What to feed to the input of the neural network? Your ideas... - page 10
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
And just above about finding the high and low, what was that?
Let's say there is only one tick, what should we do with it? What can be done with it at all?
You must have missed this post: #80
The current point is where you start. Find the current minimum and maximum. Naturally, they will coincide with Bid and Ask. 1) The 2nd tick will come, there will be a new min. or max.
In this way the current min. and max. are determined. 2) And after closing the order, everything starts from zero, i.e. from the current price, and what was earlier is not important.
To describe all this, it is necessary to make screenshots, draw charts and write several pages of text, as well as video.
I just wanted to say that you can get good results without analysing history, without MO and NS and any external indicators.
1) And here - there are already two ticks, min and max is determined on the basis of two ticks.... That's the history analysis.
2) Something from the field of "from fence to lunch".
1) And so - already two ticks, min and max are determined based on two ticks.... So much for analysing history.
2) Something from the field of "from fence to lunch".
If so, we can consider that when the very first tick comes, it is already history. And when we determine its spread, we analyse the history.
This is ridiculous :)
Even on a tester, you can get the history for at least one day (for the past day).
If so, we can consider that when the very first tick comes, it is already history. And when we determine its spread, we analyse the history.
This is ridiculous :)
Even on a tester, you can get the history at least for one day.
In a purely chronological sense, the current tick is also history, because it is real only at the moment of its arrival.
From the point of view of analysis, we can make an assumption and consider the last tick as the present, but everything that was before it will be history.
Therefore, even if only two ticks are analysed, it is already an analysis of history.
Spread is from another field. You should not cheat. Nobody makes decisions only on the basis of the spread.
The graph with the price is rubbish?
Yes, it is rubbish, because it is impossible to get anything but a retrained network.
More precisely, without emotions, a graph is a time series that is non-stationary. Non-stationary time series cannot be modelled, especially for such a lousy thing as neural networks.
In a purely chronological sense, the current tic is also history, because the present it represents solely at the moment of its arrival.
From the point of view of analysis, we can make an assumption and consider the last tic to be the present, but everything that was before it will be history.
Therefore, even if only two ticks are analysed, it is already an analysis of history.
Spread is from another field. You should not cheat. Nobody makes decisions based on the spread alone.
Why only spread ? We havecurrent price, max and min, spread size, speed of price change(not time), which are defined explicitly.
They are used to open an order and not only them. And there is no need to guess where the price will go.
But the question of the topic was "What to input to the neural network?".
Why only spread ? We havecurrent price, max and min, spread size, speed of price change(not time), which are defined explicitly.
:) They are used to open an order and not only them. And there is no need to guess where the price will go.
But the question of the topic was "What to input to the neural network?".
Spread, because it was spread that was used here recently by someone for a special argumentation. Max and min are at least two ticks, so there is a history.
To determine the speed you need two measurements, so also history is used.
:) no comments))) everything is said with a smiley face.
To feed the neural network input - data.
If so, we can consider that when the very first tick comes, it is already history. And when we determine its spread, we analyse the history.
This is ridiculous :)
Even on a tester, you can get the history for at least one day (for the past day).
A new tick is already a history.
At this price you will not open an order with a robot. On a new tick there may be a completely different price. For example, a significant gap)
In terms of trading, it does not matter which TF to choose. On any TF, the price behaviour repeats itself in the same way. Wavelike. Only the influence of the spread on the result will depend on the selected time interval. On the tick time interval in the first place.
And you want to be considered as knowledgeable developers ?