What happens when there is no price at all?

Every risk calculation in Unit 3 rested on an assumption stated once and then quietly relied upon: that prices move continuously, so a close-out at a defined level executes near that level. Markets do not always oblige. When they do not, leverage stops being a multiplier and becomes an open-ended obligation.

What a stop order actually is

A stop is an instruction to transact at the next available price once a level trades. It is not a promise to transact at that level. The same is true of a broker's automatic margin close-out. If there is no available price for a distance, the fill happens on the far side of that distance.

Three ordinary mechanisms produce that outcome:

• Slippage. In a fast market, the price at which your order reaches a liquidity provider is not the price you saw. Normal, small, and it accumulates.
• Last look rejection. From Unit 1: a provider may decline in milliseconds, and by the time the order is re-routed the level has moved.
• The weekend gap. The market closes Friday at 17:00 New York and reopens Sunday at 17:00 New York. A stop resting through a weekend event executes at the Sunday reopening price, which may be far from the level.

The documented illustration — 15 January 2015

The Swiss National Bank had maintained a floor of 1.20 francs per euro since September 2011, defending it publicly and repeatedly. On the morning of 15 January 2015 it removed the floor without warning.

EUR/CHF collapsed within minutes. Some venues printed in the mid-0.80s at the extreme — roughly a 30% move — before the pair settled closer to parity by the close. For a stretch of minutes there were effectively no bids: not a bad price, no price.

The arithmetic is the whole lesson. Retail leverage on major pairs at the time was commonly 50:1 or higher, meaning 2% margin or less:

• A 30% adverse move against a 2% deposit is fifteen times the deposit
• At the 30:1 cap the EU and UK now impose (3.33% margin), the same move is nine times the deposit

Documented consequences across the industry:

• Alpari UK entered insolvency.
• FXCM required an emergency rescue loan of around USD 300 million.
• IG Group disclosed client losses in the region of £30 million.
• Interactive Brokers reported client debit balances of roughly USD 120 million.
• Saxo Bank re-priced client fills after the fact, on the basis that the original prints were not executable.
• Multiple firms pursued clients for negative balances; some later waived them.

Note what did not happen. No stop failed to trigger. No system broke. There was simply no price between the trigger and where the market next traded — and every risk model that assumed otherwise produced a number that turned out to be fiction.

Why a peg is the sharpest version of this

A managed or pegged rate compresses volatility to almost nothing for years, which makes every backward-looking risk measure — historical volatility, value-at-risk, "maximum observed drawdown" — read as reassuringly small. The risk was not absent. It was stored, and it was released in a single print.

The general form: quiet is not the same as safe, and any risk figure estimated from a period of official suppression is measuring the suppression, not the risk.

Negative balance protection, precisely

Several jurisdictions now require negative balance protection for retail clients: the account cannot go below zero, and the shortfall becomes the firm's problem. Three precise points:

1. It is jurisdiction- and entity-specific. The same brand may operate one entity with it and another without.
2. It caps the loss at the account balance, not at the position's margin. Every other cent in the account is still available to absorb the loss.
3. It does nothing about the loss itself. Protection against owing more than everything is not protection against losing everything.

Guaranteed stops, where offered, fill at the specified level regardless of gaps, and are charged for through a premium or a wider spread. That they are priced is the useful signal: the market quantifies gap risk, because gap risk is real.

In the data

The weekend gap is literal in hourly prices. Below is euro-dollar across the last weekend of September 2026: the last hour with a price is Friday 21:00 UTC, and the next one is Sunday 23:00, forty-nine hours with no price at all.

Friday's last close was 1.13921 and Sunday's first open 1.13804, a gap of about 12 pips that no order could have traded inside. That was a quiet weekend; the January 2015 session below needed no weekend at all to jump. And because a currency pair has no counted volume, hourly FX prices cannot tell you whether an empty hour means no trading or missing data.

Try it now

1. EUR/CHF over the longest window this page holds is below, as candles. Navigate to January 2015 and Measure that session from its open down to its low. The move lives in the low, far under both the open and the close — a stop placed anywhere between them was never going to be filled there.

2. Then look at the twelve months before it, and note how little the pair had done. Compute what that one day did to a hypothetical USD 5,000 account holding one standard lot at 30:1. State the answer as a number, including the sign.
3. Find the largest weekend gap you can in the last two years of EUR/USD, below as daily candles: Measure from each Friday's close to the following Monday's open. Then compute the leverage at which that single gap alone would have consumed a full deposit. That number is the honest answer to "how much leverage is safe here" — arithmetic, not advice.