bent.WTF
Bent Measures visual showing a person measuring a future plan with a dollar ruler that bends as purchasing power changes over time.

Start Here · Numbers · Measurement Failure

The measure is not the reality

A guided inspection of definitions, scope, populations, denominators, adjustments, exclusions, revisions, and the claims a measure cannot carry.

BM0002 · The Bent Measuring Stick Mnemonic layer; claims and caveats remain in the record. Inspect the full visual transcript

Bent Measures · BM0002

The Bent Measuring Stick

The Bent Measuring Stick is the mnemonic layer for this path. Open it on its own, then return to the evidence, limits, and countercase below.

Bent Measures visual showing a person measuring a future plan with a dollar ruler that bends as purchasing power changes over time.

First, in plain language

Ask what the number was built to measure.

A number does not arrive with one universal meaning. An average describes a group under a set of rules. A rate depends on who is counted and on its denominator—the total or group used for the comparison. An index follows a defined set of items or changes; it does not become a complete biography of every household.

Most trouble starts in the sentence around the number. A careful measure gets stretched into a larger claim about typical people, fairness, success, or lived cost. Read the definition, find the population and denominator, and then decide how far the conclusion can travel.

  1. Say the question in ordinary words

    Write what you think the chart answers before reading the technical definition. The gap between those two statements is where misuse often begins.

  2. Open the label

    Find the people or things counted, the unit, the period, the denominator, and any exclusions or revisions. These details set the meaning of the result.

  3. Draw the boundary

    State what the measure supports and what it cannot establish by itself. A narrow boundary can make a number useful; it does not make the number false.

Object on the table

A consumer-price index used as though it were a complete account of one household’s cost of living.

Do not solve it from the headline. Preserve the first description, then work the route.

Bounded finding

What this path can establish

A measure should be read inside the question, population, unit, period, denominator, and method that produced it. A scope boundary narrows the lawful claim; it does not by itself make the measure false.

Try to disprove it

The strongest honest countercase

An incomplete measure may be answering a narrower question correctly. A cleaner alternative can also introduce a different omission, lag, or tradeoff.

A way to look again

Bent Number Record

Respect the measure. Refuse the extra claim it was not built to carry.

Pause here: Which of these lines would have changed your last decision if it had appeared sooner?

  1. 01Question in ordinary language
  2. 02Formal metric definition
  3. 03Unit and population
  4. 04Numerator and denominator
  5. 05Period and geography
  6. 06Adjustments and revisions
  7. 07Exclusions and alternative measure

Open the complete field guide →

Continue through the introduction

Notice it. Trace it. Use the method.

Go deeper · Bent Number Workshop

Nine inspections. One number at a time.

Nine guided inspections for learning what a number measures, where its method stops, and what evidence belongs beside it.

Boundary: CPI-U is the worked methodological case. These modules do not calculate a household inflation rate or treat statistical adjustment as evidence of manipulation.

  1. N01 · Measurement primer The Index Has a Job

    What exact question was this measure built to answer?
  2. N02 · Measurement primer The Average Household Does Not Exist

    Whose spending pattern enters the average, and whose does not?
  3. N03 · Measurement primer Shelter Is Not the House

    Is the question about consuming shelter or acquiring a housing asset?
  4. N04 · Measurement primer Substitution Without the Steak Story

    At what level does the formula allow purchasing patterns to shift?
  5. N05 · Measurement primer Quality Is Not One Lever

    What happened when the sampled product changed?
  6. N06 · Measurement primer The Revised Past

    Can this published value change after its first release?
  7. S01 · Source School Read the Footnote

    Which sentence in the method changes how the headline should be read?
  8. S02 · Source School Find the Denominator

    What exactly is the reported result divided by, weighted against, or indexed to?
  9. S03 · Source School Build the Countercase

    What is the strongest honest reason the measure may be right for its job?

Six transfer tests

Wages, school ratings, hospital scores, on-time performance, customer satisfaction, and revised data. None arrives with the lens already named.

Pause here

What can you see differently now?

  1. Can you name the population, unit, period, numerator, and denominator?
  2. Can you distinguish a methodological choice from a methodological error?
  3. Can you find the adjustment, exclusion, or revision that matters for the claim?
  4. Can you offer a better measure without pretending it answers every adjacent question?

Keep the record underneath