Principles of working with an optimiser and basic ways of avoiding fitting in. - 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
You discuss the same thing in every thread - your model. I think everyone has already commented on it more than once))
By the way, you can make money on a zigzag :) Speaking of zigzags
Actually, the aim is to shift the expectation, not to simulate a quotidian.
In the past or in the future?
Actually, the aim is to shift the expectation, not to simulate a quotient. Stationarity is essentially not needed at all.
Now that's really the salt of the matter. I agree again.
Now that's really the salt. Again, I agree.
It's like a riddle. The right answer is 112 and bang :)
"What Where When?
That's where the real experts have... what do you call it... "sense of the right answer". When they hear the right answer, they know in their gut that it's right.
I don't think I'm an expert. But I think (hopefully :) ) that the process is similar :)
This is quasi-stationarity - a change in Mo over a certain range. Maybe it's not just about mo, but in this context we are most interested in mo
So maybe it's a super complex method, but it's rough enough to estimate the regularity.) It's more a question of the number of system parameters and the sensitivity of the result to their change. If a small change in the parameter causes a change in the result, this is not good. There are other signs. I just recently wrote about it here https://www.mql5.com/ru/forum/137614/page5
Try to create a super complex method with a minimum number of parameters. The longer the formula, the more parameters it has. Of course it's not a law, but it's a good approximation to reality. Take an elementary function (model) y=ax. It has one parameter "a" which allows changing the angle of slope of the straight line. And that's it. Try to fit this model to the market. Let us take a more complicated model y = ax^2 + bx. It is more complex and has two parameters. It will definitely be better on history. Now let us break it into 2 submodels and test them separately: y = ax^2 and y = bx. Each of them shows poor results, so the sum of these results is much lower than the original model? There is a high probability that this is a fit. Not every simple model guarantees profit, but in any case simplicity reduces the probability of fitting.
I will try to describe the shift method and ways of breaking the model under test into smaller ones in more detail later.
I wonder from what principles it follows that it should be reversible (in time, or what?)...
Physicists have known for decades that there is no perfect symmetry in nature.