Finished Privacy Simulation Results

At the previous meeting I gave a preview of some of the early results from
simulating the impact of adding random codepoints to requests on privacy.
I've now put together a writeup on the finished results which I've
published to our analysis github repo. You can find that here:

https://github.com/w3c/PFE-analysis/blob/main/results/noise_simulations.md

One of the main changes from the early results I presented is the addition
of a variable length noise. In this approach the amount of codepoints added
is based on the number of codepoints in the request. The less codepoints
there are, the more noise is added. This method further improved
effectiveness.

Additionally the results now include data for additional scripts: Arabic,
Devanagari, Chinese, and Korean.