The idea of making the world a better place - Articles, Library comments

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

MetaQuotes · 2021-09-17T19:51:46.0000000Z

New article Exploring Seasonal Patterns of Financial Time Series with Boxplot has been published: In this article we will view seasonal characteristics of financial time series using Boxplot diagrams. Each separate boxplot (or box-and-whiskey diagram) provides a good visualization of how values are distributed along the dataset. Boxplots should not be confused with the candlestick charts, although they can be visually similar. Prices have a shift in the average value over time and form trends, therefore the statistical analysis is not applicable to such raw series. Percentage price changes (price increments) are usually used in econometrics to ensure they all lie in the same value range. The percentage changes can be received using the pd.DataFrame(rates['close'].pct_change(1)) method. We need average monthly price ranges. Let us arrange the table so as to receive the average values of monthly increments by years and display them on the boxplot diagram. Fig. 1. Average price increment ranges by month, over 10 years. Author: Maxim Dmitrievsky

fxsaber 2019.12.06 23:20 #91

Maxim Kuznetsov:

PS/ distractedly - Fedoseev (if I'm not mistaken in his nickname) showed a completely crazy thing this week, he didn't even realise what it was...you have to think about it :-)

Let's stop the kindergarten practice of "I know, but I won't tell" announcements. Maxim really shares his methods, which cannot help but be impressed. The other practice is flud.

Maxim Kuznetsov:

PPS/ "nodal points of seasonal fluctuations together create a strictly linear structure" perhaps old Gunn was damn right.

Flooding, that's what it is.

Discussion of article "How Discussion of article "Applying Discussion of article "Practical

Maxim Kuznetsov 2019.12.06 23:22 #92

fxsaber:

Let's stop the kindergarten practice of announcing "I know, but I won't tell". Maxim really shares his methods, which cannot but be appreciated. Another practice is flooding.

see. PPS/ - I stand corrected :-) I prefer not to speak in riddles, but I don't always have time to say everything

Fast235 2019.12.06 23:22 #93

fxsaber:

Let's stop the kindergarten practice of announcing "I know, but I won't tell". Maxim really shares his methods, which cannot but be appreciated. Another practice is flooding.

Flooding, that's what it is.

Gentlemen Administrators, please ban in view of the New Year.

Maxim Kuznetsov 2019.12.06 23:23 #94

fxsaber:

Let's stop the kindergarten practice of announcing "I know, but I won't tell". Maxim really shares his methods, which cannot but be appreciated. Another practice is flooding.

Flooding, that's what it is.

I was going to send you a message, but you're already there

Maxim Dmitrievsky 2019.12.06 23:25 #95

Guys, don't stress already, the topic will be in the tops by the number of comments. The rest is business. Like suggestions on what else to look into. Before you do that, check with a psychologist for adequacy.

Discussion of article "Applying Discussion of article "The Discussion of article "Processing

Maxim Kuznetsov 2019.12.06 23:50 #96

@Dmitry Fedoseev posted a beautiful demo "a la slopes of averages", linear structures are quite clearly visible in it. Which in turn cannot be formed by chaotic speculations, but exclusively by cyclical things.

Even before that, on the other hand, while trying to clear up the rubble of hoaxes around Gann, I encountered stable pure-blooded sinusoids in the price series (in convolutions frankly speaking) - just physically a cycle as it is.
We need to properly and completely isolate seasonals because we know about them and it can be used, @Maxim Dmitrievsky is right to suggest a good method. Interpretation of the result (in the form of TC) is controversial perhaps, but that's another issue

Discussion of article "Visualizing Discussion of article "Color Discussion of article "MetaTrader

Igor Makanu 2019.12.07 07:52 #97

Maxim Dmitrievsky:
Guys, don't stress already, the topic will be in the tops by the number of comments.

it's not a problem, I always liked the style of presentation of your articles, there are no reprints of academic material, purely personal opinion and what exactly you have researched, I think I am not the only one who can see it and cause interest ;).

Maxim Dmitrievsky:
Before doing so, check with a psychologist for adequacy.

come on, the discussion was as adequate as possible and it is not necessary to achieve your rightness at any cost, though it is rather unexpected that the obvious fitting to the hypothesis was not visible to you..... perhaps this is a psychological effect?

Maxim Dmitrievsky:
For example, suggestions that you might want to look into

something from price normalisation methods, I would love to read even Box-Cox in your interpretation, or maybe the mentioned price logarithm? - in general in normalisation there is a "certain trick" which could allow if not to transfer TS from one symbol to another (it is probably not a solvable task), then at least it would allow to compare with some accuracy the work of TS on different symbols - there is some illusion that if you apply TS on another symbol and well to optimise, then you can apply it on another symbol, but imho, it is not correct it is another TS with other parameters and comparing the results of these TS is another self-deception.

Discussion of article "Practical Discussion of article "Implementing Discussion of article "The

Alexander_K 2019.12.07 07:54 #98

fxsaber:

The only difference is the application of the time filter.

Fact, the dumbest TS shows results that can't help but make you wonder. I am baffled and want to find a catch.

:))) Finally, mate, you are starting to understand something about the market. I rejoice and pray for all those who suffer in your person.

Maxim Dmitrievsky 2019.12.07 10:38 #99

Maxim Kuznetsov:

@Dmitry Fedoseev posted a beautiful demo "a la slopes of averages", linear structures are quite clearly visible in it. Which in turn cannot be formed by chaotic speculations, but only by cyclical things.

I didn't see anything of the sort from him and I didn't understand what linear structures are.

econometrics is silent about old Gann and slopes of averages, apparently because it is absolute esoteric nonsense

Discussion of article "LifeHack Discussion of article "Studying Discussion of article "The

Maxim Dmitrievsky 2019.12.07 11:49 #100

Igor Makanu:

any of the price normalisation methods, I would love to read even Box-Cox in your interpretation, or maybe the mentioned price logarithm? - in general in normalisation there is a "certain trick" which could allow if not to transfer TS from one symbol to another (it is probably not a solvable task), then at least it would allow to compare with some accuracy the work of TS on different symbols - there is some illusion that if you apply TS on another symbol and well zoptit, you can apply it on another symbol, but imho, it is not correct it is another TS with other parameters and comparing the results of these TS among themselves is another self-deception

Logarithmisation, Box-Cox, Fisher and Lambert transformations and other nonsense have no relation to financial markets.

only the spatio-temporal structure approved by Alexander and supported experimentally is interesting

Discussion of article "Exploring Seasonal Patterns of Financial Time Series with Boxplot" - page 10