Goodhart Divergence Curve (24 Periods)Software Dev
When high-stakes KPI pressure is applied without intrinsic learning feedback, reported velocity surges while actual purpose quality collapses into metric gaming.
Team Behavior Heatmap (Multi-Agent Dynamic)
Genuine Value
Metric Gaming
Goal Displacement
Nodes represent staff members shifting from genuine task completion to local optimization (e.g. task splitting, cherry-picking easy tickets, suppressing long-term fixes).
Goodhart's Law & KPI Gaming Simulator Diagnostic Proof
Period 24 Completed State
KPI Reported Velocity
94.2
Target Metric Inflation
True Purpose Quality
38.5
Actual Delivered Value
Goodhart Divergence Index
55.7
Metric vs Reality Gap
Gaming Prevalence Rate
68.0%
Staff Adopting Gaming Strategy
Primary Dominant Gaming Strategy
Micro-task Splitting & Tech Debt Accumulation
High target pressure incentivizes team members to prioritize quantifiable easy wins while leaving complex systemic problems unresolved.
Counterfactual Strategy Analysis: KPI Target vs Metric-as-Learning
A: KPI as Incentive Target (Current)
Reported Velocity:94.2
True Quality:38.5
Divergence Index:55.7
Gaming Prevalence:68.0%
B: Metric as Learning Feedback (Mogi Thesis)
Reported Velocity:76.4
True Quality:82.1
Divergence Index:5.7
Gaming Prevalence:8.2%
Kenichiro Mogi's Insight: When metrics are demoted from high-stakes targets to reflective learning tools, true organizational capability expands and metric divergence drops to near zero.
Organizational Diagnostic Brief & Health Report
This report captures the current simulation state, calculating Goodhart's Law metric divergence and behavioral corruption rates across 24 simulated iterations.