Why a Forex Sell Order May Not Close Your Buy Position: Netting, Hedging and Spread Exposure

A sell confirmation does not, by itself, prove that an existing buy position has closed. The result depends on the account’s position accounting and the instruction’s intended effect: reduce an existing position, close a specified trade, or open another position. The useful evidence is the change in position IDs and actual units, not the colour of the trade button.

There is a second trap. Separate long and short positions can have the same net directional quantity as a smaller long position without having the same marked profit. Even equal long and short quantities do not necessarily freeze the displayed result: spread changes, financing and the way the positions are eventually closed still matter.

Check what the instruction changes, not just whether it buys or sells

An order is an instruction, a deal is an execution, and a position is what remains open. They need not map one to one. One order can generate several fills; a fill can open, reduce, close or reverse a position according to the applicable account and instruction rules.

Under a netting arrangement, an opposite execution in the same instrument can reduce the existing net position. An equal quantity can close it; a larger quantity can reverse it if the operation is allowed. Under an arrangement that permits separate opposite positions, a new opening instruction can instead leave the original long intact and add a short. That does not mean every sell in a hedging account must open another position: an explicit close or reduction instruction may have a different effect.

These are accounting arrangements to verify, not universal settings available to every forex customer. Account permissions, local restrictions and the product’s close instructions can differ. Do not infer them from the generic word hedging, or assume that an opposite order is equivalent to an instruction explicitly restricted to reducing an existing position.

Selling 6000 euros can leave two different ledgers

Use a hypothetical linear EUR/USD contract in a USD account. The original long is N = 10000 EUR, bought at the actual ask K = 1.1000 USD per EUR. At the comparison instant the bid B is 1.1020 and ask A is 1.1022. A sell of q = 6000 EUR is fully confirmed at H = 1.1020. The alternatives below are independent, not successive trades.

There are no additional commissions, financing, taxes, slippage or cash flows in this example. It is a position profit calculation, not a complete balance sheet for physically delivered euros and dollars. The long is marked at bid and a short at ask. Here one pip is 0.0001 USD per EUR, so the current spread is two pips, not 2% or two basis points of account return.

Reducing the long realises 6000×(1.1020−1.1000) = 12 USD and leaves 4000×(1.1020−1.1000) = 8 USD unrealised. The combined result is still 20. Looking only at the remaining floating figure would wrongly make the realised 12 appear to have vanished.

Opening a separate short leaves the original long’s 20 USD gain and adds a short marked at 6000×(1.1020−1.1022) = −1.20 USD. The combined mark is 18.80. Both alternatives have a net long quantity of 4000 EUR, but active gross quantity is 4000 in the reduced position and 16000 in the two-position ledger. Gross quantity is neither margin due nor a loss.

For 0≤q≤N, the combined realised and unrealised amounts are Tnet = q(H−K) + (N−q)(B−K) and Thedge = N(B−K) + q(H−A). Their difference is Thedge−Tnet = −q(A−B). It comes from marking the additional open short at its closeout side, not from an unexplained account-mode fee. Do not subtract the same spread again. An explicit reduction of the original trade in an account that allows it can produce the same 12 plus 8 as the netting example, given identical fills and costs.

If both bid and ask later rise by 0.0010 with the spread unchanged, each alternative gains 4 USD because both retain 4000 EUR net long exposure. Their totals become 24 and 22.80. This is a conditional comparison, not a claim that all account displays use the same marking convention. The formulas also do not cover q>N: selling 14000 against a 10000 long may leave a 4000 net short if reversal is allowed, rather than merely closing the old position.

Equal longs and shorts cancel a common price move, not every input

Start a separate example with the same 10000 EUR long at K = 1.1000. Now explicitly open and keep a 10000 EUR short at H = 1.1020. With both trades still open, their combined marked profit is T = N(B−K) + N(H−A) = N(H−K)−N(A−B). In this example, T = 20−10000×spread.

A parallel move of both quote sides leaves the total at 18 USD: the long gains 30 while the short loses 12. But widening the spread around the original midpoint of 1.10210 gives 18.50 on the long and −3.50 on the short, a total of 15. Net quantity remains zero and gross quantity remains 20000 EUR in all three states.

These are hypothetical valuation snapshots, not frequencies, price forecasts or worst-case limits. The cancellation requires equal actual units of the same linear contract and a common quote snapshot. It does not let you settle both positions at the midpoint. If the account uses a different documented valuation method, reconcile that method separately from the prices and charges of executable closure.

Financing, margin and closing one side remain separate questions

Add one independent financing illustration: a debit of 1.50 USD on the long and a credit of 0.40 on the short for the same charging period. The net cash adjustment is −1.10, so the baseline marked 18 becomes 16.90 after that adjustment. These amounts are invented inputs, not a quoted rate schedule. Equal units do not force the two cash charges to cancel, and an amount already included in the reported result must not be deducted again.

Zero net units do not establish zero required margin, unrestricted cash or immunity from forced closure. The actual treatment of offset positions must come from the account terms. Gross quantity, net directional quantity and required collateral measure different things; no margin percentage is implied by these examples.

Closing only one side changes the exposure. At the baseline quote, buying back the short realises −2 USD and leaves the 10000 EUR long. If bid then falls to 1.1010, the long has 10 USD unrealised profit; combined with −2 realised, the result is 8, down from 18. The earlier zero-net description no longer applies once the short is gone.

Paired closure is a specific operation, not an automatic consequence of hedging

Some eligible accounts support a paired close that matches opposite positions using their existing entry prices. If that precise operation is available and both positions in the equal-size example are successfully matched, the gross result is 10000×(1.1020−1.1000) = 20 USD. Separately selling the long at bid and buying back the short at ask at the baseline snapshot instead produces 18. The difference is one current spread on the matched 10000 EUR.

This is not a new arbitrage profit. The original fills already contain their applicable price sides, and the two entries were separated by a real assumed price difference of 0.0020. Matching cannot erase historical execution costs or financing. Nor does a new opposite opening order automatically request a paired close.

Availability, permitted sizes, fees, matching priorities and confirmation rules must be checked. A close-all button may process requests sequentially rather than as an indivisible operation; an error can leave an active leg. Different contracts, accounts or legal entities cannot simply be treated as one eligible matched pair. If the operation is unavailable, its hypothetical result cannot replace the actual separate-close result.

Audit positions, executions and the remaining protection together

First define the intended outcome: reduce, close, reverse or retain a separate opposite position. Record the account mode and the instruction’s permitted position effect. Do not change account settings merely to force a desired interpretation of an already completed trade.

Next compare the before-and-after position IDs, instrument, units and direction with confirmed executions. Check which IDs were opened, reduced or closed, the actual filled quantity and the remaining quantity. A submitted request, timeout or confirmation that an order was received is not itself proof that a position is closed; repeating an unresolved request can create unintended size.

Recalculate realised plus unrealised amounts with matching timestamps and the applicable quote sides. For that price-side check, see how bid and ask prices affect forex stop checks. Keep net units, gross units, margin and cash adjustments in separate fields, and include each commission or financing amount only once.

Finally inspect the remaining stop and target orders, their linked position IDs and quantities, and any failures in a sequence of closes. Do not assume protection automatically remains correct after a reduction or reversal. Without the instruction effect, position records and price inputs, the case remains unresolved: neither a platform error nor a safely finished trade has been demonstrated.