Add LinearLeiosTuner.html

[nfrisby]

An interactive calculator/plotter for ballparking the consequences of various L values (where L_diff is the assumed bulk).

Three general consequences are at-least-somewhat illustrated:

• Expected duration between EBs succeeding and also the variability in that duration.
• Expected resulting throughput capacity (no variability here) AND how bad the ratio of network usage could be compared to that capacity (the paragraphs under the capacity chart).
• Pull-demand on the Mempools to refill.

[nfrisby]

Rendered:

• The tool https://raw.githack.com/input-output-hk/ouroboros-leios/nfrisby/ldiff-tuner-html/docs/targeted-design-investigations/LinearLeiosTuner/LinearLeiosTuner.html
• The Markdown spec file that I bootstrapped Claude from https://github.com/input-output-hk/ouroboros-leios/blob/nfrisby/ldiff-tuner-html/docs/targeted-design-investigations/LinearLeiosTuner/LinearLeiosTuner-README.md

[ch1bo]

The probability distribution of possible certifications being a cut-off binomial makes sense, I guess?

I don't understand the max capacity calculation though.

Shouldn't it be just a multiplication with the expected value for P_certified? i.e. something like 48.8% for L=14

How does this assessment compare against the one used in the cost estimate and cost estimator tool

• https://leios.cardano-scaling.org/cost-estimator/
• Update the cost-estimate to CIP-164 protocol #872

Note that the cost estimator is flipped in it's parameterization and takes a resulting throughput (= your capacity?)