From theory to practice - page 110

New comment
 
Nikolay Demko:

What, you just gave up on me the first time?

What happened to "prove physics is cool"?

Whoever keeps stepping on the same rake without changing anything is insane. It is enough to change something and it is no longer insanity but a new experiment.

I want to draw your attention that the tick history is not deep, and it is possible that shamanism with the methods of stat processing simply lead to a fit with existing history.

In order to avoid that you need to have a piece of history, for which you can't optimize methods. And even in this case we can not guarantee that the history won't be compromised.

After all, a person during the development process: invents the processing methods, optimizes the parameters on the test leg of the history and then checks them on the unknown data. If the loop fails, it is repeated. If we imagine that the cycle is long enough, we cannot guarantee that we just don't adapt the methods to the unknown fragment as well. Because already on the second loop the unknown fragment is in principle already known. We have some information on it that some methods will not work on it although they will work on the test one.

I want to say that story tests are an important part of development, but an important but insufficient part and must be done correctly.

Let's see, Nikolay.

Of course, we will have to optimize the tool. The main subject is to monitor the distribution of tick increments. Of course, I'm going too far - now I have 18 "technical passports" of currency pairs and I need 36!!!!. I have to constantly monitor the parameters of these distributions with very large sample sizes. I stand by my opinion - the process of tick increments is stationary in nonparametric statanalysis. But one has to check, of course.

The biggest problem is determining the optimal sample size. A fraction of a percent error in the variance calculation immediately gives a difference of +-1000 or more ticks for the sample size.

It's a tough task - I'm not saying anything. And I don't think you have to give it up. A theorist is a theorist. There is no more powerful thing and will never be.

 

Well, and, of course, the method of data reception. A very important thing!!! The most important thing, I would say. I look back now - gigantic work has been done. I'm lucky that I'm an engineer in my office - I've given out tasks and go away, Vasya! If I was just an engineer, they would have kicked me out a long time ago :))))))))

 

Hello!

Alexander_K2, very interesting. I have been dealing with patterns for a long time, including on the tick chart. I have some success. I have understood how to receive "exponent" ticks, I have understood only approximately; therefore I am not ready to write a program for accumulating my own archive.

I would like at least look at the tick chart picture, or better yet, send me ticks "received via exponent" in a text file.

 
Wizard2018:

Hello!

Alexander_K2, very interesting. I have been dealing with patterns for a long time, including on the tick chart. I have some success. I have understood how to receive "exponent" ticks, I have understood only approximately; therefore I am not ready to write a program for accumulating my own archive.

I would like at least look at the tick chart picture, or better yet, send me ticks "received via exponent" in a text file.

I am just now busy compiling the data into the archives. If it is interesting - I will post it as I process it.
 
Alexander_K2:

Well, and, of course, the method of data reception. The most important thing!!! The most important thing, I would say. I look back now - gigantic work has been done. I'm lucky, I'm an inspector in my company - I've given out tasks and you can go away, Vassya! If I was just an engineer, they would have kicked me out long ago :)))))))).

I was a chief specialist in my firm, too. Graduated from a technical college, worked for many years, had to be promoted, but I couldn't be promoted to an engineering position. He was given the title of chief specialist and thus promoted.

I won't say anything about the consequences - they were horrific. Later nobody paid any attention to the diploma, and they (management was changed into managers) began to promote him further. Then, they say, the chief specialist (who was already a big boss) himself quit.

 
Nikolay Demko:

I'm going to have to get into your trial unity, if a ticking dance can by definition have a lag of 10 seconds, then you can't say that you're measuring a window until the 10 second lag has passed. How is your method then different from a 20 period wristwave set at half a period?

It's not my unity. I don't have any romantic feelings about 18.
Imagine you need a sample of a thousand values to predict some period of time. And for the forecast in smaller periods - where to get these values as not below a minute.
Quantitative analysis becomes qualitative. It's not pipsing, we're not predicting ticks by ticks, we're "predicting" (we generalize hypothetically)
larger intervals through ticks. Sometimes there is a lag of 10 seconds, it's not constant. And then it's not qualitative at this stage, but quantitative.
The main thing is to have a sample size. Small inconsistencies within the sample are not very important. What good is a 10 second lag if we're using that volume to predict
of much larger time intervals or for much larger movements than the size of the ticks themselves (whatever).
 
Alexander_K2:

Greetings, Nikolai! Think for yourself, eh? I'm too lazy to think before the New Year.

I'll tell you what I'll do.

If my TS shows profit within a month, I'll instantly post the model of this TS here on the forum for free.

Let everyone earn and Forex market go to hell - I do not give a damn. That's it!

Yep yep yep, we know we know. To get rid of this promise is like two fingers. It didn't work and there's nothing to get out of it. And if it works, we say it didn't work,
throw out some intermediate model that doesn't work, and kill the story. And you don't get caught and burn a working model)))).
Doubtful of course that you have it working. Let's see what happens. What is this thread all about in the end, and what were its aims.
The author actually have the results, or once again (branches have already been) leading to the conversation, blabbing at experienced material.
If the ultimate desire is regardless of the outcome - to lay out the material. Then
1-Why not conduct this process of research in general, laying out all the intermediate data and calculations, if there is a promise to eventually
It would speed up the process of getting to the final result, because people would be able to give more concrete advice, instead of guessing what the
the author has in mind. Not logical.
2-If you don't have results, then why should people who do have results help you. It might as well be easier for those people to just put everything
out in the open. And not make you Dartanian (Maybe he was an Armenian by the way D'Artanian).

 
Alexander_K2:
Greetings! I'm in the process of compiling data into archives. If it's interesting, I'll post it as I process it.

Thank you very much! I'll have a look, very interesting.

--

Happy New Year, everyone! And less pessimism! Remember, thoughts are material :)

 
ILNUR777:
Yeah yeah yeah, we know, we know. It's a two-fingered escape from that promise. If it didn't work, there's nothing to get out of it. And if it works, we say it didn't work,
throw out some intermediate model that doesn't work, and kill the story. And you don't get caught and burn a working model)))).
Doubtful of course that you have it working. Let's see what happens. What is this thread all about in the end, and what were its aims.
The author actually have the results, or once again (branches have already been) leading to the conversation, blabbing at experienced material.
If the ultimate desire is regardless of the outcome - to lay out the material. Then
1-Why not conduct this process of research in general, laying out all the intermediate data and calculations, if there is a promise to eventually
It would speed up the process of getting to the final result, because people would be able to give more concrete advice, instead of guessing what the
the author has in mind. Not logical.
2-If you don't have results, then why should people who do have results help you. It might as well be easier for those people to just put everything
out in the open. And not make you Dartanian (Maybe he was an Armenian by the way D'Artanian).

Ilnur, don't you pity people like the poorest Yusuf? Let them earn on their rent or some other stuff. I am convinced that Forex is for rich people. To be able to place large lots you have to have at least 2500$ on your account. Does everybody have it? Especially now, when the economic situation in the country is not good?
 

As a proof of serious intentions I enclose the model and the real working version for AUDCAD pair in the system VisSim + module of linking VisSim with MT4 (please do not laugh at the writing style - my first program in MQL4, but it works).

The model itself and .csv files should be placed in the directory C:\Forex

Files:
AUDCAD.zip  1504 kb
1...103104105106107108109110111112113114115116117...1981
New comment
Reason: