The score sees only what it counts.

CONCEPTS / VALUE & POWER

Proxies and incentives

Is a closed case a solved problem?

Follow the idea.

A concept, in three frames

Read the score.

Check which outcomes the reward counts.

Follow the effort.

Resolution and closure use effort differently.

Inspect the cases.

A closed case can remain unresolved.

A WORKING DEFINITION

Proxies and incentives

A proxy is a measure used to stand in for an intended outcome. Incentives attach a reward to a measure; the rewarded action can depend on what that measure counts.

2r + c = 20 effort units

FOLLOW ONE EXAMPLE

Name the outcome before choosing the score.

With twenty effort units, twenty administrative closures yield twenty closed cases and zero resolutions. Ten resolutions use the same effort and yield ten closed cases with ten resolutions. At the default midpoint, five resolutions and ten administrative closures yield fifteen closed cases. The two counts overlap; they are not additive categories.

Work through the investigation →

WHERE THE CLAIM STOPS

Keep the assumptions attached.

This model supplies the action costs and assumes perfect confirmation of resolutions. It does not predict how real people behave or prove that every target fails.

Inspect this model’s assumptions and definitions
  • This is a deterministic constructed model, not an empirical claim about workers or inevitable target failure.
  • Confirmed resolutions are assumed to be perfectly observed; real verification can be difficult and costly.
  • Switching the reward does not reallocate effort. Closed and resolved counts overlap and must not be added.

Definitions used in this instrument

budget
twenty-effort-units-and-twenty-distinct-cases
actions
resolution-costs-two-admin-closure-costs-one
intendedOutcome
confirmed-resolutions-perfectly-observed
closed
resolutions-plus-administrative-closures

SOURCES & SCOPE

Read beyond the model.

Manheim & Garrabrant (2019 revision), Categorizing Variants of Goodhart’s Law ↗

Primary conceptual analysis distinguishing mechanisms of metric overoptimization. This original deterministic proxy example is not an empirical study or a simulation of all those mechanisms.

All example data, figures and exact calculations on this page are original constructed teaching material. Source publications provide conceptual background, not the example dataset.

Open the instrument →