Where is the line between fitting and actual patterns? - page 56
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
The subject seems to have dried up....The edge is ephemeral......A follow-up question: Can we try to introduce some characteristic of system stability by analogy with the way it is done in radio engineering, sound engineering, etc.? Maybe in the form of bandwidth by level? Or quality factor? There is an optimized parameter (or several) and an optimum value in optimization (fitting - who calls it) was found. If there are several parameters it will be multidimensional, but we will have to somehow correlate them to time while the quality factor remains within reasonable limits.
The idea makes sense. This is roughly what I do after optimisation. Only the criteria are informal and intuitive. Like requirements or something. Like - if the system is unstable to small movements of parameters - then forget it. For parameters of a cotier will definitely move in the future. And not weakly. Further, if the OOS shows an immediate flush or even a too short (to what extent "too" is determined by eye) period of profitability - it's also in fault. If, on the contrary, the swing holds the parameters (does not subside too much) and the OOS increases equity, it enters the golden fund.
So it seems. I've never created numerical criteria. Although it seems nobody forbids it. Yeah. I have to think about it.
- Is it a glitch?
- No. It's a widget...
- Is that a butterfly or a Paukas?
- I'm both dreaming...
- Is that a butterfly or a Spider?
- I'm dreaming about both of them...
It's dough... between the cockroaches...
aha!
And Chapaev...
With Pepsi Co.
;)
The idea makes sense. This is roughly what I do after optimisation. Only the criteria are informal and intuitive. Like requirements or something. Like - if the system is unstable to small movements of parameters - then forget it. For parameters of a cotier will move in the future unequivocally. And not weakly. Further, if the OOS shows an immediate flush or even a too short (to what extent "too" is determined by eye) period of profitability - it's also in fault. If, on the contrary, the swing holds the parameters (does not subside too much) and the OOS increases equity, it enters the golden fund.
So it seems. I've never created numerical criteria. Although it seems nobody forbids it. Yeah. I need to think it over.
I do the same intuitively for the most part. "Quantitative" quality is indirectly defined as the ratio of the number of profitable runs in the tester to the total number - that's the "goodness". But it's too loose probably.
It's more than lax.
But fellow MegaDriver will teach you to understand the series...
It's important for Galton.
;)
And it's time to stop with the tester.
Demo accounts why do servers support them?
"Where's my North? ;)"
Test it...
Understandably, the internal battle of reason with excitement in the testor is set up optimally.
But that's everyone's internal problem.
So my manifesto is simple - demo, or even better - real!
There is no other way. And everything falls into place.
Or you start a debate - whether the ticks are properly generated in the batter?
And how in the real world?
Surely correct?
Why then "cheating"?
IMHO!
More than a little bit.
But fellow MegaDriver will teach you to understand the series...
It's important for Galton.
;)
If it's important, you can learn!
Thank you for your understanding!
I'm learning myself. every day.
I do the same intuitively for the most part. I define "quantitative" quality indirectly as the ratio of the number of profitable runs in the tester to the total number - that's the "goodness". But that's probably too loose.