Inferring Percentiles for Lognormal Distributions in Reserves Estimation

Calculating a percentile X, based on two other percentiles for a lognormal distribution, sounds familiar❓

That’s right, in reserves estimation, when we have any two rough estimates of high / low / medium reserves values, and you need to estimate the unknown one (if you don’t have a full probabilistic run at hand of course).

Here is a deep dive on calculating it in #Python using only #scipy and #numpy. Beyond the reserves estimation application itself, it is an interesting topic to understand a bit more about distributions and revisit some university topics 😉

Also, here is the link to the notebook version: https://lnkd.in/esPhRWaX

• psst, you can run it in your browser if you use #codespaces without installing anything, the repo already has the #devcontainer in it.

Have you done this before, and if so, what did you use to calculate the unknown percentile? Let me know in the comments 👇