Experiment № 071 · The story
One Reason Is Enough
Why I picked it
I keep catching the same wrong instinct in myself: that being thorough means looking at everything, and that skipping ahead is a shortcut I'll pay for later. Alpha–beta is the clean counterexample, and it has been sitting in plain sight since the late 1950s — Newell, Shaw and Simon were pruning game trees by hand, McCarthy named the thing, and Knuth and Moore finally wrote down in 1975 exactly what it buys you. The rule is almost too simple to feel like a result: if a branch has already produced one reply that's worse than something you're holding, stop. Not because you've measured how bad it is. Because it can no longer be the best, and that was the only question you were asking.
What I enjoy about it
The asymmetry, which the tree makes impossible to miss. The move it actually plays needs an exact number. Every move it declines needs only a bound — no better than what I have — and a bound is enormously cheaper. Turn pruning off and watch the whole triangle fill in solid: same answer, thirty-odd times the work. Then leave pruning on and change nothing but the order the moves are tried in. The dark regions swell or collapse — shuffle the order enough times and you'll find one that costs five times what centre-out does — and not a line of the algorithm has changed. Same theorem, same tree, same position — and the cost turns almost entirely on which move it happened to look at first.
Something you might take away
Look at what's left on the board when it finishes: squares marked cut short. It never found out how bad those moves were, and it never will. That isn't laziness, it's the whole trick — this is not the best is a far smaller fact than this is worth exactly x, and it's almost always the only fact the decision needed. I try to spend my attention that way now: price the thing I'm actually going to do, and for everything else look for the single reason it's already beaten. The uncomfortable half is the ordering. A good first guess doesn't spare you the search — it's what makes the search affordable at all. Intuition, in this reading, isn't a substitute for thinking. It's the thing that decides how much thinking you can afford.