Game setup
Choose a scenario, challenge mode, difficulty, and display detail.Current record
Round status
Round 1Score0points
Timer60seconds
Streak0correct
Accuracy0%branches
Keyboard controls
← left branch · → right branch
H hint · N new round
+ zoom in · − zoom out
Decision path
0 stepsStart the round to reveal each decision.
Prediction explanation
Complete the path to reveal the prediction and explanation.
Counterfactual laboratory
Change features and watch the path respond.Threshold playground
Tune numerical splits and inspect model changes.Import and custom records
Load a CSV record or a compatible JSON tree.Use headers on row one and values on row two.
Expected keys: name, type, features, nodes, records.
Analytics dashboard
Plotly charts update after every completed path.Learning dashboard
Recommended practice
Complete your first path to receive a recommendation.
Achievements
How to play and learn
1. Inspect the record
Read every feature value. Compare it with the highlighted node condition before choosing a branch.
2. Follow the rules
Choose left when the condition is true. Choose right when the condition is false.
3. Explain the leaf
Describe the rules that were satisfied and connect them to the final class, probability, or regression value.
Decision-tree terms
The root is the first rule. Internal nodes contain later rules. Leaves contain predictions. Impurity measures how mixed a node is. Information gain measures how much a split improves separation.
Interpretation limits
A path explains one tree’s local reasoning. It does not prove causality, guarantee fairness, or show relationships learned by models outside that tree.