Importance, Confidence, and Risk
NataPulse uses several scores for different questions. They should not be treated as interchangeable.
Importance
Section titled “Importance”Importance answers: How much attention should this event receive relative to other current events?
It is displayed on a 1–10 scale. Inputs can include materiality, entity relevance, source quality, corroboration, velocity, market or on-chain impact, and curated semantic signals.
Importance is a priority score. It is not confidence and it is not a forecast of price return.
Confidence
Section titled “Confidence”Confidence answers: How strong and reliable is the available evidence for this item or conclusion?
It can reflect source reliability, data quality, independent corroboration, consistency, and evidence completeness. It is generally shown as a fraction, percentage, or visual bars.
Confidence is not certainty. Several high-quality sources can still share the same incorrect premise.
Sentiment
Section titled “Sentiment”Sentiment represents the direction or tone of relevant text where applicable. It can be useful for detecting a change in discussion, but should not be confused with fact, importance, or expected return.
Quantitative confidence and quality
Section titled “Quantitative confidence and quality”Quantitative signals have their own confidence and data-quality fields. A model may produce a strong output from stale or incomplete data; NataPulse therefore exposes quality separately and can suppress invalid outputs.
Wallet risk
Section titled “Wallet risk”Wallet Intelligence can display a user-facing wallet risk score from 1 to 10 based on the available profile, labels, and activity. It describes investigation risk, not the probability of loss on an investment.
Internal scores
Section titled “Internal scores”The control plane can use additional private review and publication signals. Internal risk and publishability components are deliberately excluded from public product interfaces.
See Scores and Interpretation for the full methodology.