Discussion of article "Exploring Seasonal Patterns of Financial Time Series with Boxplot" - page 20

New comment
 
Igor Makanu:

why flud? - This is my opinion, and it doesn't have to coincide with yours, especially since I wrote my conclusions after reading the article and discussing it.

I see no point in discussing further, I am not afraid of enemies, but to explain the obvious is just not interesting, you think that the methodology works..... well, as it is known on this forum - go to the real! ;)

I'll use an example from the set of your arguments - on a bob.

 

I will summarise the article (with discussion) and remove myself from the thread.

  1. The article shows how convenient it is to do statistical research in Python.
  2. On the basis of statistical studies, a robot is written that shows a profit on the study interval.
  3. It is concluded that this is a pattern and not a fitting. Since it was not the Optimiser that was used, but a 100% interpretable statistical study.
  4. At the same time, the Expert Advisor is not demonstrated on OOS (outside the interval of the statistical study).
  5. It is presented that this statistical study is a good example of finding a market pattern.
  6. It is claimed that finding such a pattern is a very difficult task, which the author has finally solved.
  7. It is said that after the statistical research the author created an optimisation version of the robot and found good input parameters, but he did not report anything about it in the article, because it is about something else.
  8. There is no real account monitoring.

If I made a mistake somewhere above, it was not on purpose. So understood.


Now what I said in the discussion from my side.

  1. The result of any optimisation is a statistical study, even if there is no 100% interpretation.
  2. A classical statistical study is the result of implicit optimisation on the interval of the statistical study.
  3. It is wrong to talk about a pattern without a result outside the statistical interval. You need OOS, at the very least.
  4. Random luck, when the best Optimisation pass on OOS shows a good result, is possible. But it's better than nothing.
  5. An EA is published that finds the "pattern" from the article in three minutes, showing a result just as good. The EA matches the one the author did not include in the article.
  6. It is claimed that any Expert Advisor must contain a filter by time of day. Otherwise - serious omissions in the search for patterns.
  7. It is claimed that trading deviations from the MA is one of the simplest and most famous TS.
  8. A method is suggested that anyone can see a good result on OOS according to the research in the article. Whether it is luck or something else - without analysis.
  9. The example of the posted Expert Advisor demonstrates that it is easier and more efficient to conduct such studies in the Optimiser without involving other tools.

 
fxsaber:

I'll summarise on the article (with discussion) and remove myself from the thread.

  1. The article shows how convenient it is to do statistical research in Python.
  2. On the basis of statistical studies, a robot is written that shows a profit on the study interval.
  3. It is concluded that this is a pattern and not a fitting. Since it was not the Optimiser that was used, but a 100% interpretable statistical study.
  4. At the same time, the Expert Advisor is not demonstrated on OOS (outside the interval of the statistical study).
  5. It is presented that this statistical study is a good example of finding a market pattern.
  6. It is claimed that finding such a pattern is a very difficult task, which the author has finally solved.
  7. It is said that after the statistical research the author created an optimisation version of the robot and found good input parameters, but he did not report anything about it in the article, because it is about something else.
  8. There is no real account monitoring.

If I made a mistake somewhere above, it was not on purpose. So understood.


Now what I said in the discussion from my side.

  1. The result of any optimisation is a statistical study, even if there is no 100% interpretation.
  2. A classical statistical study is the result of implicit optimisation on the interval of the statistical study.
  3. It is wrong to talk about a pattern without a result outside the statistical interval. You need OOS, at the very least.
  4. Random luck, when the best Optimisation pass on OOS shows a good result, is possible. But it's better than nothing.
  5. An EA is published that finds the "pattern" from the article in three minutes, showing a result just as good. The EA matches the one the author did not include in the article.
  6. It is claimed that any Expert Advisor must contain a filter by time of day. Otherwise - serious omissions in the search for patterns.
  7. It is claimed that trading deviations from the MA is one of the simplest and most famous TS.
  8. A method is suggested that anyone can see a good result on OOS according to the research in the article. Whether it is luck or something else - without analysis.
  9. The example of the posted EA demonstrates that it is easier and more efficient to conduct such research in the Optimiser, without involving other tools.

Unfortunately, your conclusions are wrong for a simple reason: you misunderstand the essence of "regularity" and do not distinguish it from the Law.

A law is a rule that always holds and is proved mathematically.

A regularity is an observed repetition of an assumed relationship of events or values. A proven regularity is a Law.
Laws do not exist in the area of price movement, and therefore the author does not have to prove anything. He saw some regularity in the repetition of values of the observed parameters and used it in trading.

If the author had added the position of the stars to the research and found a connection with the price, he would have been right about the existence of a regularity and would have used it profitably.

 
Реter Konow:
Unfortunately, your conclusions are wrong for a simple reason: you misunderstand the essence of "regularity" and do not distinguish it from Law.

A law is a rule that always holds and is proved mathematically.

A regularity is an observed repetition of an assumed relationship of events or values. A proven regularity is a Law.
Laws do not exist in the area of price movement, and therefore, the author does not have to prove anything. He saw some regularity in the repetition of values of observed parameters and used it in trading.

Not some pattern, but a seasonal pattern. It's important.

If someone misunderstands something and appropriates the results of the study for himself, it is his personal problem, we forgive him.

As a matter of fact, he took a ready-made bot from the article and implemented it. This, of course, is his great achievement. And proves now that he himself found the pattern through genetics.

 

The point of this action is that one person wanted to appropriate the result of the research for themselves because they liked it so much. The rest of us just didn't get it.

and that's the only reason that's the only reason you have to read it

It's about psychology.

 
Maxim Dmitrievsky:

The point of this action is that one person wanted to appropriate the result of the research for themselves because they liked it so much. The others just didn't understand it.

keep it, I understand, but no thanks ;)

 
Igor Makanu:

keep it, I get it, but no thanks needed ;)

You're one of those who didn't understand anything )

What is the econometric approach to finding seasonal patterns outlined in the article?

This topic has not been fully explored and there is more to explore, but to do so you need to realise the current results
 
Maxim Dmitrievsky:

You're the kind of person who doesn't get it )

What is the econometric approach to finding seasonal patterns outlined in the article

This topic has not been fully explored and there is more to explore, but to do so you need to realise the current results

I understood everything, and searched for seasonal patterns about 6-7 years ago, alas there are none, but there is a trend that either persists for a long time or has already ended, your example of a found pattern 2017-2019 is a found trend - price drift, you can find a lot of things on it, you can even sit out a loss, which is actively being done now in signals ;)

I offered you to demonstrate the methodology of pattern search on other CRs, take crosses, this is also a CR, but it does not contain information about USD, there you will not be able to fit the methodology to what is already visible on charts ;)

 
Igor Makanu:

I understood everything, and I was looking for seasonal patterns about 6-7 years ago, alas there are none, but there is a trend, which either persists for a long time or has already ended, your example of a found pattern 2017-2019 is a found trend - price drift, you can find a lot of things on it, you can even sit out a loss, which is actively done now in the signals ;)

I offered you to demonstrate the methodology of pattern search on other CRs, take crosses, this is also a CR, but it does not contain information about USD, there you will not be able to fit the methodology to what is already visible on charts ;)

You are misinterpreting the results of the study

You should read exactly what is written, without adding new terms that have nothing to do with the topic.

There is a library for analysing seasonal patterns for any instruments, I have my own goals

 

As promised, it was a bomb - it bombed a lot).

I wanted to throw some firewood in the form of discussions on the topic of determining the significance of differences between samples on the example of Mood's exact median test. But now I see that there is enough of it here already).
1...131415161718192021222324252627...33
New comment