semi-stable foundational update, but a big step in the direction I want to go.

I thought of a method to learn from bad placements, lowering the number of retries as more bad placements are detected.

I'd call it machine learning, but it isn't a traditional ML model.

Also added /rtp parameters to generate a random location based on inputs