
Goodhart's Law as Non-Commutativity: A Mathematical Framework for Metric Tradeoffs
Original: Goodhart's Law Is a Math Problem - And We Can Prove It With 10 Lines of Python
Short summary
The article applies quantum mechanics' non-commutativity and uncertainty principles to organizational metrics, arguing that Goodhart's Law is a mathematical property rather than a human behavior problem. Pairs like velocity vs. quality or hiring speed vs. team cohesion don't commute—optimizing one first irreversibly alters the system, making simultaneous optimization impossible. Practical recommendations include rotating optimization focus by quarter, measuring the 'commutator' between metrics before deploying new KPIs, and using composite metrics that encode tradeoffs directly.
- •Goodhart's Law stems from mathematical non-commutativity of metric pairs, not just human gaming
- •Non-commutative metrics like velocity vs. quality cannot be simultaneously optimized—order matters
- •Practical fixes: rotate optimization basis quarterly, estimate metric coupling, use composite metrics
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



