Method & Confidence

Why rate at all

Flown Fact and First-Principles Forecast, on the Same Page

The subject is half built. A lunar landing Δv is measured; a Ceres space elevator is arithmetic. A reader has to know which is which at a glance.

Every quantity on this site falls somewhere between a flown measurement and a derivation with no hardware behind it. Hiding that spread would make the forecast look like the record and the record look negotiable. So each figure carries one of four ratings, and the ratings form a colormap in which luminance stands for certainty: gold is brightest and most certain, violet is dimmest and least. The eye reads brightness before it reads hue, so a page can be scanned for how solid its claims are before a single number is read.

The scale runs from what instruments have established to what only calculation supports, with two intermediate steps for things being built now and things credibly announced but not yet real. The order is deliberate: a measurement outranks a construction program, a construction program outranks an announcement, and an announcement outranks a model. Observed

The four levels

Observed, Building, Projected, Modeled

One question sets the level: how close is this to a thing that exists and has been measured?

Modeled is not a synonym for weak. A derivation can be sound and still sit here, because the rating measures distance from a realized, measured thing — not whether the arithmetic is right. What it warns against is reading a model as a schedule.

How to read the tags

The Pill Inline, the Full Rating in the Record

A colored pill sits beside a claim where the rating matters; the page’s Record block collects the ratings for everything on it.

A rating appears two ways. Inline, a small colored pill — Observed, Building, Projected, Modeled — marks the claims where knowing the certainty changes how the sentence should be read. At the foot of every page, the Record block states the confidence for the page as a whole and names the weakest link, so a reader can see what the argument rests on without hunting for pills.

The hue carries the meaning, but the luminance carries it first, which is why gold is reserved for Observed and never used for anything decorative. Page headings are set in cyan rather than gold for the same reason: gold has one job on this site, and a gold heading would blur the signal that gold otherwise sends. The colormap is treated as load-bearing, not styling.

Field status vs confidence

A Dataset’s Own Status Field Is Not the Confidence Rating

Some datasets carry their own vocabulary. It maps onto the four levels; it does not replace them.

The Corridors table marks each route Flown, Planned, or Assumed; the mission chronology uses Flown, Firm, Planned, At risk, and Notional. These are domain status fields, and they line up with the confidence scale rather than competing with it: a Flown corridor is Observed, a Planned one is Projected, an Assumed one is Modeled. The confidence rating is the site-wide scale that lets a corridor, an element, and a player be compared on the same terms; a status field is the finer, dataset-specific grain underneath it.

Ratings move

A Figure Climbs the Scale as Evidence Arrives

A Modeled claim becomes Projected on announcement and Observed on a flight.

Ratings are not permanent. A propellant-transfer demonstration moves orbital refuelling from Projected toward Observed; a mass-driver test article moves lunar launch from Modeled toward Building. The site tracks that movement as it happens. The Experiments Board is where the Projected and Modeled claims are listed with the specific result that would promote them, and the Falsifiers are the results that would instead knock a claim down. Every figure also carries a status date — the month its rating was last checked — so a stale number is visible as stale.

Record

Provenance and Data Sources