How the Nerdle Solver Works

When you play a guess, every possible answer would produce some specific coloring — greens, purples, blacks. That means a guess partitions the suspect list: every answer falls into exactly one feedback-pattern bucket, and once the real colors appear, every answer outside your bucket is eliminated instantly.

So a guess is exactly as good as the buckets it creates. Many small buckets mean that whatever feedback you get, few suspects survive. A few big buckets mean you can get unlucky and learn almost nothing. Our best opener, 48-32=16, splits the field into 1,689 buckets. A repeat-heavy guess like 11+11=22 manages only 415 — it asks the same question about the same few characters twice.

To rank guesses, the solver asks: if the answer is chosen at random, how many suspects will still be alive after I see the feedback? Big buckets get weighted by how likely you are to land in them (a bucket holding 100 answers is hit 100 times as often as one holding 1), which is why lopsided guesses score badly.

Same rules, same 17,723 suspects — but the best opener leaves you a to-do list of ~31 while the worst of these leaves ~319, ten times as many. That one choice is often the difference between solving in 3 and sweating on 5. The full ranking lives in the strategy guide.

Scoring one guess is cheap; scoring every legal guess against every surviving answer at every turn is not. So we did it once, offline, and saved the result: a decision tree whose root is 48-32=16. The tree has one branch per feedback pattern, each leading to the best reply, and so on until every branch ends in the answer.

Two details matter. First, the tree optimizes expected total guesses, not just the flashiest first split. That's why 48-32=16 beats the classic division openers, and why the tree sometimes plays probe guesses that can't be the answer (explained in the strategy guide). Second, when you use the solver, it isn't computing anything heavy — it walks this precomputed tree, which is why suggestions are instant.