Execution modes
Turn a signal
into a complete strategy
A TradingView alert on its own is just a notification that something happened. Attach an execution model to it and it becomes a strategy: how the position is built, how it exits, and what it is never allowed to do.
- 4
- execution modes
- 1
- set of limits, applied to any of them
The idea
A strategy is a signal plus a model. An alert is a signal.
Most people wire an alert straight to an order and get exactly one behaviour: fixed size, in at the signal, out at a stop. That is the simplest execution model, and sometimes it is the right one.
The other three change how the position is built without changing what fires it. The same alert can enter once, scale in on a schedule, ladder across a range, or size up after a loss — and every one of them stays inside the limits on that account.
- Signal — straight in, with a stop and a take-profit
- DCA — scale into the position on your schedule
- Grid — ladder orders across a price range
- Martingale — size up after a loss, bounded by your caps
The modes
What each one is for.
All four are configurable, and all four are checked against the same limits.
Signal
One entry at the alert, with a stop and a take-profit. The simplest model and the easiest to reason about.
Direct
DCA
Scale into a position over several entries on your own schedule, rather than committing the whole size at one price.
Averaged
Grid
Place orders across a price range and work the range as it moves. Bounded by the position cap like everything else.
Laddered
Martingale
Increase size after a loss. The most dangerous model in the category, which is why the account ceiling still applies.
Bounded
The rule that does not bend
A sequence does not outrank a ceiling.
A grid that would breach your position cap stops adding at the cap. A martingale that would breach the daily loss limit does not place the next order. The mode describes intent; the limit decides what is allowed.
Point an alert at a demo account.
Attach a DCA or grid model to a signal you already use, run it on demo, and watch where the limits cut the sequence short.