Regret Minimization in Games with Incomplete Information (CFR)

This is the closest to literate programming I have seen in non book form. It is nice.

It seems to be written with PyLit [0]. PyLit lets you use LaTeX syntax to write math in comments [1]. Then it uses MathJax [2] to render the math symbols in HTML.

[0] https://github.com/slott56/PyLit-3/

[1] https://docutils.sourceforge.io/docs/ref/rst/directives.html...

[2] https://www.mathjax.org

Someone has created an excuse, a research topic that upon investigation yields information about how these people figured out how to play video games, at school, at work, professionally, without as much as an audience!

So probably the actual content falls a bit flat? But that’s a guess, which part didn’t you understand?

The twitter thread tries to explain the concepts without the math

https://twitter.com/labmlai/status/1407186002255380484

Here is a simple implementation I made a while back with plain-english comments rather than formulas, if that's helpful.

https://github.com/nemtiax/CFR/blob/main/cfr.py

I think it's a derivative of CFR, with a lot of optimizations and abstractions. They reduce the search space by grouping different card combinations together e.g. KK and QQ are treated same. They also limit the action space and only allow half-pot or full pot bets.

This is what I remember. I went through what they published briefly sometime back. So I could be wrong here.