How Bilby works
Estimate ranges, sources and missing data
A range describes scenarios
Low, expected and high describe different cost scenarios under the selected method. They are not statistical confidence intervals or a promise that a quotation must fall between the endpoints. Expected is an explicit central scenario and can differ from the midpoint.
Separate quantity evidence from price evidence
A measured planner area can be precise while the installed unit rate remains a broad national allowance. SGU soil mapping can support a ground condition assumption without supporting a contractor price. A source link identifies evidence for its stated scope; it does not turn every model coefficient into a published tariff.
Where inputs come from
Inspect the receipt before comparing two estimates. An editable assumption stays an assumption after you change it; it does not become a measured survey.
| Input | Unit | Where from? |
|---|---|---|
| Measured | m², m, items | Planner geometry or available measurements |
| Supplied | SEK or specification | Your entry, supplier quote or project record |
| Inferred | depth, condition, distance | Mapped data and model rules |
| Assumed | quantity or rate | Disclosed fallback/default |
| Confidence and completeness | method-specific | Evidence quality versus amount of usable evidence |
Which method wins
Usable native planner or engineering results and retained project-specific costs precede national references. A researched municipal fee has a locality and scope; a national allowance is a fallback. Manual prices and package inclusions affect the counted project amount. A reference receipt beside a retained legacy amount is comparison evidence, not necessarily the historical formula that created that amount. Unknown quantities may produce provisional assumptions or an unavailable result, depending on the method.
Illustrative ranges and missing values
If a budget range supplies only SEK 100,000 low and 160,000 high, its missing expected value becomes 130,000. If only SEK 120,000 is known, all three values become 120,000. If none is known, it remains missing rather than zero. Separately, native terrain confidence can impose a minimum envelope around expected: medium confidence uses ±25%, so expected 100,000 implies bounds no narrower than 75,000–125,000 before rounding. Existing scenario calculations may make that range wider.
Confidence is not suitability
A high score with little evidence can still have low confidence. Planning can remain neutral with no usable evidence; utility completeness covers only its assessed services. Groundwork and terrain use their own confidence bands and coarse total rounding. A longer price range is not a second contingency amount. Check exclusions, price basis, included VAT, dates and assumed quantities before comparing suppliers.
Dates that mean different things
Method reviewed 18 October 2026. The national reference catalogue records 16 September 2026 research; individual source reviews can be older or need confirmation. Do not treat today’s article date as a new tariff verification. SGU documents the geological layers that support site evidence SGU. Contractor price evidence and your project quote remain separate.
The review date covers this explanation and its implemented method. Price and data dates are stated separately in the article.