Alpha-beta pruning intuition
An optimization on minimax that **skips branches** that cannot affect the final decision—same answer, less work (under standard assumptions).
What it is
An optimization on minimax that skips branches that cannot affect the final decision—same answer, less work (under standard assumptions).
Why it matters
Makes adversarial search practical deeper in the tree—classic CS teaching bridge to modern game AI.
How it works (plain)
Track best already-guaranteed scores for both sides; abandon lines that are irrelevant. Move ordering affects how much you prune.
Try it
On a tiny tic-tac-toe line, mark a branch you’d skip once a better line is known.
Myths
- ⚠️ Myth: Pruning changes the optimal value.
- ✓ Reality: Correct alpha-beta preserves minimax value.
Sources
- Course 06 adversarial-search-and-minimax
- CS50 AI / AIMA games chapters