How BullFUD scoring works

For every scan we pull recent items about the asset from a wide pool of public sources, grouped roughly into three kinds:

• Established financial and crypto press — major newswires, market publications and dedicated crypto or equity outlets.
• Primary and authoritative sources — company and regulatory filings, project or exchange announcements, and official blogs, so material facts are read from the origin rather than a retelling.
• Early-signal and community sources — on-chain flow trackers, aggregators and discussion boards, which often surface a narrative hours before it reaches mainstream coverage.

Each source carries a credibility weight. A wire report or filing counts for more than an anonymous post, and low-trust items need a stronger signal to make the list.

Raw items are cleaned up before anything is scored. We drop exact and near-duplicate headlines, discard items that only mention the ticker in passing, and cluster pieces telling the same story so one event does not fill the whole page.

What survives is ranked on four things:

• Relevance — is the asset the subject of the story, not just a name-drop?
• Source credibility — the weight described above.
• Recency — newer items outrank older ones covering the same ground.
• Materiality — how much the story could plausibly change the picture: earnings, filings, hacks, listings and large flows outrank opinion pieces.

Items that read as unconfirmed get a Speculation tag so rumour is never presented as fact.

Every surviving item gets a 0–100 score for the mode you are in. In FUD mode the score answers "how bearish is this, and how much should it matter?"; in Bull mode it answers the same question in the other direction. The score blends the strength of the sentiment in the item itself with its source weight, recency and materiality, so a strongly worded but flimsy post lands below a neutral-sounding filing with real consequences.

The number on a card is a relative ranking signal, not a probability or a price target. A 90 means "this is the loudest, best-supported item in this direction right now", not "this asset will fall 90%".

Versus mode runs both sides on the same pool and compares the summed weight of each case, which is what the tug-of-war meter shows. The sentiment history chart plots the average score per scan over time, so you can see whether a narrative is building or fading.

Headline intensity alone is a poor gauge of a narrative, so each score is adjusted by a handful of context signals, and every one of them is shown on the results page rather than hidden in the maths.

• Coverage volume — how many distinct stories ran today against this ticker's own 30-day average. Ten bearish stories in a week of silence is a different event from ten in a week of constant noise.
• Outlet bias correction — we track how often each outlet publishes bullish versus bearish coverage across all tickers. An outlet that is always negative carries less weight when it is negative again; an outlet breaking from its usual lean carries more.
• Expectedness — scheduled and widely anticipated events are discounted, because the market has largely absorbed them. Genuine surprises are weighted up.
• Organic amplification — twenty outlets reprinting one wire release is one voice, not twenty. Near-verbatim reprints are collapsed and marked Syndicated, PR-wire origins are discounted further, and only outlets that tell the story in their own words count as corroboration.
• Sector-relative sentiment — the ticker's bull share is compared with its peer group, so you can tell a company-specific problem from a sector-wide one.
• Theme decomposition — each story is bucketed into regulatory, security, adoption, macro, technical, flows or speculation, and the panel shows what share of the score each theme contributes.
• Momentum — the score history is differentiated twice, so a narrative that is building reads differently from one that has peaked and is fading.
• Narrative-price divergence — the weekly price move and the narrative tilt are both expressed in standard deviations, and the gap between them is reported as a number rather than a hunch.

Signals that need history are hidden until there is enough of it. Coverage volume, momentum and peer comparisons only appear once a ticker has been scanned enough times for the baseline to mean something.

Scans are live: hitting the button fetches sources at that moment. To keep results fast and avoid hammering publishers, a completed scan for a ticker is shared with other users for up to 24 hours, and popular tickers are re-warmed in the background roughly every 30 minutes, so what you see is typically minutes old rather than hours. Price data and market context chips update on their own schedules from their providers.

BullFUD is a sentiment discovery tool. It tells you what is being said about an asset and how strongly, not whether that view is correct. Coverage is limited to public sources we can reach, headlines can be misleading, automated ranking makes mistakes, and sentiment often lags or overshoots price.

Nothing here is financial, investment, legal or tax advice, and nothing is a recommendation to buy or sell anything. Always click through and read the original source, and do your own research before making any decision.