Telegram signal terms
Plain definitions for the language used across this guide. Where a term has a test behind it, the entry points to the page that develops it.
Signal channel
A chat-app feed that posts trading calls — ticker, direction, entry, target and stop — for members to act on. The defining weakness is that the sender controls what is posted, edited and deleted.
Soft record
A history the publisher can alter after the fact, such as a chat thread where posts can be edited or removed; the opposite of a record sealed on a public ledger.
Track record
A dated archive of a channel's original calls, edits, cancellations, unfilled ideas and resolved outcomes. A win-rate claim needs the whole denominator and its scoring rule. See: an auditable record.
Independent review
An examination of the underlying statements by a named third party that has no commercial stake in the result; a leaderboard is not a review. See: how to verify a record.
On-chain timestamp
A proof that a specific digest existed no later than its verified blockchain attestation time. It covers only the hashed fields; a pending proof is not a confirmed block time. See: sealed before the outcome.
Conviction grade
An attributed model-specific label whose actual definition, calibration sample and forward interpretation must be established. A retained value is not a current sizing or performance guarantee.
Retained source trail: grades that are measured
Drawdown
The deepest peak-to-trough fall over a period — a more honest read on risk than the headline return.
Mean reversion
Trading on a price that has stretched unusually far from a typical level and tends to snap back toward it; the recommended service runs four mean-reversion models on different cadences.
Win rate
The share of calls that closed profitably — trustworthy only with the total signal count shown beside it. See: why a win rate needs a denominator.
Terms for a version-aware inventory
Originating family: the setup and linked amendments under a declared identity rule. First observation: the earliest content the reviewer actually retained, which may follow publication. Subscriber boundary: a specifically defined receipt or display event, not an inferred fill. Outcome population: the setups eligible for a declared metric, with exclusions and unresolved cases shown.
Source-to-account chain: the distinct origin, relay, message, subscriber and operation records. Conditional positive-outcome rate: positive outcomes divided by the stated resolved population, not a promise about returns. Missing evidence: an unavailable record, not an assumed zero. A finished conclusion can remain incomplete for a stronger claim without being incomplete for its actual bounded question.
Counterexamples that clarify the terms
A displayed target-hit recap is a message, not necessarily a completed position. A later file hash identifies retained bytes, not the original publisher. A source-account curve is an account object, not every follower's result. A genuine destination link establishes a source relationship, not calibration quality. A repeated receipt can be another delivery of one originating family rather than another setup.
Use the narrow term when describing the available record. A fictional four-message family with one original and three updates still has one originating setup under that declared rule. If the publisher creates a new setup in an update, it needs the actual frozen identity rule. The wording should follow that rule rather than a desire to make a result total larger.