The essence of optimisation - page 6
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
It's just a fashion. It's a new one. To shit on those who have a lower rating. Do not worry, save your rating. Or come back in five years, maybe the trends will change.)).
If you have an inferiority complex based on low ratings, instead of themes with ex5 would put something useful in the base.
And here the manner of behavior of the topicstarter personally reminds me of the manner of hrenfx: said "A" and does not want to say "B".
No one here (in this thread) is shitting on anyone. The participants in the discussion have expressed their attitude to optimisation without looking at the ratings.
This does not apply to you. And to TheXpert too. Told you a mod. A new one.
And there is a fashion. It's hard not to notice. By the time you get rid of a bunch of empty remarks, everyone has already forgotten the question.
If you have an inferiority complex based on low ratings, instead of themes with ex5 would put something useful in the base.
When last time I was unranked, I did not say a word. For I do not think it is important. It, by the way, was not returned, and so on. So don't.
I am not exposing code to the kodobase because of my own reasons. Well, I have no desire to.
It's just a fashion. It's a new one. To shit on those who have a lower rating. Do not worry, save your rating. Or come back in five years, maybe the trends will change)).
No kidding.
I figured it was all about ratings, thanks for clarifying.
I figured it was all about ratings, thanks for clarifying.
Oh, come on. Maybe I overreacted. I was being cordial. There's a really literate crowd in this thread...
The general tendency is just that. To trash, to shut down any topic. The battle for the top is on... Mostly by newbies.
Who's offended, I apologize.
Oh, come on. Maybe I overreacted. I was being cordial. There's a really literate crowd in this thread...
The general tendency is just that. To trash, to shut down every topic. The battle for the top is on... Mostly by newbies.
I'm sorry if I offended anyone.
Okay. I'm half-joking about the rankings too, I don't really care.
In general, one of the ideas is that we analyze "window testing" chart, "walk forward" as it is also called, but in this case it is "walkbackward", i.e. for example, having a million bars of the tested symbol we increase it by 100 000 shifts by 10 bars and look at dynamics of parameters.
So, we get something similar:
You can clearly see the regularity of the "migration" of extrema in the process of window shift, this regularity being quite "slow" in the sense of "inertial", i.e. predictable for some strategies and quite random for others.
Dynamics of extremum shifts for "qualitative" types of strategies I have figured out how to approximate and predict ahead, though frankly speaking it is not as trivial as it may seem at first sight.
If such research makes sense for someone, I can continue thinking, if not, I can't.
...
One can clearly see the regularity of the "migration" of extrema in the process of window shifting, and this regularity is rather "slow" in the sense of "inertial", i.e. predictable for some strategies and rather random for others.
Dynamics of extremum shifts for "qualitative" types of strategies I have figured out how to approximate and predict ahead, though frankly speaking it is not as trivial as it may seem at first sight.
If this kind of research makes sense to someone, I can continue thinking, if not, I can't.
Of course this is all very interesting. Thank you. Carry on. Any research is interesting, even if it doesn't yield positive results.
Then immediately a question on the extremum picture.
You can clearly see that the backtest intervals overlap. This means that "slow" regularity can be ensured simply by these overlaps. Almost like a mashup.
What is going on with the forwards?
Then let us immediately ask a question about the picture of extrema.
You can clearly see that the backtest intervals overlap. This means that "slow" regularity can be ensured simply by these overlaps. Almost like a mash-up.
What is going on in the forwards?
Yes, and they overlap significantly, in this case 99.9% overlap, so the analogy with mashka is relevant, you can also draw parallels with window spectrograph, window correlogram, etc.
SMA differs from MO in the same way that windowed testing differs from single-sample testing.
The point is precisely the specifics of the "mountains and troughs" dynamics.IMHO
As it was mentioned above about "correct strategies", so the more "correct" a strategy is, the more predictable is the dynamics of extrema in cocoon testing and the less noisy is their picture, there are still raw quantitative considerations how to calculate that.
As a forward, the question is rather subtle, depending on what to consider a forward ... In my opinion, to fit forward or backward makes no difference, and one-time validation by forward is often not very representative. So, in general, it's still in development))) Too many technical points still need to be solved.
So far there are many fragmentary tests, some with super performance and on forward, but everything is too "on the nose" at the moment.
The main idea is that such "windowing" in a sense "differentiates" the series to a state where it is relatively stationary and therefore predictable.
toxic:
In my opinion, it does not make much difference to fit forward or backward.
Your view is wrong, because what the parameters are adjusted to is called a semple (training sample) or backtest, and what they are not adjusted to is called an out of sample (outside the training sample) or a forward.
I.e. any part of the history, to which the fitting (optimization) is carried out, is a backtest and by definition cannot be a forward.