Where is the line between fitting and actual patterns? - page 18
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
Hmmm...)
Why does the AOS have to be after the training period? That doesn't make any sense, do it before the training.
Next, the training period itself, why does it have to be immediately before the current moment? What is this heresy? Suppose the tool is falling for a week, then we optimize it and what do we get? If it still loses, it's good, but what if it rises?
For example, I take a "synthetic" training period composed of pieces of movements that I expect in the future, or simply of pieces covering all possible variants of behavior.
I think I've made up a picture too.Malevich rests.
I'm adopting a style of communication on this forum that is well established and carefully cultivated.
I agree, posts like this are very easy to write, I'll get to the big numbers under my avatar soon enough.
Question is, how much is this car worth?
Fitting is an attempt to make the best possible use of the patterns that actually exist at a given time interval. So there is no margin and there is no difference.
Whether you can use the patterns you find and beyond is another question. Only the forward test will reveal the truth.
Well, if the topicstarter himself doesn't mind turning this thread into another dump (sorry...),
you're on, gang!!!
Let's have a blast....
Let's have a blast....
Nothing changes, it's always Friday here) And I found a cool game http://www.worldoftanks.ru/, it's been 100 years since I played it, but I can't tear myself away for 3 hours...
By the way, it's a start.
Ask me specific questions and I will try to answer them.
In the meantime, I've asked you five or so specific questions.
In the meantime, I asked you five questions specifically.
And I didn't think they applied to me. Not much specificity either. Oh,come on -..... ))
Why does the AOS have to be after the training period? It doesn't make any difference, do it before the training.
1) I didn't say that, and "synthetics" like yours are good too, and interesting to test the stability of the TC.
2) But, you won't really be trading in the past. Time moves from the past to the future. This is the pattern shown in the picture. No more than that.
Further, the training period itself, why does it have to be immediately before the current moment? What is this heresy? What is the point?
25 again!
On the screenshot my yellow plot, clearly signed "Real, demo or new history plot"
Move your point on the history timeline denoting "current moment" as you wish. As long as there is a profit on the OOS.
Move your point on the timeline of history denoting the "current moment" as you wish. As long as there is a profit on the OOS.
You have interestingly written "Real, demo, or new piece of history". So, if you do have a piece of history, then your interpretation is no different from the "classic neural network". Let's exclude real and demo from the OOS generalisation though.
I thought from the picture that you have learned (somewhere in the shaded area) to do without classical OOS, i.e. without checking the results of training on a piece of history other than OV (training sample). That's what I wanted to know:)
You have interestingly written "Real, demo, or new piece of history". Well, if you do have a piece of history, then your interpretation is no different from the "classic neural network". Let's exclude real and demo from the OOS generalisation though.
I thought from the picture that you have learned (somewhere in the shaded area) to do without classical OOS, i.e. without checking the results of training on a piece of history other than OV (training sample). That's what I wanted to know how).
You're thinking right. It's just a little bit left for our mutual understanding.
Your "classic OOS" is an illusion.
In fact there are only two periods(rays) and a point separating them on the timeline.
- Past and future
- Sample and OOS, OV and TV
The point (or "current moment") we can virtually move to the left ( to the past) as we wish. To the right (to the future) we cannot - because there is no data.
If you don't mince words, in general, do you agree?
............................
Then the question is: What are the fundamental differences between OB(sample) and your "classic OOS"?
The fact that your TS didn't see OOS? That answer I know.
................
This is all the same plot from which a piece has been "cut off" and given a big name -- OOS.