Institutional-grade signal scoring
A twelve-component signal engine — attention, FinBERT sentiment, source credibility, prediction-market move, liquidity, market participation, price confirmation, Fed-rate path, FRED macro confirmation, fundamentals and 3- and 7-day signal momentum — blended into a single score that updates every cycle and is then discounted by how crowded the story already is. Weights redistribute when a component has no data, so scores stay comparable across topics.
Always-on multi-source discovery
Hundreds of feeds across news, social, prediction markets and price anomalies, scanned every five minutes. Nothing reaches the board without two or more independent sources.
Forward-tested, self-correcting, public
Every prediction is tracked at 1-day, 1-week, 1-month and 3-month horizons, with directional alignment rates, excess returns against benchmarks and per-topic scorecards. Weights recalibrate every six hours from realised outcomes, and Brier scores against Kalshi and Polymarket are published on a public leaderboard.
From signal to action in one pipeline
Narrative detection, ticker exposure, price and options confirmation, prediction-market pricing and broker deep links — every step visible, without switching tools.
What-if scenario analysis
Ask what happens to a sector under a given shock and get structured analysis: affected tickers, sector rotation and prediction-market pricing, grounded in live signal data. A Markets view surfaces where the engine systematically disagrees with what Kalshi and Polymarket are pricing.
From headline to the bottom line
Most platforms stop at sentiment. Narratick reads each company's filings and earnings on the way through — margin shifts, guidance revisions, leverage spikes, free-cash-flow inflection, dividend cuts — and surfaces them as micro-narratives on the affected ticker.