KnowraGoodhart's lawLinked fromLinked fromThe 29 pages that link to Goodhart's law, each with the reason it gives.All 29Related 24Compared with 5Regression analysisRelated: Regression-based metrics can invite incentives that undermine what they were meant to measure.AI alignmentRelated: It explains why optimizing a human-chosen metric can undermine the value that metric was meant to represent.Lucas critiqueRelated: Both ideas warn that intervention can change the behavior behind a useful statistical relationship.GamificationRelated: When a score becomes the goal, participants may optimize it instead of the intended behavior.Mathematical modelingRelated: Optimizing a model’s proxy objective can distort the real-world goal it was meant to serve.Portfolio optimizationRelated: Optimizing a narrow historical metric can encourage portfolios that exploit the metric rather than achieve durable goals.BibliometricsRelated: It explains how incentives tied to bibliometric indicators can distort research behavior.Educational measurementRelated: Schools may improve measured outcomes while neglecting the underlying learning.Financial return on investmentRelated: Using ROI as a rigid target can encourage choices that improve the metric but harm broader goals.IncentiveRelated: Tying rewards to a metric can prompt behavior that raises the score while undermining the goal.Journal impact factorRelated: Treating impact factor as a target can encourage behavior that raises the score without improving research.PhilanthrocapitalismRelated: Tying funding to indicators can encourage organizations to optimize scores instead of social outcomes.Quantitative dataRelated: Turning a quantitative indicator into a target can distort the behavior it was meant to track.Research assessmentRelated: When assessment scores drive rewards, researchers and institutions may optimize the metric rather than the underlying goal.Body countRelated: Pressure to raise reported totals weakened their reliability as measures of battlefield progress.ExtrapolationRelated: Extrapolated performance trends can fail when incentives change the behavior being measured.Rotten TomatoesRelated: Marketing pressure around scores illustrates how a reception measure can acquire incentives of its own.Social engineeringRelated: Social plans tied to metrics can invite gaming and distort the behavior they aim to improve.Statistics (numerical data)Related: When statistics guide incentives, people may optimize the metric rather than the underlying goal.Pareto principleRelated: Applying a simple concentration metric as a target can distort the behavior it was meant to guide.Programme for International Student AssessmentRelated: High-stakes attention to PISA scores can encourage systems to optimize measured outcomes.CliodynamicsRelated: Historical indicators may distort conclusions when their proxies are mistaken for the phenomena themselves.DataismRelated: It exposes a recurring failure mode in systems built to optimize numerical indicators.Benevolent dictatorshipRelated: Rulers judged by selected performance measures may distort them while claiming success.
KnowraGoodhart's lawLinked fromLinked fromThe 29 pages that link to Goodhart's law, each with the reason it gives.All 29Related 24Compared with 5Regression analysisRelated: Regression-based metrics can invite incentives that undermine what they were meant to measure.AI alignmentRelated: It explains why optimizing a human-chosen metric can undermine the value that metric was meant to represent.Lucas critiqueRelated: Both ideas warn that intervention can change the behavior behind a useful statistical relationship.GamificationRelated: When a score becomes the goal, participants may optimize it instead of the intended behavior.Mathematical modelingRelated: Optimizing a model’s proxy objective can distort the real-world goal it was meant to serve.Portfolio optimizationRelated: Optimizing a narrow historical metric can encourage portfolios that exploit the metric rather than achieve durable goals.BibliometricsRelated: It explains how incentives tied to bibliometric indicators can distort research behavior.Educational measurementRelated: Schools may improve measured outcomes while neglecting the underlying learning.Financial return on investmentRelated: Using ROI as a rigid target can encourage choices that improve the metric but harm broader goals.IncentiveRelated: Tying rewards to a metric can prompt behavior that raises the score while undermining the goal.Journal impact factorRelated: Treating impact factor as a target can encourage behavior that raises the score without improving research.PhilanthrocapitalismRelated: Tying funding to indicators can encourage organizations to optimize scores instead of social outcomes.Quantitative dataRelated: Turning a quantitative indicator into a target can distort the behavior it was meant to track.Research assessmentRelated: When assessment scores drive rewards, researchers and institutions may optimize the metric rather than the underlying goal.Body countRelated: Pressure to raise reported totals weakened their reliability as measures of battlefield progress.ExtrapolationRelated: Extrapolated performance trends can fail when incentives change the behavior being measured.Rotten TomatoesRelated: Marketing pressure around scores illustrates how a reception measure can acquire incentives of its own.Social engineeringRelated: Social plans tied to metrics can invite gaming and distort the behavior they aim to improve.Statistics (numerical data)Related: When statistics guide incentives, people may optimize the metric rather than the underlying goal.Pareto principleRelated: Applying a simple concentration metric as a target can distort the behavior it was meant to guide.Programme for International Student AssessmentRelated: High-stakes attention to PISA scores can encourage systems to optimize measured outcomes.CliodynamicsRelated: Historical indicators may distort conclusions when their proxies are mistaken for the phenomena themselves.DataismRelated: It exposes a recurring failure mode in systems built to optimize numerical indicators.Benevolent dictatorshipRelated: Rulers judged by selected performance measures may distort them while claiming success.