How a Trading Signal Should Be Scored

Signals are locked at publication and scored against real OHLC candles. When target and stop fall inside the same candle we record a loss, because the candle cannot show which came first — a rule that can only make our numbers worse. Signals reaching neither level within their horizon expire and are scored on close-versus-entry.

Scoring a trading signal after the fact sounds trivial: did it hit the target or the stop? In practice almost every ambiguous case can be resolved in a way that flatters the result, and the choices a provider makes there matter far more to the headline number than the strategy itself.

Here is exactly how we resolve each one, and why.

The signal is locked when it publishes

Entry, target and stop are written at the moment the signal fires and never touched again. No moving a stop after the fact, no re-entry, no “we would have exited here”. If the levels were wrong, the result is wrong, and it counts.

The same-candle problem

The hard case: within a single candle, price touches both the target and the stop. The candle records a high, a low, an open and a close — but not the order in which the high and low occurred. So it is genuinely unknowable from that data whether you would have been stopped out or paid out.

The alternative — scoring it a win, or discarding it — would raise our win rate at the cost of making it untrue. A provider that quietly resolves these in its own favour can add several percentage points to a headline figure without changing a single thing about its strategy.

Signals that do neither

Plenty of signals never reach either level. Rather than drop them — which would silently remove every mediocre outcome from the record — each has a fixed horizon:

• Daily signals: 30 days.

If neither level is reached in that window, the signal is marked expired and scored on its signed close-versus-entry return. A signal that drifted 0.4% in the right direction is recorded as +0.4%, not thrown away. Expired signals are excluded from the win rate (they neither won nor lost) but included in average return and expectancy, so they still drag on the numbers exactly as they would drag on an account.

Publishing how far it ran either way

For every resolved signal we also record the furthest it moved in your favour and against you before resolving. A losing trade that first ran 3% in profit tells you something a bare “loss” does not, and so does a winner that spent the whole time underwater first. Both figures appear on every asset page.

This is the part almost nobody publishes, because it exposes how much of a result was luck of sequencing rather than edge.

How to check any provider

• Are losing signals published, or only winners?
• What happens when target and stop fall in the same candle?
• Are signals that hit neither level counted, or quietly dropped?
• Are entry, target and stop fixed at publication, or described afterwards?

Our answers are above; the full rules are in the methodology, and the resulting numbers — including the losses — are on the track record.

Educational content, not financial advice. Trading carries risk, including loss of capital.