Recursive Stochastic

The Self Referencing Stochastic Oscillator

The stochastic oscillator bring values in range of (0,100). This process is called Feature scaling or Unity-Based Normalization

When a function use recursion you can highlights cycles or create smoother results depending on various factors, this is the goal of a recursive stochastic.
For example : k = s(alpha*st+(1-alpha)*nz(k[1])) where st is the target source.

Using inputs with different scale level can modify the result of the indicator depending on which instrument it is applied, therefore the input must be normalized, here the price is first passed through a stochastic, then this result is used for the recursion.

In order to control the level of the recursion, weights are distributed using the alpha parameter. This parameter is in a range of (0,1), if alpha = 1, then the indicator act as a normal stochastic oscillator, if alpha = 0, then the indicator return na since the initial value for k[1] = 0. The smaller the alpha parameter, the lower the correlation between the price and the indicator, but the indicator will look more periodic.

Comparison

Recursive Stochastic oscillator with alpha = 0.1 and bellow a classic oscillator (alpha = 1)



The use of recursion can both smooth the result and make it more reactive as well.

Filter As Source

It is possible to stabilize the indicator and make it less affected by outliers using a filter as input.



Lower alpha can be used in order to recover some reactivity, this will also lead to more periodic results (which are not inevitably correlated with price)

Hope you enjoy

For any questions/demands feel free to pm me, i would be happy to help you