Discussion of article "Exploring Seasonal Patterns of Financial Time Series with Boxplot" - page 7

 
This is the best article for the last 2 years on the forum. But only not only everyone is able to see the Truth in it...
 
Best article in the last 2 years on the forum. i'll enjoy it too, hopefully
 
fxsaber:
In the appendix the result of Optimisation.

BestInterval failed because it is not finalised for the opening price mode. That's why I took a frontal variant via GA. It took 78 seconds

optimization done in 1 minutes 18 seconds
shortest pass 0:00:00.030, longest pass 0:00:00.918, average pass 0:00:00.071


And the chart with OOS.

inPeriod=20||5||5||250||Y
inLow=-50||-500||10||100||Y
inHigh=30||-100||10||500||Y
inStartHour=0||0||1||23||Y
inCountHours=7||2||1||23||Y
inMaxAbsoluteDD=2000
inMinTrades=500


ZY Wild profit at midnight is embarrassing.

 
Alexander_K:
This is the best article for the last 2 years on the forum. But only not only everyone is able to see the Truth in it....

Let's not flatter the author, the article is not brilliant.

but it's the best in 3 years.

--

either he was writing with his sleeves up or he was trying to get into "mainstream".

I mean, just from the forum, I could have done better, and I didn't say everything I wanted to say.

 
Maxim Kuznetsov:

Let's not flatter the author - the article doesn't shine.

but in three years, it's the best.

--

or he was either writing too much or he was trying to get into mainstream.

just from the forum, I could have done better, and I didn't say everything I wanted to say.

The author's not bad, I like him, but if you look at you.

 
fxsaber:

BestInterval failed because it is not finalised for the opening price mode. So I took the frontal variant via GA. It took 78 seconds


And the chart with OOS.


ZY Wild profit at midnight is embarrassing.

Normich, so it can be improved.

Saber took a half-candy and made a candy out of a half-candy.

I'm not in condition today to perceive wholesale files, I'll look later )

 
Maxim Kuznetsov:

Let's not flatter the author - the article doesn't shine.

but in three years, it's the best.

--

or he was either writing too much or he was trying to get into mainstream.

just from the forum, I could have done better, and I didn't say everything I wanted to say.

I wanted to see how different timings worked. Saw that as sampling increases they become 50/50, any seasonal cycles. Was researching to improve dopamine (filter neuromodule, similar to Saber's bestinterval). AI bot takes an unrealistically long time to finish, several years. The article is logically finished, but I can think of ways to look at other cycles, I haven't thought about it yet.
 
Fast235:

Maxim it's not highly optimised and scientifically useless?

What exactly? The cycles are the real ones found. Optimised is found through an optimiser, in my understanding. Nothing has been optimised here
 
Maxim Dmitrievsky:
Which is what? The cycles are real found. Optimised is found through an optimiser, as I understand it. Nothing has been optimised here

I didn't have time to delete it, I don't want to go into it yet, it will be difficult for me.

 
Fast235:

I didn't have time to delete it, I don't want to go into it yet, it's hard for me.

Econometrics - search for seasonal patterns, Google. There is plenty of theory and practice too. If I suddenly need to :)