Basketball Simulation with RL
A small probability-update experiment that visualizes a shot in Pygame and steers trials toward a 48-degree target angle.
Purpose
This project pairs a physics-based basketball shot animation with an early experiment that selects an angle through repeated trials. The README gives little detail, so this description follows the source. Despite “reinforcement learning” in the title, it is not a general RL algorithm or full environment; it explores the idea in a small example.
Simulation
The Pygame component calculates a ball trajectory from speed, gravity, and angle, then draws a parabola. A separate Python script samples angles from 1 to 89 degrees using a probability distribution. It sets 48 degrees as the target, updates probabilities over successive trials, and saves the result to a pickle file.
Limits
This is an educational experiment in how choices and feedback can change probabilities. It is not a trained policy, validated model of real basketball, or benchmark of advanced RL methods.