Should an EA use a fixed take-profit or a trailing stop?

It depends on the shape of the strategy. Mean-reversion and breakout systems have a natural level price should reach, which a fixed target expresses well. Trend systems have none, and a fixed cap cuts off the moves that pay for the losers.

What is a good take-profit distance?

One that gives a defensible ratio with the stop and stands well clear of the spread. The 18 uniform-target records here ship 25 to 300 pips, median 107, with a median declared ratio of 1.5 times the stop.

Why did my take-profit not trigger when price reached it?

Because a long closes at the bid and a short at the ask. On a short the ask has to reach the target, even though the chart usually draws the bid.

Does a take-profit work if my terminal is closed?

Yes. Once accepted it sits on the broker's server and executes there, which is why it beats closing at a level in EA code when the terminal may not be running.

How often does an EA's take-profit actually close the trade?

It depends on what else can. Across 3,965 winning backtest trades on 18 published records, 42.02% closed within 2% of the shipped target. On the 7 builds with no trail, timed close or break-even move, 97.87% did.

Strategy

Take-profit

also: TP, profit target, target

A take-profit is a price level attached to a position at which the broker closes it automatically to bank a gain. It lives on the broker's server like the stop-loss, and the trader decides it before the trade opens.

Updated 2026-09-18 · Reviewed2026-09-18

At a glance

Term type: MT5 platform
Difficulty: Beginner
Used in: Strategy design · MT5 operation · Risk management
MQL5 function: CTrade::PositionModify(ticket, sl, tp)

In plain English

The price at which the trade has been right enough. The stop says how wrong you will let it be; the target says how right is enough.

Why it matters

The target is where a strategy collects its edge, and the number traders move most often after the fact. Fixed at entry, it turns the risk/reward ratio into a claim you can test.

With the stop it fixes the breakeven win rate: 1 ÷ (1 + target ÷ stop). Move either number and the bar the strategy has to clear moves with it.
It executes on the broker's server, so a target hit overnight banks the gain whether or not the terminal is running.
Its distance sets the character of the record. Near targets win often and small; far ones wait through long flat stretches for the trades that pay for the rest.

In MetaTrader 5

Where it appears in MT5

New Order dialog (F9) → Take Profit — set at entry, as a price
Market Watch → right-click the symbol → Specification — the Stops level row is the nearest distance the broker accepts for a target too
Strategy Tester → Inputs — an EA's target is an optimisable input, and one of the easiest to overfit

How EAs use it

Sent in the tp field of MqlTradeRequest at entry, as a price. The EA converts its intended distance into a level itself.
Attached after the fill with TRADE_ACTION_SLTP when the strategy needs the real entry price first.
Derived from the stop when the strategy speaks in ratios. The 18 uniform-target records here declare 0.88 to 4.19 times the stop, median 1.5.
Rarely the only exit. 11 of those 18 also ship a trailing stop, a timed close or a break-even move, and on those builds the target closed 25.85% of the winners.

Typical settings

Setting	Typical value	Note
Ratio-based target	1.5–3 × the stop distance	The most common EA formulation; the shipped median here is exactly 1.5.
Shipped target, 18 uniform-target EAs here	25–300 pips, median 107	What these builds send. Three ship a target narrower than their stop — 100 against 114.
ATR multiple	2–4 × ATR(14)	None of the 17 ATR-named inputs in this catalogue controls a target; all of them gate an entry.

Common operational problems

The bid reaches the target but a short closes at the ask, so a target set exactly on a level can sit unfilled by the width of the spread.
Points and pips collide on 5-digit quotes: a 100-pip target goes out as 100 points, and the test prints a suspiciously high win rate.
The server rejects a target inside the stops level and the position runs with no exit on the profit side. Read POSITION_TP back after the request.

Related MT5 functions

MqlTradeRequest.tp: The target price sent with the order; zero means no target.
CTrade::PositionModify(ticket, sl, tp): Adds or moves the target on an open position.
PositionGetDouble(POSITION_TP): Reads back the target actually in force.

Example

One entry rule, two target distances, two very different hit rates to clear.

Stop distance: 30 pips
Target A: 30 pips (1:1); Breakeven win rate 50% before costs.
Target B: 90 pips (3:1); Breakeven win rate 25% before costs.
Spread paid on each trade: 1.5 pips; 5% of target A, 1.7% of target B — the same toll, a different share.

Target A wins far more often and hands most of it back in costs and in losses the size of the wins. They are not the same product.

Calculation 1 ÷ (1 + 1) = 50% · 1 ÷ (1 + 3) = 25%

Result Two products from one entry rule

How it is used

Read the target beside the stop and beside the realised results. What the EA achieved says more than what it aimed for.

Range	What it means
Average win well below the configured target	Something else closes the trade first. On the 11 records here that declare another exit, the median win runs from 0.5% to 45% of the target.
Average win close to the target	The target does what it says. On the 7 target-or-stop-only records, 96.43% to 100% of winners closed within 2% of it.
Target within a few times the spread	Costs take a large share of every win; the strategy needs an exceptional hit rate to survive them.
No target, exit by trailing stop only	Legitimate for trend strategies, but the risk/reward has to come from realised trades, not a declaration.

Compute the realised ratio from the ledger — average win over average loss in money — and set it beside the declared one. On 15 of the 18 records here the money ratio comes in below it.
Add the spread to the target distance when the target is small. A 10-pip target on a 1.5-pip spread gives up 15% of the gain before anything else.
Optimise the target on a different window from the entry; fitted together, the pair share the same noise and fail together.

Every EA here publishes its exit rule and its trade ledger; the figures on this page come from those backtest ledgers (EXP-TARGET-REACH-001).

Common mistakes

Moving the target after entry

Extending a target because a trade is going well rewrites the ratio you took the trade on. The honest risk/reward is the one that existed when the position opened.

Ignoring which side of the spread closes the trade

A long closes at the bid and a short at the ask. A target placed exactly on a round level can sit unfilled while the chart, which draws the bid, appears to have touched it.

Reading the declared ratio as the realised one

On the 7 builds here where the target closes 96–100% of winners, the money ratio still landed below the declared ratio on 6 of 7, lowest 0.90×. Spread, commission and swap take that much; a trail or a break-even move takes far more.

In depth

What 3,965 published winning trades did against the target meant to close them

Every listing ships its closed-trade list, so each winning exit sits beside its target. Eighteen of the 31 records send one target distance on every order.

18 uniform-target records, 3,965 winning trades	Figure
Winners closing within ±2% of the shipped target	42.02% (1,666)
Winners closing short of it	56.77% (2,251)
Winners closing past it	1.21% (48)
7 records with no trail, no timed close, no break-even move	97.87% at target, lowest 96.43%
11 records that declare one of those	25.85% at target
Declared ratio, target ÷ stop	0.88–4.19, median 1.5
Money ratio below the declared ratio	15 of 18

Rows four and five split row one. Three trailing builds ship a 100-pip target behind a 114-pip stop and a 34-pip trail; the trail closes 83–86% of their winners at a median 21–24 pips, and their money ratio lands at 0.32–0.36 against a declared 0.88. Four break-even builds post a median win of exactly 1 pip. Row three is the favourable side of the fill: a gap through the level fills at the next price, the largest 112.6 pips against a 67-pip setting.

On a build that declares a trailing stop or a break-even move, treat the target as a ceiling and read the realised average win from the ledger. Pair it with the stop-loss that sets the other half of the risk/reward ratio, weigh it against a trailing stop when the strategy has no natural level, and read the exit rule each expert advisor states in the catalogue.

Frequently asked questions

Should an EA use a fixed take-profit or a trailing stop?: It depends on the shape of the strategy. Mean-reversion and breakout systems have a natural level price should reach, which a fixed target expresses well. Trend systems have none, and a fixed cap cuts off the moves that pay for the losers.
What is a good take-profit distance?: One that gives a defensible ratio with the stop and stands well clear of the spread. The 18 uniform-target records here ship 25 to 300 pips, median 107, with a median declared ratio of 1.5 times the stop.
Why did my take-profit not trigger when price reached it?: Because a long closes at the bid and a short at the ask. On a short the ask has to reach the target, even though the chart usually draws the bid.
Does a take-profit work if my terminal is closed?: Yes. Once accepted it sits on the broker's server and executes there, which is why it beats closing at a level in EA code when the terminal may not be running.
How often does an EA's take-profit actually close the trade?: It depends on what else can. Across 3,965 winning backtest trades on 18 published records, 42.02% closed within 2% of the shipped target. On the 7 builds with no trail, timed close or break-even move, 97.87% did.