Knowledge center · Factor library
News Sentiment
Aggregate tone of recent news headlines, shrunk toward neutral with limited coverage.
Formula
Average headline sentiment, shrunk by effective-evidence weighting (more/fresher headlines carry more weight; duplicates and stale items are down-weighted)
Interpretation
Captures whether recent news flow skews positive or negative — a fast-moving, qualitative complement to the slower-moving fundamental and quality factors.
What good looks like
Sustained positive sentiment across multiple independent sources.
What bad looks like
Sustained negative sentiment across multiple independent sources.
Typical range
-1 to +1 before shrinkage; pulled toward zero with limited or stale coverage.
Limitations
Only as good as the underlying news feed and its sentiment labeling; single sensational headlines are deliberately down-weighted so they can't dominate the score, which also means the factor reacts more slowly to genuinely fast-breaking news than a human reader would.
How OmniSignal uses it
A fast-moving signal the slower price- and fundamentals-based factors can't capture on their own.
Where it enters the score
News family
References
Standard NLP sentiment scoring applied to financial headlines; shrinkage estimation is a standard technique for noisy, variable-count evidence
Related concepts