CALCULATORS / RELIABILITY AND ENGINEERING STATISTICS

Measurement sample-size estimate calculator

Zero observed failures binomial demonstration count for supplied confidence/reliability.

reliability-sample-size.napkin / editable

SCENARIO COMPARISON

Compare the numbers.

This source describes an analytical model. Its quantities and checks are the view.

Scenario	samples
	29
	10

All calculated quantities are listed with the inputs. Switch to Source to inspect the equations.

SCENARIOS / CALCULATED VALUESSAME SOURCE

Change an input to see the consequences.

test sample budget29.00 ✓ passes

BASELINE INPUTS

reliability0.9 confidence0.95

CALCULATED

samples29

Edits change the baseline source.

THE SOURCE

Edits below update the same model.

# Measurement sample-size estimate
# Scope: Zero observed failures binomial demonstration count for supplied confidence/reliability.
# Status: illustrative worked example; engineering review pending.
# Inputs are example values. Checks are editable example requirements, not standards.

reliability = 0.9
confidence = 0.95

samples := ceil(log(1-confidence)/log(reliability))

check test sample budget: samples <= 30

scenario alternative: reliability = 0.72

Calculated from your source · not saved

Baseline · Baseline

test sample budget29.00 ✓ passes

alternative · alternative

test sample budget10.00 ✓ passes

Illustrative first-order example. Checks are editable requirements, not design standards. The broader topic may require additional equations, data, and validation.

Formula and units

This example uses the following equations. Input names and units match the editable source above.

samples: samples = ceil(log(1-confidence)/log(reliability))

Original worked example

These are the baseline values in the downloadable source. Edits above let you test your own values.

Given

reliability: 0.9
confidence: 0.95

Calculated

samples: 29

Assumptions and limits

Zero observed failures binomial demonstration count for supplied confidence/reliability.

The requirement checks use the limits written in this example. Passing a check does not establish that the model covers every load, material behavior, or operating condition in a real design.

Download this .napkin example ↓

Take it further with your agent

Napkin keeps the inputs, equations, checks and assumptions together in one study. Your agent sets it up; you change the numbers.

Try with Claude ↗

codex 'Use Napkin (https://napkin.engineering/mcp, or the local Napkin plugin) to build a study: Measurement sample-size estimate. Inputs: variance; target interval width; confidence convention. Outputs: required sample size. Model scope: Explicit distribution, independence and sample assumptions. Read napkin_capabilities and the current snapshot. Ask me for the values that matter, label assumptions with units, author .napkin source, validate it and apply it with napkin_source using the current revision, then open the study. Add checks for the limits I supply. Keep unsupported analyses explicit.'

Claude needs the Napkin connector; Codex needs the Napkin plugin. Set up your agent ↗

The broader study

The topic brief below describes the broader scope. The worked example above states the particular calculation it implements.

Inputs: variance, target interval width, confidence convention
Outputs: required sample size
Modes: calculate, compare, sweep
Drawing: Distribution and reliability plot

EXECUTABLE WORKED EXAMPLE · FIRST-ORDER MODEL · CHECK THE SCOPE & ASSUMPTIONS