Locking a Losing Trade (Same-Symbol Hedge): Why It Equals Closing, Only Dearer

Locking a trade means opening an opposite position of the same size on the same symbol while the original is losing. Say you hold a losing Buy of 0.10 lot on gold, so you open a Sell of 0.10 lot to “stop the bleeding” and plan to unlock at a better moment. Thai traders call it ล็อกไม้, and it is popular because it feels as if you haven’t really lost yet. This article works through the numbers to show that a locked trade is a closed trade in every way that matters, except that it costs more and asks you to make two harder decisions later.

The math of a lock: the total freezes

Every example here uses gold (XAUUSD) at 0.10 lot. One lot of gold is 100 ounces, so a 1-dollar price move makes or loses 10 dollars on 0.10 lot.

• Buy 0.10 at 4,000. Price falls to 3,980. Floating loss = (3,980 − 4,000) × 10 = −200 dollars
• Lock it with a Sell 0.10 at 3,980

Let P be any price after that.

• Buy leg = (P − 4,000) × 10
• Sell leg = (3,980 − P) × 10
• Total = (P − 4,000 + 3,980 − P) × 10 = (3,980 − 4,000) × 10 = −200

P drops out of the equation. Whether price goes to 3,900 or 4,050, the total stays at −200. Check by plugging in: at 3,900 the Buy is −1,000 and the Sell +800, total −200. At 4,050 the Buy is +500 and the Sell −700, total −200.

Your exposure to price is zero, exactly as if you held nothing. The −200 is not waiting for price to come back. It is already locked in. The only difference from pressing Close is that the Balance figure hasn’t moved yet.

A netting account proves it another way

MT5 has two account types.

• Netting: a symbol can have only one position. If you hold a Buy 0.10 and send a Sell 0.10, the platform closes the Buy on the spot and the −200 goes straight into your Balance
• Hedging: you can hold several positions on one symbol, including opposite ones, so the Buy and the Sell sit side by side (MT4 works this way)

Same order, same market. One account calls it “closing”, the other calls it “locking”. The risk left over is zero in both. Only the display differs. Some countries, the United States for example, do not let retail forex accounts hold opposite positions in the same currency pair, on the grounds that it does not reduce risk, it only adds cost.

A locked trade is a closed trade whose receipt hasn’t been printed yet, and it still pays rent every night

Costs that a plain close doesn’t have

One extra spread. The Sell you use to lock is a new trade and pays a full spread (the gap between the buy and sell price). If your broker allows the MT5 Close By command, you can close the two opposite positions against each other and save one spread, but if you unlock one leg at a time, you pay in full. See how to count this in Spread cost per trade.

Swap on both legs. Swap is the overnight financing charge. On many symbols, gold included, swap on both the Buy side and the Sell side is often negative. A locked trade can then lose money every night while having no chance to make any. Real values vary by broker and move with interest rates, so check the contract specification. Details are in Swap and overnight fees.

A wider spread deepens the floating loss. A Buy is valued at the Bid price and a Sell at the Ask price. When the spread widens, for instance at rollover or on news, a locked pair loses on both legs at once. If gold’s spread widens from 0.30 to 3.00 dollars, the floating loss temporarily deepens by (3.00 − 0.30) × 10 = 27 dollars.

Margin depends on your broker’s hedged-margin rule. Some charge full margin on both locked legs, some charge part of it or only one leg. Look at the Hedged margin field in the MT5 contract specification. If it is charged in full, locking doubles the margin in use, and swap shrinks Free margin further every night. A thin account left locked for a long time can still be stopped out. Try the numbers with the margin calculator and read Leverage, margin and Stop out.

Example: close now versus lock and unlock

Continuing the example, the floating loss is −200 at 3,980. The costs below are assumptions for the calculation, not any broker’s figures.

• Spread 0.30 dollars, so 3 dollars per round trip on 0.10 lot
• Swap per night on 0.10 lot: Buy −2 dollars, Sell −1 dollar (counted once per night, no triple swap)

The choices compared:

• A, close now: −200 and done, no extra cost (the Buy’s own spread is already paid in every case)
• B, lock and unlock: open a Sell at 3,980 (spread −3), stay locked for 3 nights (swap 3 × (−2 − 1) = −9), so the lock itself costs −12, then unlock depending on what happens
• C, close now and open a new trade: close at 3,980, then open a fresh trade at the same price and in the same direction as the leg B keeps after unlocking

Four scenarios, each worked with (exit price − entry price) × 10 × direction (Buy = +1, Sell = −1):

1. Well-timed unlock. Price falls on to 3,960. You close the Sell for (3,980 − 3,960) × 10 = +200, hold the Buy 2 more nights (−4), and price returns to 4,000, where you close the Buy at 0. B = 0 + 200 − 12 − 4 = +184 · C = −200 + (4,000 − 3,960) × 10 − 3 − 4 = +193

2. Unlock, then the trend continues. Close the Sell at 3,960 (+200), hold the Buy 2 nights (−4). Price drops to 3,920 and you give up on the Buy at (3,920 − 4,000) × 10 = −800. B = −800 + 200 − 12 − 4 = −616 · C = −200 − 400 − 3 − 4 = −607

3. Wrong leg released. Price bounces to 4,000, you close the Buy at break-even (0) with relief and keep the Sell 2 nights (−2). Price keeps rising to 4,020 and you close the Sell at (3,980 − 4,020) × 10 = −400. B = 0 − 400 − 12 − 2 = −414 · C = −200 − 200 − 3 − 2 = −405

4. Stuck locked for 20 nights. You can’t decide, and finally close both. The price part is still −200, plus costs of −3 − 20 × 3 = −63. B = −263 · C, with no new trade = −200

Scenario	A Close now	B Lock and unlock	C Close and re-enter	B − C
1. Well-timed unlock	−200	+184	+193	−9
2. Trend continues	−200	−616	−607	−9
3. Wrong leg released	−200	−414	−405	−9
4. Locked 20 nights	−200	−263	−200	−63

The last column is the answer. B equals C in every row, minus the swap paid while locked. The good result in row 1 doesn’t come from locking. It comes from a new trade opened at 3,960 that turned out right, which you could have done just as well after closing. Rows 2 and 3 show the other side: the leg left after unlocking is a new trade with no Stop loss, often bigger than you would normally risk because it was opened while you were rattled.

The mental trap: one decision becomes two

People don’t lock trades because the numbers say it pays. They lock because they don’t want to see the loss become real. Economists call this the disposition effect (Shefrin and Statman, 1985): selling winners too soon and holding losers too long. It sits on top of loss aversion (Kahneman and Tversky, 1979), where losing hurts more than an equal gain pleases. A lock feels like “I haven’t lost yet”, even though mathematically you already have. More in Trading biases and psychology.

The problem is that unlocking is harder than cutting a loss.

• Cutting a loss: one decision, out or not
• Unlocking: choose which leg to release and when, then decide where the remaining leg exits, all while the account is in the red and you are not calm

If you couldn’t make the first decision, giving yourself two more in the same state doesn’t make it easier. Row 4 of the table, staying locked indefinitely, is the most common real-life outcome, because every moment to unlock looks risky.

Our test results: refusing the first loss made things worse

We didn’t backtest locking itself, because the math already says the total is fixed. We did test its close relative, the rescue trade: when the first trade goes wrong, don’t cut it, add a trade twice the size in the same direction, then exit the whole group together. Both ideas come from the same wish, to avoid accepting the first trade’s loss.

In our backtest on gold (XAUUSD) M1, 2026, with real tick data and 0.10 lot per trade, the first trade entered at the 61.8% Fibonacci retracement of a large swing. If price went wrong as far as 78.6%, a rescue trade of double size was added.

Version	Hours (Thai time, UTC+7)	Trades	Net profit ($)	Profit factor	Drawdown ($)
Single trade with Stop loss	11:00–17:00	59	+4,002	1.60	1,479
Rescue, near group stop	11:00–17:00	97	−337	0.98	2,473
Rescue, far group stop	11:00–17:00	97	−3,844	0.79	5,678
Single trade with Stop loss	24 hours	225	+8,754	1.35	4,416
Rescue, near group stop	24 hours	355	+5,549	1.13	8,275

Every rescue version did worse than the single trade that simply accepted its Stop loss: less profit, and a Drawdown (the drop in equity from its peak) close to twice as large or more. The reason is that the level where the rescue was added had no edge of its own, so more lot there only added risk. Profit factor is total profit divided by total loss. All of these are backtest results. Past results do not predict future results, and they are not returns you should expect.

A real-money account in our trade-log data showed the same picture. The first trade was placed by hand on gold, rescue trades were added as price moved against it, and the whole group was closed at a small profit. In the first 2 days (6–7 July 2026), 15 trades made +21.02 dollars. As trading continued the account turned negative: the average winner was only 0.40 times the average loser, and there was a run of 5 losses in a row. The mechanics of grids and rescue trades are covered in Grid, Martingale and averaging down.

When hedging does make sense

A sensible hedge reduces the risk of something you intend to keep and can’t close, or would lose more by closing. It is not a way to freeze a trade that went wrong.

• Across instruments or markets: you hold physical gold or crypto long term and open a temporary opposite contract in a futures market around a big event, because selling the real asset is costly or slow. Know that the two markets don’t move in perfect lockstep (basis risk), and the hedge contract has its own carrying costs
• At portfolio level: you hold several pairs that all move with the dollar and reduce the combined risk with another position. See how to think about it in Correlation and portfolio risk
• Real business: an exporter expecting dollars in 3 months fixes the exchange rate in advance. This is where the word hedge comes from

If all you want is less risk on a trade you hold on one symbol, the cheaper and more direct way is to close part of it. Closing half gives the same risk as hedging half, without paying swap on two legs. See Scaling in and out, partial close.

Before you lock, ask: “if I had no position right now, would I buy at this price?” If not, close it

Common mistakes

• Believing a locked trade hasn’t lost yet. The total froze the moment you locked. Equity is your real money, not Balance. Read Drawdown: equity versus balance
• Forgetting swap on both legs. A lock held for a month pays every night while the account has no way to profit
• Releasing the winning leg and leaving the other without a Stop loss. The remaining leg is a new trade and needs a plan like any new trade. Size it from the Stop loss distance, as in Position size from Stop loss
• Locks on top of locks. Unlock, get it wrong, open another opposite trade. More open positions, more cost, and the loss hasn’t gone anywhere
• Using a lock instead of a Stop loss. If you know you struggle to cut losses, set the Stop loss when you open the trade and set a daily loss limit. See Daily loss limit and revenge trading