Game setup
Choose a mode, round count, and difficulty before starting.
Standard
Player 1
Round 1/10
25
Decision map
Category Drag the centre token into a zone, or use the decision buttons.
Scenario
Scenario title
Available evidence:
Sample size
Response target
Cost to build
Consequence of error
Data quality
Rule stability
Select up to four reasons.
- Recommended approach
- Why it fits
- Useful alternative
- Practical implementation
Game complete
Decision performance report
Plotly performance dashboard
Charts update from your completed rounds.Achievements
Review and learning plan
Decision guide
Use these signals without treating them as absolute rules.
Machine learningUseful for changing patterns, prediction, perception, personalisation, and large labelled datasets.
Fixed rulesBest when logic is explicit, stable, exact, auditable, and inexpensive to maintain.
Human judgmentEssential for empathy, legal discretion, rare cases, contested evidence, and accountable value choices.
Hybrid approachCombines predictions, hard constraints, uncertainty thresholds, expert review, and appeal mechanisms.
Important: A technically accurate model may still be unsuitable when fairness, privacy, safety, accountability, or contestability are weak.