- I peer reviewed one of my peers.
- I changed the data structure of the board from an array of arrays to a single long array.
- 2x performance gain
- I cleaned the code of the minimax algorithm.
- I implemented alpha-beta pruning
- 35x performance gain!
The 35x performance gain of implementing alpha-beta pruning was a surprise. I had previously seen some article that featured a formula for the performance gain of alpha-beta pruning. Using that formula I got a result that in the best case scenario I should get a 10x gain. Now I'm wondering why did I get a performance so much higher?
- Optimize the minimax algorithm and the program in general with many tricks I have thought of.
- Implement an algorithm for choosing the starting position.
- Right now the program assumes that you provide it with a board with the starting stacks in place.
- Write testing and implementation reports.
- Implement the player vs AI mode.
22h