This week, we started our proof of concept. Our meeting revolved around delving deep into our learning algorithm, primarily via gaming examples. We started with a very simplistic understanding of Blackjack and soon found out that we had to incorporate a strategy that could optimize player wins. Questions bubbled up about the intricacies of gameplay strategies—when to hit when to stand—and how various factors, such as the cards played, influence these decisions.

The Tic-Tac-Toe example showed a good but simplistic implantation of our algorithm, acting as another lens to comprehend it’s nuances. While exploring these examples, an underlying theme became evident: identifying optimal strategies is pivotal, and nodes represent various strategies a player could employ.

Our project architecture also took center stage. We dissected our system’s architecture, pinpointing where our learning algorithm examples could mesh within the established framework. One question really revealed a flaw in our plan:  “Why aren’t the Red and Blue Teams mirror images of each other?” This led to enlightening discussions on game dynamics, code reusability, and the significance of consistent terminology.

Coming next week, we should expect to add strategies to our Blackjack Proof of Concept, update our system architecture, and lastly gear up for our peer presentations. Stay tuned as our project quickly develops!