Wordle Decision Trees: Planning a Whole Game in Advance
A decision tree is a map of every result a guess can produce and the best response to each. Computer solvers build complete trees; humans build partial ones. This article shows how to think in branches so that your second and third guesses are decided before you ever see the clue, dramatically lowering the variance of your results.
The anatomy of a tree
A decision tree starts at your opener. Each branch corresponds to a possible color pattern. Under each branch sits the response you have pre-selected for that pattern, then its own sub-branches, and so on. The tree encodes a policy: a complete rule for what to do in every situation, rather than a single line of play.
The value of a tree is consistency. When your responses are planned, you remove in-the-moment errors and emotional decisions. You also build pattern memory faster, because you repeatedly traverse the same structure and learn which branches lead to trouble.
Building a human-scale tree
You cannot memorize the full tree for a 2,300-word answer list, but you can build the trunk. Fix one opener. Plan a default second guess for the most common outcome — typically 'all grey,' which happens often with a strong opener. Then plan responses to the two or three highest-frequency partial patterns. This covers the majority of your games with a rehearsed line.
For the long tail of rare patterns, fall back on general principles: maximize information while candidates are many, switch to probability when they are few. The trunk handles the common cases; principles handle the rest.
Reading a tree for weaknesses
Once you have a trunk, audit it. Look for branches where your planned response leaves a large, undifferentiated candidate group — these are your variance hotspots. Replace the response with a higher-information word, even one that cannot be the answer, to flatten that bucket.
Auditing your own tree after a difficult game is the fastest route to improvement. The loss almost always traces to a single weak branch, and fixing that branch upgrades every future game that passes through it.
Key takeaways
- A decision tree encodes a complete policy, not a single line of play.
- Memorize the trunk: one opener and responses to the most common patterns.
- Audit branches that leave large candidate groups and replace weak responses.