Where is the line between fitting and actual patterns? - page 26
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
I haven't seen any discussion about the profit factor. There is only discussion about its usefulness, its application, etc. Why?
Because there is a generally accepted definition.
Everyone uses the phrase "OOS" as he likes. I've even mentioned "ten consecutive OOC's".
Plus there's confusion about where on the timeline: is it the real future, or a virtual one?
And everyone's right.
In the financial markets, everyone "pulls the blanket" over themselves - most cannot be in profit, so there are unlikely to be common definitions - it does no one any good.
Why don't we talk about quality? ........ We will! I wrote "not yet" )))
And the timing depends on how you in your 2H point estimate the right sample. Looking at all balance curves? ......... Yes, looking at all the curves and even bits and pieces of them, but programmatically. (saving my eyes).
It's longer than criterion filtering. Here you are prying the wonder of approaches from others, while you yourself are talking in riddles... ..... Why, I tried to describe it more or less here, but the thread died immediately afterwards. With my activity in this thread, I'm trying to resuscitate that one. (An old friend is better than two new ones.)
Instead of wasting a lot of time on your scheme, you would describe failure of "traditional approach" (by the way, why is it traditional?) and all the charms that are easily seen in your approach. I.e. how from
How will you filter them out PF and others in the 1/5 area? Do you count them separately? ...............................
1) You can also do it separately if you need a PF of 0.01.
What are you worried about? Isn't the machine counting what it will?
I'm worried about you, though, that you analyse the OOS ten times.
2) There are many other criteria.
In the financial markets everyone "pulls the blanket" over themselves - most cannot be in profit, so there are unlikely to be common definitions - it does no one any good.
I don't think this applies to forex
We have a "blanket without measure", and he who "stays awake" and always keeps an eye on the temperature in the room earns).
Figar0:
You say that you are preparing data for training. Please elaborate, how long have you been using these methods? Something in your words is very familiar, I remember, I suggested, like in the branch about the context, to prepare a synthetic data with the required parameters for optimization, so you can change the data parameters and see the response of the TS. I think in de like you were just agreeing with me, but suggesting a slightly different option to mine - prepare data from real pieces of history, is that right?
I use for a long time, from real pieces (I tried to generate something, it turns out worse, it is difficult to catch and convey the characteristic features of the instrument movement, and they are), the principle of selection can be different sometimes it is just pieces of different types of movements of the required length, sometimes they are filters based on the description of the current situation, like in the example above. The process is creative) but rewarding.
preparing the data for training by some rule is simply introducing an additional filter into the system.
Yes and no. This filter is not used to make trading decisions, is not trained, and is not part of the trading system. Although you are right in part.
I don't think this applies to forex
We have a "blanket without measure", and he who "stays awake" and always keeps the temperature in the room earns)
Ask any of the successful managers for the source code of TS - you know in advance what you'll get, right?
So all the same "blanket" is not dimensionless :)
Even more try to cherish the "theory of TS building", because they believe that having a "successful theory" they can always create a "successful TS" or even many successful TS.
That is why all theoretical discussions ultimately result in banal rubbish - everyone is trying to benefit from the discussion, while maintaining their own "big, big secret" :)
Literally everyone demonstrates temporary successes, but strenuously hides how this success was achieved.
In general, conditions for "brainstorming" are not observed on any of the well-known financial market forums.
Everyone is trying to benefit from the discussion, but keep their own "big, big secret" :)
Ask one of the successful managers for the source code of TC - you know in advance what the answer will be, don't you?
So all the same, the "blanket" is not dimensionless :)
Even more try to cherish the "theory of TS building", because they believe that having a "successful theory" they can always create a "successful TS" or even many successful TS.
That is why all theoretical discussions ultimately result in banal rubbish - everyone is trying to benefit from the discussion, while maintaining their own "big, big secret" :)
Literally everyone demonstrates temporary successes, but strenuously hides how this success was achieved.
In general, conditions for "brainstorming" are not observed on any of the well-known financial market forums.
On the subject, I told you how I do it myself, I don't see any secrets in it. All the more so because there has already been a lot of discussion about optimization. And speaking of codes, who will just give up his well-worn and long-worn experiments and bumped heads? Therefore, we see mainly tester toys which will show different results on the real market.
So anyway, no one can say in advance whether his "hard-earned good" will fail in the future or not :)
I'm not worried about putting out the code, because I know that almost no one will use it anyway - for fear of losing it.
I have even conducted an experiment - I have tried to give a commercial Expert Advisor as a gift, but they have refused to take it. So now no one will ever know how it will work in 2011.
It's a laugh and a sin :)
Ask for the source code of TC from one of the successful managers - you know in advance what they will answer you, don't you?
So all the same, the "blanket" is not dimensionless :)
Even more try to cherish the "theory of TS building", because they believe that having a "successful theory" they can always create a "successful TS", or even many successful TS.
That is why all theoretical discussions ultimately result in banal rubbish - everyone is trying to benefit from the discussion, while maintaining their own "big, big secret" :)
Literally everyone demonstrates temporary successes, but strenuously hides how this success was achieved.
In general, conditions for "brainstorming" are not observed on any of the well-known financial market forums.
+10.
What if we apply optimisation? Lately I've been thinking more and more about this.....
Let me explain.
..............................
Scanning known finmarket forums.
Selecting adequate, interesting, listening and listening people.
A private, 'closed forum'.
Brainstorming.
Everyone has the same goal. Motivation is there. And the blanket won't rip... )
That's for sure. And some talk in riddles, like software looking at pieces of balance curves, and telepathically calculating the parameters of a strategy run on 1/5 of that same run and giving everyone a "fail". Others are talking about some 10 OOS and a bright future...
Well, come on....
What's your problem? You want 10 OOS? OK!
Approximate solution:
Take the Sample range, single and indivisible ))
We divide it into 11 parts and for each part in the EA, calculate the linear regression and the "variance" ("or your unique criteria").
At 2H, you programmatically analyse all this goodness.
And that's it.
I think it's actually better than looking at it with your eyes.
Or don't you agree? Or are there mysteries left?
............................
Fractal post again.....
Oh, this TA.... ))