The reason to do this is reducing the amount of calculations the script goes through, to keep it within the limits of PineScript.
The iterates will split up momentum lenghts from the largest available 5400 to the smallest usefull approx 30
This can be done Linear, dividing the steps equally, Logarithmic or Exponential.
Favor goes out to Exponential, as longer momentum lenghts equal each other more.
In the above image you see momentum indexed from -1 to 1 colored red to green, split exponentially in rows of each iterate lenght.
Short motivation for fluid transformations vs arbitrary settings
Add formula and image. // Should be scaled to timeframe
Add formula and image // Focus Should be adjustable
A factor (a) limited from -1 to 1 will be applied on the iterates with following formula:
A note here is that a percental based (a) would have its limits from -1 to ∞ If this is what should be applied, the math of the formula needs to be changed accordingly.
This transformation could be applied flat to all iterates. But this could mean an unjust bias is applied to the smaller momentum lenghts.
A simple proposition to improve this, is applying a decay curve to the bias factor. A complex curve could later be calculated to deviate from this decay curve based on lenghty backtesting and optimization results with ML.
An example of the decay curve, simple and measured.
Add formula and image, Use-case examples
Use-case examples
Use-case examples
Add image
Add image
Add image + Use-case
Add image + use-case
Add image + use case