Any single signal source carries a meaningful false-positive rate. Governance activity can spike for mundane procedural reasons. On-chain announcements can be routine. Social momentum can be manufactured. But the probability that all three independently produce the same false signal, on the same entity, in the same window, by coincidence, drops sharply — which is exactly why cross-source corroboration carries the most weight in the entire scoring model.
What "three sources" actually means here
This isn't three mentions of the same headline being reposted across platforms. It's three structurally independent collection mechanisms — each watching a different layer of activity — separately flagging the same entity within the same observation window. Governance forum activity is a different data layer from on-chain announcement monitoring, which is a different layer again from protocol development signal tracking.
How this feeds CIA scoring
- Each corroborating source contributes to the Collective Intelligence Asset score independently
- Source diversity is tracked explicitly — not just signal count, but how many structurally distinct sources are confirming
- Recency and frequency compound together over time rather than scoring any single moment in isolation
Why this matters more than any individual signal's strength
A single very strong signal from one source is, on its own, still a single point of failure — it could be an anomaly, a data error, or a coordinated but ultimately hollow promotional push. Three moderate signals from three independent sources, confirming each other, is structurally more reliable than one strong signal alone. This is the principle the entire detector stack is built around, across every domain on the platform.