Melega Score — canonical core specification
One ecosystem-wide project quality and trust rating. The same canonical evidence pack, read under the same spec version, produces the same Melega Score on every platform.
What Melega Score is
Melega Score is a project quality and trust rating on a 0–100 scale. 100 is representable: the scale is not artificially capped. It expresses how much verifiable evidence supports a project, not how a token will perform.
Melega Score is not investment advice, a guarantee, an audit or a certification.
Canonical version
The canonical version is melega-score-v1. Behaviour under a published version is immutable: weights, axis definitions, normalisation and publication gates cannot change inside a version. Any change to them requires a new version string and a changelog entry. Historical scores are never recomputed in place under a different version.
- Scope
- ecosystem-wide
- Baseline lineage
- swift-shift::trustScore
- Paid influence
- false
Formula and axes
The baseline mathematical lineage is swift-shift::trustScore: six weighted axes whose weights sum to 1.00.
The antiSpamConfidence multiplier is a source-proven property of the historical swift-shift baseline. It is a project-quality/anti-abuse signal, not evidence completeness, and it must never be read as missing evidence being multiplied down.
The six-axis trustScore family is the adopted mathematical baseline of melega-score-v1. The historical antiSpamConfidence multiplier is disclosed above as lineage, and evidence-confidence gating is an added Core publication contract — not a seventh weight and not merged into any axis.
volatility — Exists in the historical baseline as a diagnostic-only signal. It carries no Core v1 weight.
Momentum is not part of Melega Score. The dynamic momentumScore remains a separate local metric and never enters this formula.
Historical tier lineage
The bands below are the historical tier lineage recorded in the canonical report. They are documented as candidate banding for the ecosystem core; they do not rename or override the bands Melega Space currently shows on its own surfaces.
EvidencePack contract
Platforms collect and normalise evidence differently, but the scoring input is a single canonical EvidencePack. The contract is: same canonical EvidencePack + same Core Spec version = same Melega Score, independently of whether the consumer is BLACK, DEX, SPACE or a future platform.
The platform adapter boundary sits strictly before the core: an adapter may discover, normalise, timestamp and attribute evidence. It may not weight, adjust, floor, cap or override the resulting score.
Confidence and publication gates
The Core Quality/Trust Score, Evidence Confidence and Publication State are three distinct outputs and are published as such. Evidence confidence describes how reliable and how fresh the observations are; it is not a statement about project quality. Missing evidence is never treated as a negative fact about a project.
Unknown is never zero. No unknown axis, indicator or field is imputed to 0, 70, 74 or to any other fallback number. When evidence or confidence is below the publication requirement, no number is shown at all.
Evidence Confidence gates publication and is published beside the score. It is never multiplied into the score as a project-quality penalty, and missing evidence is never treated as a low-quality fact.
Local metrics are not Melega Score
Only the canonical core output may be labelled “Melega Score”. Every platform-local metric keeps its own name.
| Local metric | Scope | Why it is not Melega Score |
|---|---|---|
| Momentum | Local / dynamic | momentumScore is explicitly outside Melega Score. |
| Platform Contract Score | Platform-local | Integration-level contract compliance, not project quality. |
| Civilization Readiness | Ecosystem-internal | Operational readiness of an organ, not a project rating. |
| Launch Completeness Estimate | Platform-local | Checklist completion, not evidence-weighted trust. |
| Evidence Quality | Evidence layer | How good the observations are, not how good the project is. |
| Market Quality | Market layer | Liquidity/market telemetry quality only. |
| Runtime Readiness | Platform-local | Service/runtime health, unrelated to any project. |
| Creator / discovery metrics | Marketplace-local | Creator and discovery ranking inputs, never a project trust rating. |
Commercial neutrality
No commercial action can improve a Melega Score. Paid promotion, featured placement, listing purchases, visibility boosts, referral activity, SmartDrop participation, marketplace spend and claim state are all absent from the core input by construction — there is no field for them to be read from.
Transparency
A score is never published alone. Every published score carries:
Relationship to the legacy SPACE formula
Melega Space previously designed and ratified its own seven-dimension formula, space-score-v1, together with the space-foundation-1 evidence model. That work is historical and SPACE-specific. It is preserved, not deleted, and it is not the ecosystem-wide core: where it differs from this specification, melega-score-v1 is the canonical direction.
The SPACE-specific historical design remains readable at the SPACE evidence model and the SPACE scoring methodology. Those pages describe what Melega Space currently computes on its own surfaces; this page describes the canonical ecosystem standard.