Goodhart's Law
The principle that once a measure becomes a target, it ceases to be a good measure, because people optimize the metric rather than the intended outcome.
As defined in Jennifer Pahlka
Other definitions
The principle that when a measure becomes a target it ceases to be a good measure, because people optimize the proxy rather than the underlying goal.
Alison Gopnik
Where it comes up · 8
- Caleb Watney - America's Innovation Engine1 claim
- Compounding Through Operational Excellence1 claim
- #58 John Bogle: Enough: True Measures of Money, Business, and Life1 claim
- Alison Gopnik1 claim
- Jennifer Pahlka1 claim
- The Dark Side of Conscious Ai | Daniel Schmachtenberger1 claim
- Terence Tao1 claim
- 10 year anniversary presentation -- Russell Napier on lessons from financial history1 claim
Most important claims · 8 of 8Search all claims for “Goodhart's Law” →
- Goodhart's Law — that once a measure becomes a target it ceases to be a useful measure — appears everywhere in government; the 1990s reinventing-government and new public management reliance on setting and following goals had real value but eroded as people met targets without meeting the intended outcomes, which is why a faster-iterating digital approach that can revise targets is needed.Jennifer Pahlka
- A listener raises the possibility that when governments target any specific statistical number (credit growth, M2, inflation), the system bypasses that number and it becomes irrelevant (Goodhart's Law), suggesting Napier should add this to his lessons.10 year anniversary presentation -- Russell Napier on lessons from financial history
- Goodhart's Law states that as soon as a metric becomes an optimization target, it ceases to be a good metric for what it was supposed to measure because people will find ways to fulfill the metric that no longer provide the intended benefit or actively produce harmsThe Dark Side of Conscious Ai | Daniel Schmachtenberger
- Daniel Yankelovich's four steps describe how over-reliance on measurement goes wrong: measure what is easily measured, disregard what cannot be measured, presume the unmeasured is unimportant, then conclude the unmeasured does not exist, which is suicide.#58 John Bogle: Enough: True Measures of Money, Business, and Life
- Goodhart's Law states that when a statistic is used to incentivize performance, it becomes gamed and is no longer a useful measure—relevant to why academics are wary of using contribution metrics (like lines of code in Lean projects) for tenure review.Terence Tao
- China is plausibly good at catching up to the technological frontier but is much less likely to actively push the frontier outward, partly because Goodhart's-Law-style metric-chasing (e.g., mandating 50 'world-class' AI textbooks) misses what actually makes science productive.Caleb Watney - America's Innovation Engine
- Current schooling is a clear illustration of Goodhart's law: because children are highly motivated to become skilled, they get extremely good at the proxy—taking tests and 'going to school'—which we assume correlates with the broad adult capacities we actually want, but the proxy ceases to track the target, leaving students unable to design a new experiment or do creative independent work.Alison Gopnik
- Although Luca is a fan of data, almost any data can be manipulated, and metrics like revenue per employee can be inflated by simply not hiring enough people.Compounding Through Operational Excellence