Combining trend and flat strategies into one TS = grail? - page 14

 
ChachaGames:

I'm already leaning towards statistics too, but then all the beauty of the idea is lost, you get a fit....

And ugly is wrong!

It doesn't matter, as long as there is a profit.
 
paukas:
It doesn't matter as long as there is a profit.
Well true too, although statistics are a pseudoscience :)
 
ChachaGames:
Yes, automatically, for clarity and testing, I even draw them visually, naturally along the extrema, of which the so-called "fractals" and then the secret, six months of thinking, trying, testing, and programming.
Excellent, you can check it elementary.Test the EA over the entire history - Dabbing inside the channel, breakdowns, rebound, if successful, it means that the algorithm meets the criteria of the history
 
Techno:
Excellent, it is elementary to check it.Test the EA on the whole history - Dangling inside the channel, breakdowns, rebound, if successful, then the algorithm meets the criteria of the history
So it is, but the indicator does not make money - yes, it's cool, but the Expert Advisor is losing everything on the flat that it has earned on the trends
 
Roman.:
C-4:
I propose to use a simple effect of volumes accumulation as a flat criterion (if you need to identify it that much). If the accumulated volume exceeds a certain amount of standard deviation over a given time interval, then we are in a flat.
If possible, more details, if it is not a "secret of Madrid", because it is not clear... the topic is interesting.

There is no secret here. Accumulated volumes are also called "Market Profile". This concept is used by a whole caste of traders. You can read about how it works here: https://www.mql5.com/ru/articles/17. You may be interested in the effect of volumes accumulation during the consolidation phase. The distribution peak is the median of the flat, the limiting deviations of volumes from the median volume will be indicative of its borders. Here is an example (figure taken from Dmitry's article):


Z.U. The same principle can be used to determine the trend. If volumes cannot be accumulated in sufficient quantities at a given interval of time, we have a trend.

 
C-4:

There is no secret here. Accumulated volumes are also called "Market Profile". This concept is used by a whole caste of traders. You can read about how it works here: https://www.mql5.com/ru/articles/17. You may be interested in the effect of volumes accumulation during the consolidation phase. The distribution peak is the median of the flat. The limit deviations of volumes from the median volume will be indicative of its borders. Here is an example (figure taken from Dmitry's article):


Oops! beautiful, let's study
 
ChachaGames:
So it is, only the indicator does not make money - yes, it's cool, but the Expert Advisor is losing everything on the flat that it has earned on the trends
Well, a flat is a wobbling between the borders of the channel, let the EA switch accordingly...
 
C-4:

There is no secret here. Accumulated volumes are also called "Market Profile". This concept is used by a whole caste of traders. You can read about how it works here: https://www.mql5.com/ru/articles/17. You may be interested in the effect of volumes accumulation during the consolidation phase. The distribution peak is the median of the flat, the limiting deviations of volumes from the median volume will be indicative of its borders. Here is an example (figure taken from Dmitry's article):


Z.U. The same principle can be used to determine the trend. If volumes cannot be accumulated in sufficient quantities over a given period of time, then we have a trend.


Sens, I'll take a look, thank you.
 
Techno:
Well, a flat is a wiggle between channel boundaries, let the EA switch accordingly...
The formula determines the change of state, during a trend it can also wiggle between the boundaries.... although.... maybe that's an idea! check it out
 
ChachaGames:
The formula determines the state change, during a trend it can also dangle between the boundaries.... although.... maybe that's an idea! I'll have to check it out.
ok, degree of slope, if big trend, almost horizon - flat. You could prooptimise the transition boundaries on the history.