How it works · the radar methodology
Plain-language methodology.
What the radar reads, how the Council reviews each story through structured lenses, how the Radar Score ranks news by market relevance, and how the system learns. The Radar Score measures market relevance, not expected return — nothing here is a prediction, a recommendation, or investment advice.
01 · the input
What the radar reads.
Every day the radar gathers market news across thousands of sources — wire services, company disclosures, macro releases, regulatory filings, and credible financial commentary. The goal is not to read everything; it is to find the handful of stories that actually matter for markets and explain why.
Coverage runs across several streams that are logged separately — never conflated:
- Council Daily — a multi-LLM review of the day's most market-relevant stories; a focused daily digest, quiet days expected.
- Intraday Pulse — a lighter hourly pass that surfaces developing stories as they gather momentum through the session.
- Catalyst Stream — a near-real-time news poll that flags material events the moment they break.
Each stream keeps its own log and its own review history, so you can follow what each one surfaced independently — no cross-stream averaging.
Before the Council weighs in, each story is given a first-pass relevance read — how widely it is being reported, how directly it bears on tradable assets or themes, how fresh it is, and how credible the sourcing is. This produces the composite_score that the Council then reviews.
02 · the deliberation
The Council.
For each top-scoring story, the radar runs a five-role review of independent reasoning models:
- Risk — identifies how the story could be overread or mis-framed, and what would make it matter less than it first appears.
- Builder — makes the strongest case for why the story is genuinely market-relevant.
- Strategist — frames the story in macro / catalyst / second-order context.
- Contrarian — dissents from the consensus reading; pressure-tests assumptions.
- Chairman — synthesizes the four roles into a single Council view, with preserved dissent, of why the story matters and what to watch next.
The chairman's output is what the radar publishes. Every role's reasoning is stored alongside it so any radar item can be inspected at /track-record → click the inspect link on any recent item.
03 · the lenses — how relevance is judged
Eight lenses on every story.
The Council reviews each story through eight structured lenses. Stories are ranked by market relevance across macro, asset/sector, catalyst, narrative velocity, source quality, positioning, second-order, and uncertainty lenses. The lenses are how the Radar Score gets built — and the Radar Score measures market relevance, not expected return.
Lens 1 · Macro
Does the story bear on rates, inflation, growth, liquidity, or policy — the forces that move the whole market, not just one name?
Lens 2 · Asset / sector
Which specific assets, sectors, or themes the story touches, and how directly.
Lens 3 · Catalyst
Is there a discrete event — an earnings print, a ruling, a data release — that gives the story a clear before-and-after?
Lens 4 · Narrative velocity
How fast the story is spreading and gathering attention across sources right now.
Lens 5 · Source quality
How credible and primary the reporting is — original disclosure and wire reporting weigh more than aggregation and speculation.
Lens 6 · Positioning
Whether the story confirms or cuts against what the market already seems to expect — the surprises that matter most.
Lens 7 · Second-order
The knock-on effects — who else is affected downstream once the obvious read is priced.
Lens 8 · Uncertainty
What is still unknown or unconfirmed — preserved openly, so the Council's view never reads more certain than the facts allow.
04 · the grade
Radar Score — 0 to 100.
Every story is given a Radar Score that combines:
- The first-pass relevance read from the composite score
- How directly the story touches tradable assets or themes
- Agreement across the eight Council lenses
- Narrative velocity — how fast it is spreading
- Macro and second-order reach
- Source quality and freshness
- Penalties for thin sourcing, contradiction, or missing context
The Radar Score ranks news by market relevance. It is not a forecast of price, return, or direction — a high score means a story is worth your attention today, nothing more.
06 · the retrospective
Did the story actually develop?
A radar is only useful if it surfaces stories that turn out to matter. So every item is followed up: did the story keep developing, spread further, or fade by the next session?
For each published item, the radar tracks three things over the days that follow:
- Develops — the story keeps generating credible follow-up coverage.
- Fades — attention drops off and the story does not develop further.
- Window — the follow-up horizon over which we check (next session, several days).
This is a relevance retrospective, not a trade outcome — we are scoring whether the radar called the right stories, not whether any price moved a certain way.
These follow-ups accumulate into the relevance-by-tier view at /track-record. Every item the radar surfaces is on the record, win or lose — no cherry-picking. None of this is a performance claim.
07 · continuous self-evaluation
The ranking is held to its own record.
The radar's relevance ranking is not static. It is continuously re-evaluated against what actually developed in the market — which surfaced stories went on to matter, and which faded. A change to how relevance is scored is only adopted when replaying the radar's own history shows it would have ranked the stories that mattered more accurately than the current approach. Nothing is adopted on intuition: a candidate has to beat the incumbent on the record before it ships.
This is a relevance-ranking discipline, not a performance engine, and it makes no claim about returns. It exists for one reason — so the radar keeps getting better at the only job it has: surfacing what is worth your attention today.
08 · what the score means
The Radar Score measures relevance, not return.
This is the most important line on the page: the Radar Score is a measure of market relevance — how much a story is worth your attention today — and nothing else. It is not a forecast of price, return, or direction, and it is not a buy, sell, or hold call.
A story can score high because it is genuinely important and still be followed by a market that goes the “wrong” way, or nowhere at all. Relevance and outcome are different questions. We only claim to be good at the first one.
No performance claim
AnotB makes no claim about returns, outperformance, or trading edge. It does not pick stocks, predict prices, or make trade recommendations. It is a daily market radar — a market intelligence journal that explains what matters and why.
The public framing is a Council-reviewed market radar— the news that matters most for markets, ranked by relevance.
09 · disclosures
What this is and isn't.
- This is commentary, not advice. Radar items describe the Council's read of public market news. They are not individualized investment recommendations.
- Relevance is not a prediction. The radar log shows which stories were surfaced and how they developed. It does not forecast prices and does not guarantee what happens next.
- No picks, no positions. AnotB does not pick stocks, set price targets, or recommend trades. No custody, no brokerage, no order flow. The output is context for your own process — you decide what to do with it.
- No performance or alpha claim. Nothing here is a statement about returns, outperformance, or trading edge. The Radar Score measures market relevance, not expected return.
- Model-driven. Council role outputs come from frontier reasoning models accessed via OpenRouter (or a local Apple-Silicon model when configured). No human writes or edits the Council's text after the review finishes.