Stop-out is a sequence, and the order decides the loss
A close-out is a loop that liquidates one position, recalculates, and decides whether to liquidate another. Which position it picks first is a configuration default almost nobody has changed.
A stop-out is not a moment. It is a loop that liquidates one position, recalculates, and decides whether to liquidate another. The order in which it picks positions is a policy choice, it is usually left at whatever the platform ships with, and on a gapping market it is the difference between a closed account and a negative balance the broker absorbs.
The loop, precisely
Margin level is the ratio of what the account is worth to what it has committed. Every platform computes it the same way, and every close-out policy is a rule about what to do when it crosses a line.
margin_level = equity / used_margin × 100Two thresholds sit on that number. The margin call level is a notification. The stop-out level is where the platform begins closing positions on the client’s behalf. Under the retail rules derived from the ESMA 2018 intervention, the close-out obligation for retail CFD accounts sits at 50 percent of required margin, applied on a per account basis.
- 01Margin level crosses stop-outevaluated on every tick
- 02Select one positionby the platform ordering rule
- 03Close at the available pricenot the trigger price
- 04Recalculate equity and used marginrealised loss now in balance
- 05Still below? Repeatuntil the level is restored
Why the ordering rule is a risk decision
Closing a position does two things at once: it releases margin, which raises the ratio, and it crystallises an unrealised loss into balance, which lowers equity by nothing at the instant of closing but removes any chance of recovery. Different selection rules trade those off differently, and the common ones are not equivalent.
- Largest loss first. Removes the position doing most damage to equity, but if that position is also the one most likely to retrace, the account is stopped out of its own recovery. Common default.
- Largest margin first. Restores the ratio in the fewest iterations, which minimises the number of fills taken in a fast market. Often closes profitable positions.
- FIFO. Required in some jurisdictions. Bears no relationship to either margin efficiency or loss, so it usually needs the most iterations.
- Highest margin to loss ratio. Releases the most margin per unit of crystallised loss. Rarely the shipped default, and the closest thing to an optimal rule for the client.
The ordering rule is set once, in a configuration screen, and then decides the outcome of every stop-out the broker will ever run.
Hedged positions and the relief trap
Platforms grant margin relief on offsetting positions in the same symbol, because a long and a short together carry no directional exposure. The relief is correct while both legs are open. The stop-out loop closes one leg at a time.
The instant one leg of a hedge is liquidated, the relief disappears and the required margin on the remaining leg jumps. An account can therefore be pushed further below the stop-out level by the act of closing a position intended to rescue it, which is how a cascade accelerates rather than converges.
An account holds four positions and opens the session at 96 percent margin level. The market gaps against it. Stop-out is 50 percent. The only difference between the two columns is which position the engine selects first.
| Largest loss first | Highest margin to loss | |
|---|---|---|
| Positions liquidated | 4 of 4 | 2 of 4 |
| Fills taken in the gap | 4 | 2 |
| Realised loss crystallised | $41,200 | $23,850 |
| Margin level after loop | n/a, flat | 71% |
| Closing equity | −$3,400 | $13,950 |
| Absorbed by the broker | $3,400 | $0 |
Illustrative figures, chosen to show the mechanism. Your own numbers will differ.
Gaps, and why the trigger price is not the fill price
The stop-out level is evaluated against a price that exists. The liquidation is executed at the next price available, and in a weekend gap or a scheduled release those can be far apart. This is the single largest source of negative balances on a retail book, and it is not a failure of the stop-out logic; it is the logic working on a price series with a hole in it.
Negative balance protection for retail clients means the residual is the broker’s. That converts a client credit risk into a direct P&L item, which is why the ordering rule and the relief policy belong to risk rather than to whoever installed the platform.
What to actually watch
Before the event
- The distribution of margin levels across funded accounts, not the count below margin call. A book where 300 accounts sit between 60 and 80 percent going into a weekend is a different exposure from one where 30 sit at 52 percent.
- Concentration of those accounts in a single symbol, because they will all be liquidated into the same gap, against your own book, at the same moment.
- Open hedged pairs on accounts near the threshold, which are the accounts whose margin requirement will rise during their own stop-out.
After the event
Reconstruct the loop, not the outcome. For every stop-out, the sequence of positions selected, the margin level after each iteration, and the gap between trigger price and fill price. Aggregated over a quarter, that tells you whether the ordering rule is costing you money, and it is the only evidence that will survive a client complaint or a regulator asking why an account closed the way it did.
The data is all in the platform journal. It is not in any report the platform ships, because reconstructing a cascade requires holding position state, account state and the tick series together and replaying them in order. That is the work.