How we audit signal groups

• We read each tracked group's channel as it posts. A signal is stored with the time we received it, so it cannot be back-dated.
• Each group has its own parsing rules that pull out the symbol, side (long or short), entry, stop-loss, take-profit targets and leverage. Where a group has it switched on, messages the rules cannot read go to an AI parser; chart images can be read the same way for groups that post signals as pictures.
• A message that repeats a signal recently recorded for that group is ignored, so a re-post does not count twice.
• Every recorded signal stays in the record. A group cannot edit or remove a signal once we have it.
• Every closed or canceled signal is published in full on the group's page, with no time limit, so any score can be checked against the signals behind it. Signals still live are shown in time on paid plans and after 24 hours on the free plan.

• A new signal starts as pending. It becomes open once the market price trades between the entry and the stop-loss, that is, at the entry or anywhere past it on the stop side.
• If that does not happen within 24 hours the signal is canceled. Canceled signals are not scored, but they lower the group's fill rate.
• Prices are an average across the exchanges we monitor, sampled throughout the day, not one exchange's wick.

• An open signal closes at the stop-loss the first time price reaches it, or when every take-profit target has been reached.
• Take-profits reached along the way are recorded (they count towards the win rate), but the result is taken at the closing price. A signal that hits TP1 and then the stop is scored as a stop-out.
• On each price sample the stop is checked before the targets, so when both are crossed at once we assume the worse outcome.
• The stop is never moved: no break-even, no trailing, no partial exits. Groups that tell their members to do those things later are scored on the levels they originally posted.

Win rate alone is misleading: a group can win 80% of the time and still lose money if its losses are much larger than its wins. So we score every closed signal in R, the result measured against the risk the signal itself defined.

• A stop-out is about −1R. A signal that gains twice what it risked is +2R.
• Leverage is divided out. A 20x signal and a 2x signal with the same price levels score the same, so posting high leverage cannot inflate a result.
• Expectancy is the average R per closed signal: what you would have made or lost per unit risked, on average. We also show a 95% range around it, which narrows as the sample grows.
• Max drawdown is the largest fall in cumulative R from a peak, in the order signals closed. It tells you how deep a losing run following the group would have been.

Sample size matters. We show expectancy and drawdown only once a group has at least 100 closed signals. Below that a group's page says insufficient data: a few dozen trades are not enough to tell skill from luck.

Tiers. The tier list and the provider badge group ranked providers by expectancy: S from +0.50R, A from +0.30R, B from +0.15R, C from 0R and D below zero. Win rate is shown alongside for reference, never used to rank.

• We don't pick which signals count. Every parsed signal from a tracked group is in its record.
• We don't apply anyone's after-the-fact updates, "close early" messages or re-entries to the score.
• We don't model fees, funding or slippage, so real results following any group will be somewhat worse than its score.
• We don't hold your funds. Automated trading, if you use it, runs through API keys on your own exchange account.
• We don't give investment advice. The audit describes what happened; it is not a recommendation to follow anyone.