Just as an entry Trigger stages the open, an exit trigger stages the close. Your exit rule is the setup ('the fast EMA re-crossed the slow one, I want out'), and the exit trigger is the confirmation ('but only once a bearish candle commits'). The exit rule arms the trigger; the trigger closes the trade.
The exit trigger reuses the same seven leaf kinds as the entry side — Pullback, rule, time_delay, confirmation_candle, Retest, divergence, and liquidity_sweep — so everything you learned about entry confirmations transfers directly. What changes is the shape of the block and the direction it is evaluated in.
One flat block, not two
An entry trigger is configured per direction (long: and short: sub-blocks), because a long setup and a short setup are different trades. An exit trigger is a single flat block — there is only one open position to close, so there is only one block. It carries max_wait_bars, a combine of AND or OR, an on_expiry mode, and a rules: list of leaves.
The block is evaluated against the opposite of the open position's direction, because closing a position is the reverse trade. Closing a long is a sell — so a confirmation_candle exit looks for a bearish_engulfing. Closing a short is a buy, confirmed by the same block read in the bullish direction. You write the block once; the engine flips it to match whichever side you are holding.
on_expiry: close vs. hold
If the confirmation never arrives within max_wait_bars, on_expiry decides what happens to the armed exit. close honours the original exit signal anyway — the soft exit fired, so close at expiry rather than silently swallow it. hold abandons the armed exit and keeps the position open, betting the confirmation's absence means the move still has legs. Use close when the exit rule is your real reason to leave and the trigger is just for a better fill; use hold when the confirmation itself is the only thing that justifies the exit.
Hard guards always bypass
This split is the whole point. A Stop loss is a hard guard: it is the line past which the trade is wrong, and nothing should ever stand between price reaching it and the position closing. A Structural stop behaves the same way. The exit trigger lives entirely on the other side of that line — it can refine when you take a discretionary profit or bail on a thesis, but it has no say over invalidation.
A worked example
Here the soft exit (the EMA re-cross, defined in the exit: block) arms the trigger, which then waits up to 20 bars for a committing candle before closing:
exit_trigger: combine: OR max_wait_bars: 20 on_expiry: close rules: — kind: confirmation_candle, pattern: bearish_engulfing (closes a long) — kind: confirmation_candle, pattern: bullish_engulfing (closes a short)
If neither candle prints inside 20 bars, on_expiry: close closes the position anyway — the exit rule did fire, and a fired exit should not be quietly forgotten. Wire one up in the Strategy Designer, which now exposes the exit-trigger block alongside the entry trigger.
If entry triggers are still fuzzy, read that lesson first — the exit side is the same machinery pointed the other way.
Reflexion
For a trade you hold, what is the difference between the price that proves you wrong (a hard stop) and the event that simply tells you the edge is spent (a soft exit)? Which of the seven kinds best captures that second event?