bigtoga / Blackjackbot5000

This project analyzes data created by training an RL machine learning algorithm to play a simulation of blackjack. The website will allow the user to play blackjack, provide statistics for win probability of the current hand, and provide a recommendation for what move should be made next. There is further analysis documenting the optimal strategies based on the algorithm, win percentage tracked across 1 million simulations, and more.

Black Jack & Machine Learning

Application & Presentation Website:

https://blackjackbot5000.herokuapp.com/

---

Repositories:

https://github.com/Austin-Potts/Blackjackbot5000

https://github.com/jpicca/blackjackML

---

Group team members:

• Austin Potts
• Chris Armstrong
• Joey Picca
• Lance Westlake
• Ming Gao
• Shane Gatenby

---

Key Files:

1. Reinforcement Learning:
   • blackJackReinforcement_final_v1.ipynb
     • Important Instructions: Instructional README.md
     • Outputs:
       • Q_policy
       • win_pct
       • my_csv.csv
       • Q_values.csv
       • blackJackGameData.csv
       • blackJackGameTrackerData.csv
2. Tableau/Visualization:
   • BlackJack Outcome Viz.twb
   • RL_models_v3.twb
   • EDA Desktop View.twbx
3. Webpage/Presentation/Application:
   • html: /templates/index.html
   • css: /static/style.css
   • javascript: /static/controller.js
   • javascript: /static/chart.js

---

---

• Please see the Instructional README.md file located in the /final_RLandEDA/BlackJackRL/ directory for notes and instruction for using the blackJackReinforcement jupyter notebook file