Telegram Signal Trial Ledger: Two Records, One Decision
Keep every publisher call beside receipt, actionability, costs and your own account outcome during a signal-channel trial.
Create two linked rows per call
The publisher row stores the original channel message: identifier, timestamp, ticker, direction, entry rule, initial stop, target, grade, later edits and stated outcome. The subscriber row stores when you received it, the venue and order type available to you, whether you acted, your average fill, fees and eventual result. Link the rows with one call ID. A channel outcome and a subscriber outcome can both be honestly reported without pretending they are the same trade.
Write 'missing' for an absent stop or target. Do not infer it from a chart. Keep a screenshot or export reference for every version of the message. If a provider supplies an independent timestamp, add the digest scope and confirmed attestation time as another column, not as a replacement for the two rows.
Protect the denominator
Log every formal call in the observation period, including ones you could not act on, cancelled calls and losers. Record why each was skipped: outside trading hours, stale at receipt, unavailable instrument, exceeded risk budget, missing terms or discretionary choice. Do not remove these from the publisher count simply because they never reached your account.
Keep a separate count for informal commentary so a casual market observation is not quietly scored as a formal signal only when it wins. Ask the provider how it identifies a new call versus an update. Apply that rule to the complete trial, not one favourable week.
Score only under frozen rules
Before the trial, decide how to score a range entry, partial target, moved stop and expired call. If the channel changes those rules, record the change date and do not re-score earlier messages retrospectively. A win rate without these conventions is a marketing summary, not a reproducible statistic.
For an illustrative spot buy with reference entry 100, stop 96 and target 108, the published plan has 4 units of reference risk and 8 of possible upside. A subscriber fill at 100.8 has 4.8 of risk to the same stop and 7.2 to the same target before fees. The difference is a teaching example, not a result from a channel.
Decide whether the service earns its cost
At the end, report the full call count, how many were actionable for you, the observed delivery delays, unresolved edits, missing fields and total subscription and trading costs. A useful conclusion may be 'delivery fits my schedule but the outcome history is unverified' or 'the archive is checkable but my venue cannot execute these prices.' Both are more informative than a single star score.
Use the message-audit procedure for edits and the delivery test for timing. An external receipt can support a specified publication claim, while your ledger answers the separate question of whether the service was usable and valuable to you.