Ethylene margin collar simulation model

I set up the simulation model. The margin follows a mean-reverting process.

I seperarted the data into two category, margin <0.04 and margin <0.04.
Then I estimate the mean reverting speed seperately for these two data sets.
I got higher mean reverting speed than that I estimated using the whole data
set.

The high MR speed surpresses the probability at high payout side.
Since the MR speed is sensitive to where I divide the data, so Bob will run
a few senarios.

I put the overal settlement cap and floor into the montly premium
calculation, so the
the result in E18 on the summary page is the ultimate answer to the deal
pricing.
I also calculate the undiscounted payout distribution and overall collar
worth.

Relax the overall cap and floor will have a direct comparison with the spread
option
approach that Bob and Lee set up.

Look like we got a reasonable model.

Stinson:
I'd like to have you check my set up for the simulation model.

Lee and Douglas:
You can play with the model, and let me know what do you think.

Bob:
We need run different price curve senarios using the simulation model.
plus different MR speed.

Zimin