00:16 - blackjack rule applied in this project
05:50 - database structure using XML
08:01 - base theory
27:40 - result chart
28:49 - result by different card counting setting (when greater/less than 10)
my machine learning showed, when card counting number is greater than 10, you lose about 0.5% of original bet(100%) on every game.
In other hand, when card counting number is less than -10, you earn about 3% of original bet on every game, which indicates if you play only best, the lower card counting, higher winning rate.
you can play blackjack against my ai.
project:
https://github.com/justdoit7724/Blackjack_BasicMachineLearning
