Capital at risk. Leveraged forex and CFDs are high-risk products and most retail accounts lose money on them. Nothing here is advice or a trading recommendation.
Ask a retail trader what their stop-loss does and you will usually be told it limits the loss to a set amount. It does not, quite. It is an instruction that becomes an order at a level you choose, and what happens after that depends on the market rather than on your instruction.
Understanding that gap is the difference between a risk plan that works and one that works most of the time, which is not the same thing.
The order types, and what each one gives up
Every order makes a trade-off between price certainty and execution certainty. You cannot have both, and each type picks a side.
Market order. Execute now, at whatever the price is. Maximum execution certainty, zero price certainty. Fine in a deep market, expensive in a thin one.
Limit order. Execute only at my price or better. Maximum price certainty, no execution certainty at all — if the market never reaches your level, or trades through it too quickly, nothing happens.
Stop order. When the market reaches this level, execute. The level is a trigger, not a price: once touched, the order goes to market and fills at whatever is available. Execution is near-certain, price is not.
Stop-limit order. When the market reaches this level, place a limit order at that price or better. This looks like the best of both and is the most dangerous of the four in exactly the situation stops exist for — in a fast move it may not fill at all, leaving the position open while the market runs away. A stop that does not execute is worse than no stop, because you believed you were covered.
Trailing stop. A stop that follows the price at a fixed distance as the position moves in your favour, and stays put when it moves against you. It manages the level automatically. It changes nothing about execution: once triggered it is an ordinary stop with all the same behaviour.
| Order type | Controls price | Controls execution |
|---|---|---|
| Market | No | Yes |
| Limit | Yes | No |
| Stop | No | Nearly always |
| Stop-limit | Yes | No, and least when it matters |
| Guaranteed stop | Yes | Yes, for a fee |
Why the stop does not fill at your level
Two separate mechanisms, and they get confused constantly.
Slippage is the ordinary case. Your stop triggers, the order goes to market, and the best available price has already moved past your level — because the market is moving and there is a delay, however brief, between trigger and fill. In a liquid pair at a busy hour this is often negligible. In a thin hour it is not, and our piece on what slippage costs puts numbers around it.
Gapping is the severe case. The price does not move past your level — it never trades at your level at all. It jumps. This happens across the weekend break, around scheduled announcements, and when something breaks. There is no fill at your level because there was no market at your level, and no order type other than a guaranteed stop can change that.
The January 2015 Swiss franc move is the reference case: stops placed sensibly filled hundreds of pips away, some accounts went past zero, and brokers themselves took losses. The Swiss franc shock of 2015 covers what happened and the regulation that followed.
The one order that really is guaranteed
Some brokers offer a guaranteed stop-loss order, which does exactly what its name says: it fills at the level you set regardless of gaps, because the broker takes the gap risk instead of you.
This is a real product and it is not free. Brokers charge for it either as a premium at the time of the order or as a wider spread on positions carrying one, and the charge is usually only applied if the stop is actually triggered — though this varies, so read the specific terms.
The decision is genuinely situational.
- Worth considering: positions held over a weekend, through a central bank decision or a major scheduled release, or in pairs and instruments prone to gapping.
- Usually unnecessary: intraday positions in deep major pairs during the London and New York overlap, closed before the session ends.
The useful way to frame it: an ordinary stop covers the normal case, a guaranteed stop covers the abnormal one, and the premium is the price of the abnormal case not being your problem.
What the regulator guarantees, and what it does not
Retail clients of UK and EU-regulated firms have two protections in this area, and both are frequently mistaken for something stronger.
The margin close-out rule requires the firm to close a retail client's open positions when net equity falls below 50% of the margin requirement. The wording matters: the firm must do this "as soon as market conditions allow" (COBS 22.5.13R). In a gap, market conditions do not allow it until after the gap. The rule is not a floor under your losses; it is a required action taken as soon as it can be.
Negative balance protection limits a retail client's liability for these products to the funds in that account (COBS 22.5.17R), and the guidance clarifies that those funds mean the cash in the account plus unrealised net profits on open positions (COBS 22.5.19G).
Read together: you cannot end up owing your broker money, and you can absolutely lose everything in the account. The protection sets the floor at zero, not at your stop. The margin close-out rule explained and ESMA rules explained go through both.
Placing them so they do their job
- Put the stop where the idea is wrong, not where the loss is comfortable. A level chosen to make the risk figure look acceptable is a level the market has no reason to respect.
- Size the position from the stop distance, never the other way round. Decide where the stop belongs, then work out the size that makes that distance an acceptable loss. How to calculate position size is the arithmetic.
- Keep tight stops out of the rollover window and away from the thin hours, where a spread widening alone can trigger them.
- Prefer stop over stop-limit for protection. Use stop-limit for entries if you like; a protective stop that can fail to execute is not protection.
- Do not move a stop further away. Widening a stop because the trade is going against you converts a planned loss into an unplanned one, which is the single most reliable way accounts are lost.
- Assume a gap once a year and size for it. If a gap of a few percent would be terminal, the position is too large whatever the stop says.
The honest summary
A stop-loss is a good tool used badly by almost everyone, because its name promises something it cannot deliver. It stops you from watching a losing position hoping. It does not stop the market from moving past your level while you sleep.
Treat it as an instruction that usually works well, occasionally works badly, and very rarely does not protect you at all — and size the position so that the rare case is survivable. That last part is the whole of risk management, and it is the only part no order type can do for you.
Regulatory detail in this article is from the FCA Handbook, COBS 22.5, read on 19 August 2026. Guaranteed stop availability and pricing differ by broker and instrument — check the specific terms before relying on one.
Frequently Asked Questions
Is a stop-loss guaranteed to fill at my price?
No. An ordinary stop-loss becomes a market order once the trigger level is reached, and it then fills at the best price available. If the market gaps past your level, the fill is at the first available price, which can be materially worse. Only a guaranteed stop, offered by some brokers for a fee, fills at the specified level regardless.
What is the difference between a stop order and a limit order?
A limit order specifies a price and will not execute worse than it, so it controls price but not certainty of execution. A stop order specifies a trigger and then executes at whatever is available, so it controls certainty of execution but not price. Choosing between them is choosing which of the two you are willing to give up.
How does a trailing stop work?
It moves the stop level automatically as the position moves in your favour, keeping a set distance behind the best price reached, and it does not move back when the price retreats. It locks in progress but does not change what happens on execution — once triggered it is still an ordinary stop, with the same slippage risk.
Is a guaranteed stop worth paying for?
It depends entirely on gap risk. For a position held across a weekend, a central bank decision or a major data release, the premium buys certainty about the worst case, which is the one scenario an ordinary stop cannot cover. For intraday trading in a liquid pair during the London and New York overlap, it is usually paying for a risk you are not meaningfully carrying.
