I want to simulate paths for a commodity price. I use the historic data in the following way:
$X_t$ is the price.
$\ln\left(\frac{X_t}{X_{t-1}}\right)$ is the daily return.
I calculate the slope of the regression for six prices.
I calculate the slope of the slopes found in step 3.
For the simulation I use the Ornstein-Uhlebeck equation How can I produce smoother paths in order to have a robust model, i.e. I want to calculate the slope of every six observations and then again the slope of the slopes?
