SPL-T

A Python implementation of the iOS game SPL-T. This is a work in progress, it's not finished yet.

See the game's website for details.

The eventual goal of this project is to find out the moves to play the perfect game of SPL-T, achieving the highest possible score. I plan to accomplish this either by:

1. Building a game tree that shows all possible moves from a given state, looking ahead only 3-4 moves. Then, choosing the optimal path at each turn
2. Using some kind of learning algorithm to train itself to play SPL-T. I'm not sure how realistic this is because of my lack of knowledge in this field.

I'm working on this very intermittently, since it's pretty difficult logic-wise and I've got projects I want to work on more.

TODO

• Game Logic
  • Splitting tiles
  • Forming point tiles
  • Falling tiles
  • New tiles falling from ceiling
• Game interface
  • Leverage Pythonista's scene module to build a touch interface for testing.
• Game solution (either Minimax or Machine Learning)
  • Creating a method for assigning a numerical value to an arbitrary game state
  • Building game tree
  • Finding optimal moves