dealt with the noise issue.
Paukas and TheExpert keep repeating that there is no noise. I listen to their opinion, but they themselves, I believe, have not created a model that perfectly describes the cotier and leaves no errors. They simply postulate the possibility of reaching such an ideal at some point. Maybe that's roughly how they trade by hand.
Noise is a mask, an allusion used when a predictive model works with errors - yes, I agree. We think a fair coin falls out at random, meaning we can't predict with a probability other than 50%, although theoretically we can predict with better accuracy.
It's just all naked theory, a kind of technical faith. Looking into the future, and in the present reality, TA models are built that give errors.
Try articulating what noise is.
Well, you're describing the grail...
Tomorrow we`ll continue our conversation, OK? I just have to go.
Of course, it's not a fact that I'll show you something that in every way fits the description, then I'm not jumping off the topic. Just force majeure.
I'll see you tomorrow.
We'll continue when it suits you ;)
It's not the stats, of course, the show.
It was in the framework of one project. Internet Business TV. For 2.5 months there was a real trade. Here are 4 of the 5 final programmes. You'll be exhausted listening, but you'll have to overpower yourself;) So these are just the second halves of the summaries. All the programs where entries/exits/stop transfers/partial closes were made during this time, with all the explanations, goals, I can ask for it all. But, be lenient and spare me that ;)
All the work was done on TA only.
Alexey!
Forgive me, please! I've flooded your whole thread. An interesting topic.