GARComp

• 

  Screenshot of example battle

Inspiration

I wanted a way of evaluating my Pokemon Go battle algorithms by playing against it

What it does

Not only does this work well for evaluating the strength of the various algorithms by playing against it, but the visualization aspect of this helped me find three seperate bugs in my model, just in the course of development

How we built it

It was made in GameZero.jl in the programming language Julia

Challenges we ran into

I had never developed a game before, so this was all a very new experience to me. Handling user input, asynchronous timing, and animations were all challenging aspects of putting this together.

Accomplishments that we're proud of

Not only did it play full games, but I am somewhat proud (and ashamed) that it also beat me playing randomly

What we learned

What's next for GARComp

• plugging into other existing algorithms such as the Nash solver and Monte Carlo Tree Search
• space above is left open for recording moves in PvPoke-style notation
• space below is left open for selecting pokemon without having to change the code

Built With

• gamezero.jl
• julia