How we know
Methodology
Vice Atlas exists to answer one question honestly: what is actually known about GTA VI, and how sure can we be? This page explains how we get there.
Claim-first internally, intent-first publicly
Internally, every piece of information is recorded as a structured claim — one assertion, with its sources, evidence strength, and current standing. Publicly, you never browse that database. You read briefings that group related claims into plain-language analysis. The rigour is underneath; the reading experience is on top.
Every public page is a projection
Nothing on this site is authored directly against the public. Public pages are generated from a public-safe projection of the internal record. That projection deliberately strips internal reasoning, source-reliability assessments, raw numeric scores, unapproved material, and anything not cleared for publication. A claim only ever appears publicly inside a briefing — never as a raw database row or internal identifier.
How we label confidence, evidence, and standing
We use plain-language bands, never raw internal scores. Nothing below “Official” is stated as fact.
Status
- Confirmed Established as fact for the site. Requires E1/E2, OR >=2 independent E3 sources. Never reachable by E4/E5/E6 alone.
- Inference Not officially stated, but strongly supported by reputable reporting or sound deduction from confirmed facts.
- Speculation A reasoned possibility with weak or indirect support; clearly the author's analysis, not reporting.
- Rumor A circulating claim with weak or uncertain provenance; tracked because the audience is asking, not endorsed.
Confidence
- Very Low
- Low
- Moderate
- High
- Very High
Evidence strength
- Official primary
- Official media
- Reputable secondary
- Single-source / insider
- Leak / datamine
- Community inference
Confidence is a subjective assessment, not a measured frequency. A claim's operational standing (for example: active, watching, contradicted, superseded, resolved) is shown transparently — if something was contradicted or replaced, we say so rather than quietly dropping it.
The no-leak policy
We classify publicly circulating information; we do not host, mirror, link, or reconstruct leaked or copyrighted material. We may record the real-world fact that a leak circulated, framed neutrally — but we never deliver the leaked artifact, and we never link to leak repositories, datamine dumps, or file-sharing of copyrighted content.
Human approval and under-claiming
Publication is manual. A human reviews every briefing before it goes public; no automated system publishes on its own. When evidence is thin, we under-claim rather than over-claim — rumor and speculation are always visibly labeled and never presented as confirmed.
See our legal & disclaimer notice for affiliation, trademark, and takedown information.