Coding Poker Hand Evaluation In Python

This video covers how hand evaluation is done in poker programs and the basic techniques for learning the attributes of a hand which will be used for clustering poker hands together. The previous video was an overview of some of the challenges of coding poker AI bots.
Code a Poker Bot Playlist: • Code a Poker Bot
suffe.cool/poker/evaluator.html
github.com/HenryRLee/PokerHan...
en.wikipedia.org/wiki/Lookup_...
en.wikipedia.org/wiki/Perfect...
github.com/Fossana/discounted...
www.cs.cmu.edu/~sandholm/pote...
en.wikipedia.org/wiki/K-means...
web.archive.org/web/201501130...

[@justingolden87]

So I'm probably a little out of it here, but the goal is to be super performant so you can calculate everything and generate a 52 choose 7 size lookup table? Because all the checks for what suit and such would be super fast even if done poorly in a slow lang, but the bitwise stuff makes it super fast because you're iterating many, many many times, right?

[@stackwild1803]

I put a few things into this vid that are prereqs for later videos. Fast hand evaluation is its own problem, mostly for any simulation that needs to run millions of hands quickly. I brought it up because a lookup table may be needed to check what bucket a hand belongs in for solving the abstracted game, depending on the bucketing technique. The bitwise stuff isn't really an optimization, it's just how most poker codebases represent cards and hands because it's simple once you get used to it.