ORIN uses ordinary trading words with precise, code-enforced meanings. This glossary defines each one as the engine defines it, so a term on a verdict and the same term in a statistic cannot quietly mean two different things.
Reading a verdict
The vocabulary that appears on an analysis.
- Invalidation — the structural price at which the thesis is wrong. Stated before any target, everywhere it appears.
- Confluence score — how much of the evidence checklist agrees with itself, expressed 0 to 100. Not a probability.
- Calibrated confidence — the score checked against resolved outcomes. A different question from the score, and frequently a much lower number.
- Suspended — the newest bar is too old to stand behind. The read is not wrong; it is no longer current.
- No-setup — the read completed and refused, naming its single deciding reason.
- Signal — one of strong-buy, buy, no-trade, sell or strong-sell, derived deterministically from the engine’s numbers and never written by a language model.
- Aggregated leg — a higher timeframe derived by rolling up base bars rather than fetched separately. It can discount a score but never veto a trade.
Risk and entries
The vocabulary of the plan.
- R — one unit of the risk you defined, from entry to stop. Every outcome in the product is expressed in R rather than currency.
- Zone entry — an entry that waits at a resting limit at the edge of a pattern’s zone instead of entering at the analyzed close.
- Pending — a zone-mode take whose limit has not filled. No position exists yet.
- Lens — the rulebook an analysis is graded through: the platform default, your own strategy rules, or a partner pack. A lens can only ever be stricter.
The record
The vocabulary of outcomes and performance.
- Resolved outcome — the market reached either the stop or the first target. Used identically wherever the word appears.
- Unknown — the deciding bars fell outside the fetch window. Excluded from every statistic rather than counted as a loss.
- Shadow R — what a pass or a watch would have earned had it been taken. This is how declining a bad setup earns credit.
- Adherence — whether a take followed your rules or overrode them. Overrides are a no-trade signal, fired rule flags, or a low-grade read.
- Decision Quality Score — the average realised R of resolved high-grade takes minus resolved low-grade takes.
- Expected calibration error — how far the published confidence numbers sat from what actually happened.
- Heat — a live count behind a door: unread conversation in the Insider room, or setups the platform issued today. The only warm accent on the desk; it carries exactly those two numbers and means nothing else.
Updated Sep 1, 2026