MARPLA methodology essentials: sources, derivations, privacy

Make the claim reproducible and proportionate

Public collection is bounded and rotates across saved games. Rate limits and partial responses can create gaps, so coverage accompanies interpretation. Derived session estimates, growth changes, confidence bounds, and the scouting score help prioritize review; their assumptions and risk labels travel with the value. They do not prove profitability or future success.

Private analytics requires the owner or a role with sufficient permission. Use minimal access, protect credentials, and retain only what the decision needs. For every conclusion, write the exact claim, evidence, inference, limitation, action, and next check. If another reviewer cannot reconstruct the claim from those fields, the methodology is incomplete.

When a source or formula changes, append a correction rather than overwriting the earlier decision record. Reproducibility includes the history of changed assumptions.

1. Classify the data layer.
2. Record source, definition, time, and coverage.
3. Separate observation from inference.
4. Attach a decision, limitation, and next check.

Primary sources

• Roblox Analytics dashboard
• Roblox API keys
• Roblox Discovery