Discussion of article "Exploring Seasonal Patterns of Financial Time Series with Boxplot" - page 18

New comment
 
fxsaber:

I now know the name of the author of the trade MAshka.

That's why I see only grails around, which die after 2 months, in dts with negative spreads.

Where are we compared to them?

 
Maxim Dmitrievsky:

That's why I see only grails around, which die after 2 months, in dts with negative spreads

Where are we compared to them?

I understand your move to another topic.

 
fxsaber:

I understand your convention on another topic.

Well, like, I didn't get anything useful for me either, though I kept silent.

 
Maxim Dmitrievsky:

Lube arguments from you and other opponents as to what it is:

  1. just Mashka and over-optimisation.
  2. The pattern was not found through boxplots.
  3. you would have easily found it yourself through optimisation (without knowing where to look at all).
  4. this is not a pattern

Doesn't stand up to any criticism, just nitpicking due to some specific misunderstanding of the material.

1. alas, this MA is no more and no less, and if we take into account that the time used is on the border of the day's close/opening, it is most likely that the colour of the daily bar can be used at all.

2. if it is a pattern, it should work on several time intervals, or there should be a method of cutting off when this "pattern" will not work.

3. alas, I agree! - but only because the volume of input data (OHLC) is huge (9 TFs * per number of bars in the history * number of time intervals.... in general, the problem of combinatorics with factorials, if memory serves, combinations with repetitions) + search on the TS itself, the code of order opening by time and simply by the proposed direction of trade I posted? - it did not show anything, so you found a TS that works by indicator + by time interval + on a particular section of history. And the statement that this section and TS was found with the help of mathematical method.... is very doubtful, I would say it happened by chance that the results of the research coincided with the results of testing in the tester.

4. it's usually always like this, it works or it doesn't, if the method works, then .... ask the participants to give you the names of other symbols and you will quickly check your method and the participants will check yours ;)

 
Statistical regularity doesn't require breaking spears. One day it's there, the next day it's not. That's what makes it different from a law. But, sometimes statistics finds something. In general, if a pattern is detected, you can use it at your own risk. But to prove it.
 
Igor Makanu:

1. alas, this MA is no more and no less, and if we take into account that the time used is on the border of the closing/opening of the day, then most likely we can use the colour of the daily bar at all

2. if it is a pattern, it should work on several time intervals, or there should be a method of cutting off when this "pattern" will not work.

3. alas, I agree! - but only because the volume of input data (OHLC) is huge (9 TFs * per number of bars in the history * number of time intervals.... in general, the problem of combinatorics with factorials, if memory serves, combinations with repetitions) + search on the TS itself, the code of order opening by time and simply by the proposed direction of trade I posted? - it did not show anything, so you found a TS that works by indicator + by time interval + on a particular section of history. And the statement that this section and TS was found with the help of mathematical method.... very doubtful, I would say that it just happened by chance that the results of the research coincided with the results of testing in the tester.

4. it's usually always like this, it works or it doesn't, if the method works, then .... ask the participants to give you the names of other symbols and you will quickly check your method and the participants will check yours ;)

I didn't expect such nonsense from Igor at all. He couldn't even make a proper bot for checking, he doesn't understand at all :)

So, the level of the article is unattainable for 99%, you need to write even simpler.

clogging the brain with unnecessary words

 
Maxim Dmitrievsky:

I didn't expect such nonsense from Igor at all. He couldn't even make a proper bot to check it, he doesn't understand at all :)

So, the level of the article is unattainable for 99%, you need to write even more simply

clogging the brain with unnecessary words

of course...

and so half of the article and the lion's share of discussions is spent on explaining what a box-plot is :-))

MA doesn't go any further than that.

 
Maxim Kuznetsov:

of course...

and so half of the article and the lion's share of the discussion is spent on explaining what a box-plot is :-)

MA doesn't go any further than that.

Some characters have a hidden protest: "How so, we have been optimising all our lives, falling kernels, and this upstart has shown how easy and playful it is to find a seasonal pattern".

Denying the results of the article, trying not to recognise them is a subconscious defence. Moreover, the method is so simple and intuitive that it stings twice as much and makes you feel bombarded by it

 
Maxim Dmitrievsky:

Some characters have a hidden protest: "How is it, we have been optimising all our lives, falling kernels, and this upstart has shown how easy and playful it is to find a seasonal pattern"

Denying the results of the article, trying not to recognise them is a subconscious defence. Moreover, the method is so simple and intuitive that it stings twice as much and makes you feel bombarded by it

I never thought I'd hit my inner strings so hard..... What nonsense you attribute to others. Be stronger.

 
fxsaber:

I never thought I'd hit my inner strings so hard..... What nonsense you attribute to others. Be stronger.

I'm almost never wrong on this side of things. And I'm not just talking about you

I'm just responding to comments, sharing my opinion.

1...111213141516171819202122232425...33
New comment