A trading signal is a structured statement of a trade idea. At minimum it names an instrument, a direction, an area where the idea is considered valid to enter, a level at which the idea is considered wrong, and one or more objectives. Everything else — commentary, charts, confidence language — is context around those five fields.

The five fields that define a signal

  • Instrument. The exact market, for example a currency pair quoted as base/quote.
  • Direction. Long means the idea profits if the base currency strengthens against the quote currency; short is the inverse.
  • Entry area. The price region in which the idea is considered available. Real execution can differ from the stated area.
  • Invalidation. The level at which the reasoning behind the idea no longer holds. This is the single most important field, and it is what a stop-loss order attempts to enforce.
  • Objective. Where the idea is considered complete. Multiple objectives imply a rule for what happens to the remainder of the position.

What a signal is not

A signal is not a prediction that a price will be reached. It is a conditional plan whose expected value depends on the size of the loss when it fails, the size of the gain when it works, and how often each occurs across a large sample. A single signal carries almost no information about any of those three things.

A signal also does not describe your risk. Risk is a function of position size, account equity and the distance to invalidation — three inputs the signal itself does not contain. Two traders acting on an identical signal can experience completely different outcomes.

Costs change the arithmetic

Spread, commission, financing on positions held overnight, and slippage during fast markets all reduce realised results relative to the levels quoted in a signal. Any assessment of signals that ignores these costs is incomplete.

Questions to ask of any signal source

  • Is every issued signal published, including the ones that failed and expired?
  • Are entries, exits and timestamps recorded at issue time rather than reconstructed afterwards?
  • Which pricing source is used, and does it match what a retail account would actually receive?
  • Are spread, commission, slippage and financing assumptions stated?
  • Is the sample period complete, or does it start at a convenient date?

If those questions cannot be answered, historical results from that source cannot be interpreted, and should not be treated as evidence of anything.