Algorithmic ''centrifuge'' - page 16

New comment
 
Aleksei Stepanenko:

Peter, it is impossible to devise a system that works out all the waves because their amplitude and period are not known in advance.

There are three variants of price movement from the upper point to the lower one. As a result, the price has passed the same distance in all cases for the same time. Only the size of the pullbacks differed. In the first variant, there was nothing to catch on, and there were normal entry points in the third one. But how can I know about it beforehand? Neither Zig Zag, nor any of your creations in the future will solve this question. The future is unknown.

You are saying all the right things. Only we are looking at history. We are looking for the areas of ideal deals at history in order to automatically select the indicators for TS according to these areas.

I pointed out that exactly for this purpose, LA is not suitable, and therefore the PRINCIPLE itself is not suitable: "The ideal deal on the history is a deal with the maximum profitability". This principle is erroneous and will not give the right indicators.

 
So on history too, automatically, you can't fit the indicator to perfect points, because there are few parameters and many different cases.
 
Aleksei Stepanenko:

...

You tune the indicator to one type of price movement and fly by another type of movement. The only possibility is to tune to a more frequent type of price movement. And get a statistical advantage.

That's absolutely right too. So we need to distinguish between the patterns of price movement and select indicators for each.

 
Aleksei Stepanenko:
In this case, I propose to divide the history into sections with price movement patterns (signatures) and select indicators for them.

In this case, I propose to divide the history into sections according to the nature of the price movement (signatures) and select indicators accordingly.

 

Sometimes the pullbacks look self-contained, and sometimes, although decent, they stick strongly to the adjacent price movement line. Should we consider them an entry point or not? Sometimes you don't even know.


 
Реter Konow:

In this case, I propose to divide the history into sections according to the nature of the price movement (signatures) and select indicators accordingly.

Manual markup?
 
Aleksei Stepanenko:

Sometimes the pullbacks look self-contained, and sometimes, although decent, they stick strongly to the adjacent price movement line. Should we consider them an entry point or not? Sometimes you don't even know yourself.


You need to understand what is happening in the Market at that moment. Close pullbacks are "not normal" (I believe). It looks like a hammering. Stops are being pulled down. In any case, the second pattern of market behaviour is less stable than the first, which means the risk on the trade is higher. This should be taken into account and check what the indicators show.

 
Aleksei Stepanenko:
Manual marking?
No, automatic. It's not an easy task.
 
I agree. I have an opinion, not necessarily correct, that it is impossible to create a profitable system by being in the market all the time, i.e. turning over. Maybe I'm wrong, but I don't think so:) So, the system should wait for an entry point with a higher probability of making a profit than a loss. But to close the trade, you cannot wait for the same probability in the opposite direction. That is, we enter when we are sure of winning, and exit as soon as there are first doubts that something goes wrong. This is the difference between entry and exit probability estimates.
 
Реter Konow:
No, it's automatic. The task is not an easy one.
Pay attention to the code I wrote for you. There is only one parameter there - the minimum number of points between extrema. Add the minimum time between extrema there as well and use these two parameters to manage the size of desired waves. This is better than Zig-Zag.
1...91011121314151617181920212223
New comment
Reason: