The essence of optimisation - page 3
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
Has anyone tried to correlate robustness of optimized parameters and optimization results with figures?
IMHO:
The most important factor when analysing optimisation results is the degree of "smoothness" of the hypersurface in the parameter space under test with e.g. Sharpe's coefficient as the target function.
The same also determines the robustness.
The degree of regularity ("smoothness") as well as the hypersurface topology itself in the parameter space is a powerful tool for analysis and strategy synthesis, though very exacting as to the sanity of "stylized facts" upon which the trading algorithms are based as well as to the machine power.
Has anyone tried to correlate robustness of optimised parameters and optimisation results with numbers?
In general - yes, I've tried. There is my product for this purpose in the Market (it's a small device for robustness catching).
Read my opinions. Most of them are actually wrong (sorry, I'll take the liberty to say so). "To have a good idea" or "smoothness of hypersurface" - nonsense (I've got sick of this nonsense myself).
Optimisation implies a decision-making process aimed at changing the process/appearance being managed towards the desired results. Hence - an in-house optimiser is not an optimiser, but essentially a batch tester. And in "brute force" mode, the "optimisation" criteria, such as "profitability" and all the others, have no meaning at all.
Have read the opinions. Most of them are actually wrong (sorry, I'll take the liberty of saying so). "To have a good idea" or "smoothness of hypersurface" is nonsense (I myself have had this nonsense).
Having an illness does not mean that you are cured. It is possible that it is a chronic illness...
Flood flood flood...
Everything is much less dichotomous and much more structured and hierarchical than you all think.
Partly right are both those who say that optimization is an illusion, in the context of simple algorithms, and those who argue that it is fundamentally important, when the logic of algorithms is more adequate.
I am not a quantum, I only command them:) So I will not spout here about all sorts of DA,EP tests on adequacy of models and Montecarl optimizers, making super weapons out of simple models. There is a lot of open information about it on the web . But what would it work and bring in the money, just use them directly will not work, you need a team of inventors who understand all these models and modify the current reality, and one suchRhD lamb costs about $ 100K a year (in Russia) and need them 3-5 copies to rent for a year at least, but most likely for 2-3 years. This is not you here "advisers" to trade at 100 Bakinsky))).
About the capital requirements I'm even silent. So as not to spoil anyone's mood here.
Therefore, in civilized countries, suckers are protected by the state from investment activities, as a priori they have no chance.
Flood flood flood...
Everything is much less dichotomous and much more structured and hierarchical than you all think.
Partly right are both those who say that optimization is an illusion, in the context of simple algorithms, and those who argue that it is fundamentally important, when the logic of algorithms is more adequate.
I am not a quantum, I only command them:) So I will not spout here about all sorts of DA,EP tests on adequacy of models and Montecarl optimizers, making super weapons out of simple models. There is a lot of open information about it on the web . But what would it work and bring in the money, just use them directly will not work, you need a team of inventors who understand all these models and modify the current reality, and one suchRhD lamb costs about $ 100K a year (in Russia) and need them 3-5 copies to rent for a year at least, but most likely for 2-3 years. This is not you here "advisers" at 100 Bakinsky to trade))).
About the capital requirements I'm even silent. So as not to spoil anyone's mood here.
Therefore, in civilized countries, suckers are protected by the state from investment activities, as a priori they have no chance.
Another hoaxer who thinks you need a hadron collider no less to trade...
J.B:
the hypersurface topology itself...
Robust strategies can only be in near-market hyperspace.
Partnerships.
Robust strategies can only be in near-market hyperspace.