This is excellent, creative and could also serve as a much less boring than pong project for starting reinforcement learning.
You don't need RL to compute the trajectory of a falling object.
51–60 of 189 posts
This is excellent, creative and could also serve as a much less boring than pong project for starting reinforcement learning.
One thing to consider to increase the difficulty is to make jumping add to the vertical velocity a la flappy bird, rather than fix it to a certain positive vertical velocity. Right now a reliable strategy is to let the ball fall to a certain angle below the paddle, then spam jump to get a fixed trajectory to the paddle.
26
Source code?
why does this have so many upvotes???????????????????????????