TC1-27 Fermi estimation and order-of-magnitude checks (back-of-the-envelope estimate)
Breaks an unknown quantity into factors that can be estimated, multiplies them, and produces a number with a plausible range used as a bound on a detailed result.
What it does
Breaks an unknown quantity into factors that can be estimated and multiplies them. Artefact: a number with a plausible range, used as a bound on a detailed result.
When to use it
Use when you need to know quickly whether a figure is the right order of magnitude or to bound a quantity that has no direct data.
When it does not fit
Independent routes differ by more than an order of magnitude, or they share a bias, so errors do not cancel.
Result format
The result lists the factors the unknown quantity is broken into, an estimate for each and the product of those estimates. It ends with a number and a plausible range, stated as a bound against a detailed result.
Requirements on the material
| Property | Requirement type | Requirement |
|---|---|---|
| Content type | Always applies | requires=text; requires=tabular |
| Unit of analysis | Always applies | requires=topic |
| Variable types | Raises a warning | prefers=ratio |
| Appraises source credibility | Applies in some cases | no |
Not restricted on: Structure, Time structure, How the sample was chosen, Level of detail available, Number of sources, Whether sources can be joined, Amount of material, Known data problems, Producer interest, Source dependence, Origin, Document genre, Language, Personal information, Security classification, Licence.
| Minimum material | A quantity that can be split into at least two factors with rough values and units. Two independent routes are preferred. |
|---|---|
| Additional sources needed | none |
| Not for (rules) | none |
| Not for (text) | Wrong choice when independent routes differ by more than an order of magnitude, or when they share a bias so that errors do not cancel. |
Choosing and ordering
| Kind of technique | Verification technique, runs in step 6 |
|---|---|
| Phase tag | validate |
| Other phase tags | analyse |
| Kind of finding | difference; contradiction |
| Question answered | how much |
| Strongest statement it supports | describe |
| Needs as input |
|
| Makes as output |
|
| Must exist first | None |
| Loop target | proceed with limit stated |
Library checks by stakes
Each stakes level includes the checks of the level below it. Checks that need a person apply at high stakes only, so the process leaves them out at lower stakes (decision 38). Repeatability: judgement dependent.
| Stakes | Check | Person involved | Counts as a different method |
|---|---|---|---|
| exploratory | code execution for arithmetic | none | no |
| decision-informing | code execution for arithmetic | none | no |
| decision-informing | sensitivity analysis | none | no |
| high-stakes | code execution for arithmetic | none | no |
| high-stakes | sensitivity analysis | none | no |
| high-stakes | second technique with different assumptions | none | conditional |
| high-stakes | independent review | needed | no |
Failure modes and limits
| Model suitability | model plus code |
|---|---|
| Typical model failure | Multiplies factors with inconsistent units or correlated errors and reports the product to more precision than the inputs allow. |
| Error sources | input data; statistical assumptions; expert opinion |
| Certainty limits | The estimate is a bound and not a measurement. Certainty drops when the routes share a bias or when factor values are guesses with no source. |
Related techniques
- pairs with TC1-28 Dimensional and unit checks: unit checks confirm that the factors multiply to the right units
- pairs with HE1-21 Distribution, outlier and base-rate checks: base-rate and distribution checks supply the rough factor values
Evidence and tag basis
Tested. Tested for teaching only: a classroom study reported medium effects on transfer, with difficulties in units and notation. No test of research accuracy was found. (Wikipedia, Fermi problem; arXiv 2405.16480, teaching estimation)
Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.