What is becoming technically possible
Papers, preprints, clinical trials, patents, citations, and lab-to-market signals.
The data layer behind the calls
Vaticinus maps the evidence behind a forecast before the model takes a side. The point is not a big corpus number. It is checking the right parts of the world.
What the system checks
Every forecast is routed through the evidence that should move first if the thesis is true. These are the checks, not a vanity count.
What cannot scale on demand
Trade flows, mineral data, component capacity, shipping routes, and supplier concentration.
What is actually getting built
Permits, interconnection queues, project filings, land records, and construction evidence.
Where the market has already priced it
Public filings, prices, funding rounds, procurement language, and capex commitments.
Where rules move the constraint
Export controls, sanctions, subsidies, procurement rules, and licensing bottlenecks.
Whether the thesis is becoming true
Resolved events, price anchors, production data, disclosure changes, and kill signals.
How the data changes the call
The substrate is a constraint map. It tells the forecaster where the bottleneck lives, who owns it, whether supply can answer price, and what observation would make the claim fail.
No prompt-only answers
The answer has to touch source data before it becomes a forecast.
No stale evidence
Dates matter because forecasts are only real before the outcome.
No unpriced edge
If the market already sees the constraint, the call is not useful.
The source families
The same call may need scientific evidence, supply-chain evidence, market evidence, and policy evidence on one page.
papers and preprintspatents and assigneespermits and docketsfilings and disclosurestrade and customsmarket pricespolicy and sanctionsclimate and earth systemsclinical and health datapublic forecast boardsentity and ownership graphsresolved outcomes
Data checks are evidence inputs, not investment advice.