Lab

GTA VI Lab — reproducible tests bound to protocol + platform + GameVersion + date + method. Tested after launch starts here.

Every Result is derived from Measurements via a stated aggregation (mean/median/two-way mean), shows uncertainty, and carries reproducibility state (UNTESTED → SINGLE_RUN → REPRODUCED → CONTESTED → BLOCKED). Only REPRODUCED powers leaderboards.

Vehicle Lab

Speed, acceleration, braking — mph, seconds, two-way mean

View family →

Police Lab

Recognition, witnesses, search duration, escalation

View family →

Living World

NPC reactions, persistence, world simulation

View family →

Economy Lab

Dollars per hour, profitability, variance

View family →

Performance Lab

FPS, frame time, platform/mode comparisons

View family →

Weapon Lab

Dormant until M8 — shipping data required

View family →

Research workflow

Question → ResearchTicket → TestProtocol → TestRuns (append-only, ≥1 per batch) → Measurements (immutable, units validated) → Result (derived, uncertainty required) → Review → Publish → retest via VersionChange. Evidence media lands in R2 evidence/tests/<protocol>/, metadata in media_assets.

Researcher console: admin/lab · Methodology: /methodology

Launch-readiness

At M7, a researcher can define a protocol, record repeated runs with platform/version binding, produce a versioned result with methodology, retest later, and compare — without engineer intervention. Shipping measurements are BLOCKED_EXTERNAL until the game ships; the framework is ready.