Tic-Tac-Toe

How the ML agent chose its move

The model predicted cell {{ mlExplanation.predicted_cell }} and chose cell {{ mlExplanation.selected_cell }}. The predicted cell was occupied, so it used the first available cell instead.

Test accuracy: {{ (mlExplanation.accuracy * 100).toFixed(1) }}%.

Cells are numbered left to right, top to bottom. X = 1, O = -1, empty = 0.

The highlighted route shows the rules the decision tree followed:

1. Question {{ index + 1 }} Is the value in cell {{ rule.cell }} ({{ rule.value }}) ≤ {{ rule.threshold }}?
   {{ rule.went_left ? 'Yes — left branch' : 'No — right branch' }}
2. Decision Play cell {{ mlExplanation.selected_cell }}