Pokerbot Part 2 - Game ON
Let's continue with the Pokerbot! After the first tests were quite successful and I could fight up my account as a beginner from 100 Satoshi (~0.1 Rappen) to 2000 several times, one thing was clear: Playing poker is not so easy. The hard-won capital can be destroyed within a few rash moves.
Well then, let's move on. From a relatively dull predictor to an intelligently acting system that learns poker and not just bluntly evaluates probabilities. but how?
First, the AI must be able to participate in a game and play. Very much and quickly can play. Unfortunately there were only half-baked open source solutions on the internet. So I had to cook my own soup. But that's how you learn.
After all, you can customize the whole thing to your requirements. What the processors and memory give, I can hereby play in parallel up to 9 AI's per table. All of them will learn independently and compete against each other in a league. But that's all still up in the air. First all bugs have to be removed.