CAL Litepaper.

CAL is a forecasting engine for tokenized equities on Robinhood Chain. It publishes a fair_value for every covered ticker, emits a signal when an edge and a dated catalyst coincide, writes both on-chain before the catalyst resolves, and scores itself when the window closes. This document specifies the method, the data contract, and the integration surface.

A fair value is an independent estimate of where a tokenized ticker should trade at the close of a named catalyst window, derived from primary filings and order-book evidence rather than from the prevailing quote. Five passes run on every ticker; the output is a price with a distribution, not a point guess.

A signal is emitted when three conditions hold simultaneously: an edge against fair value, a dated catalyst inside the window, and a defined exit. Absent any one, CAL publishes the fair value only.

Every published signal enters the record. No curation step, no retraction, no re-basing of an entry after publication. Measurement runs from reference_entry to window close, or to the exit price when take_profit is reached sooner.

Each forecast is written on-chain at publication, before the catalyst resolves. Forecast, reference entry, exit, and close land on the same ledger the tokenized equity settles on, which is what makes the edge verifiable rather than asserted.