Skip to content
Safely-rounded Confidence Sequence Method
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
attic
.gitignore
LICENSE
README.md
confidence-sequence-method.asd
csm.c
csm.h
csm.lisp
csm.py

README.md

Safely-rounded Confidence Sequence Method

An implementation of Ding, Gandy, and Hahn's Confidence Sequence Method for dynamic termination of Binomial tests on Monte Carlo simulations, as well as code to compute Bayesian credible intervals on the underlying success rate.

The code in C, Common Lisp, and Python should give bitwise identical results, and the results should always be safely (conservatively) rounded; the implementation of CSM in floating point should not cause any false positive. N.B., The reason I care about round-off errors isn't that I expect them to make a big difference, but simply that I want peace of mind, and not have to wonder if weird numerical issues are causing whatever my system reports.

See this blog post for more information.

I'd particularly like to receive suggestions or pull requests to improve the usability of the tools. I'm thinking improvements to the report format, graphical reports, and CLIs tools that are easier to insert in testing pipelines or in ad hoc shell scripts.

You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.