blackjack_MDP_generalizer

Author: Michael. Combines adversarial lookahead and Markov Decision Processes to formulate a game of blackjack where card values can range from 1 to a chosen N (instead of 1 to 10 (Jack)) and winning values can be a window of some numbers L to U (instead of just 21).

How to use:

What it does:

A with-replacement generalized game of blackjack with an inclusive winning-bracket instead of a single value.

To run:

python bj.py <NCards> <LTarget> <UTarget> [--concise]

(--concise option will print out grid without details, as in Davis' specification.) Python 3.7. No external packages required.

For any discrepancies email mll469@nyu.edu. Thank you for checking it out.