Skip to signal lessons
Trading Signals ExplainedMessages / Versions / Evidence

Conviction grades explained

A practical method note with explicit evidence limits.

Original edition: 9 October 2026. Expanded edition: 13 October 2026.

Educational explanations, fictional scenarios and personal worksheets. No live calls, paid-account test or provider performance certification.

A label needs a definition

An A-D grade is not a personalised size or a calibrated probability. Ask for the frozen rule and all grade-by-outcome counts.

Do not infer account advice

A grade says nothing by itself about account funding, existing exposure or a subscriber risk limit.

Source documents

Documentation establishes the stated mechanism, not tested trading performance. Checked 9 October 2026; a later version can differ.

Detailed research

A label needs its own calibration population

A conviction grade is not automatically a probability. Its meaning depends on the producing model, horizon, grade definition, eligible event population, outcome and calibration method. The same letter can have different distributions across products and periods. A financial interpretation requires complete observations, costs, frozen definitions and an independent evaluation, not just favourable examples. The fictional journal contains no conviction labels or calibration sample. Passing its identity and timing contract therefore cannot validate any grade. A reader should ask for the appropriate calibration source rather than reusing the journal's event count or a founder's separate historical return as supporting observations.

Filled interpretation finding

Imagine two invented products each use grade A. One labels an entry-recovery objective and another labels a longer-horizon research state. Even if their completed outcome shares happened to match, the economic and operational subjects would still differ. A fixed grade definition and actual eligible ledger are needed before comparing them. Entry candidates must retain their raw objective and frozen identity rather than being selected by exit-adjusted payoff. A later exit can validate, reject, size and measure overlap under the authorised process, but does not retroactively define what the original grade meant. This is a conceptual example, not a calibration claim for a named platform.

What the record still cannot answer

The lesson does not invent confidence intervals, observed return distributions or broker quantities. A-D labels attributed in a supplied product record remain unconfirmed here as independent probabilities. Actual scope, fees, quantity and execution eligibility require separate evidence. Missing calibration is a specific gap rather than proof that every grade is useless or fraudulent. Useful original explanations stay in place while the expanded guides show how to retain a stable source request. No complete form, timestamp, imported file or educational reasoning score becomes a financial certificate merely because its visible labels are consistent.