Glossary · Orders & execution

OCO order in gold trading

An OCO order links two pending gold orders so that when one fills the other is cancelled automatically, used for range breakouts in either direction.

Also called One cancels the other, OCO bracket, Order-cancels-order

An OCO (one cancels the other) order is a pair of linked orders: as soon as one is executed, the other is cancelled. It is used when a trader wants to act on whichever of two outcomes happens first, without ending up in both.

How it applies to gold trading

The classic gold use is a range breakout: a buy stop above the Asian range high and a sell stop below its low. Whichever side breaks first opens the trade and the other order disappears. Without OCO, a whipsaw that breaks one side and then the other can fill both orders, leaving a hedged or doubled position.

Not every platform supports OCO natively. MT4 and MT5 do not have a built-in OCO type; traders use an expert advisor or cancel the second order by hand. Some futures and web platforms offer it as a standard order type. Stop-loss and take-profit orders attached to one position already behave as an OCO pair, since one closing the trade removes the other.

Worked example

Example (hypothetical numbers): the Asian range is 3,996 to 4,008. You place a buy stop at 4,010 and a sell stop at 3,994 as an OCO, each 0.20 lots with a $6 stop. London breaks higher and fills the buy at 4,010; the sell stop is cancelled. Your risk is $6 × 100 × 0.20 = $120 on one trade, not two.

A common mistake

Placing both breakout orders without linking them and forgetting the second one after the first fills.

See also: Pending order, Bracket order, Asian range, Breakout.

Frequently asked questions

Does MT5 have OCO orders?

Not as a built-in type. Traders use an expert advisor or cancel the second pending order manually after the first fills.

Is an OCO order safe around gold news?

It prevents both sides filling, but a news spike can still fill one side with slippage and then reverse. Check the straddle risks before using it on data releases.