The referendum on 23 June 2016 was known about for months. The date was fixed, the question was fixed, and the two outcomes were obvious. None of that made it tradeable in the way a scheduled data release is tradeable, because the distribution was not a curve around a central estimate. It was two lumps a long way apart, with almost nothing in between.
That shape explains everything that happened that night. Options pricing on GBP had been showing elevated implied volatility for weeks. Brokers had been raising margin on sterling pairs and warning clients that spreads would widen. The market was not surprised by the event. It was surprised by the answer.
How the night actually unfolded
Cable rallied into the close of European trading as late polling and betting markets pointed towards Remain. It reached the mid 1.50s. Then the count began, area by area, through the small hours UK time. The early results from north east England came in with larger Leave margins than models had assumed, and sterling started giving ground in steps rather than in a slide, because each declaration was a discrete piece of information arriving at a specific minute.
By the time the overall outcome was clear, GBPUSD had fallen into the low 1.30s, the largest one-day fall the pair had recorded. GBPJPY, which carries the volatility of both legs, moved further. Equity index futures fell hard overnight and the UK index recovered much of its loss within days, while sterling did not. That divergence is worth remembering: a shock can be repriced quickly in one asset and permanently in another.
Why the fills were so bad
Between roughly 22:00 and 06:00 UK time, the deepest sessions are closed. The Asian session carries sterling but not in size. On a normal night that thinness is invisible because nothing happens. On that night, every liquidity provider was quoting defensively, widening its two-way price to protect against being picked off by whoever saw a declaration first.
The result was a market that gapped between prices rather than trading through them. Traders with stops inside the range got filled well past their level. Traders with limit orders sitting in the path sometimes did not get filled at all, because price traded through the level without a resting counterparty at that price. This is the same mechanism as the SNB floor removal in 2015, though it came from a scheduled event rather than an unannounced policy change.
Widened spreads and raised margin around a known event are not a broker penalising clients. They are the broker passing on what its own liquidity providers are charging it, plus a buffer against gap risk it will otherwise carry itself.
What positioning for a binary event should look like
The common retail approach was to pick a side and use a wider stop than usual. That does not work on this kind of event, for a specific reason: a wider stop does not protect you if the price never trades near your stop level. It just books a larger loss when the fill arrives. Widening the stop while keeping the same lot size increases risk instead of reducing it, which is the opposite of the intention.
The approaches that survive a binary event are duller. Reduce size to the point where the worst plausible gap is an acceptable loss, and accept that this makes the trade small enough to barely matter. Or stay flat through the event and trade the reaction afterwards, once two-way pricing has returned and there is structure to work with. Or, for firms rather than individuals, hedge the exposure in an instrument with a defined maximum loss. Retail CFD traders generally do not have that third option, which is why the first two are the practical menu.
The reaction trade deserves more attention than it gets. On 24 June there was a full trading day at levels nobody had seen for decades, with normal liquidity restored by the London open and a clear directional context. That is a far better risk environment than the count itself, and it required no view on the referendum at all. The same logic applies to smaller scheduled events covered in news trading.
There is a second, quieter lesson in how the two sides of the trade were financed. Traders who were short sterling into the count and correct made less than the arithmetic suggests, because they entered on a widened spread, paid a raised financing charge on the position, and in many cases could not add to a winner while margin requirements were elevated. Being right on a binary event pays worse than the chart implies. Being wrong pays exactly what the gap says. That asymmetry is the real reason to reduce size rather than to pick a side with conviction.
What carried forward into how firms operate
Brokers changed their event playbooks after that night. Margin requirements on affected instruments are now routinely raised days in advance of referendums and contested elections, sometimes doubled. Some firms move affected instruments to close-only for a window around the event. Client notices go out a week ahead rather than an hour ahead. These measures are unpopular with active traders and they exist because the alternative is client debit balances the firm has to absorb.
For traders, the durable lesson is about the difference between volatility and gap risk. High volatility means the price moves a lot and you can still get out roughly where you intended. Gap risk means the price relocates and your exit level was never available. Position sizing built on volatility measures like ATR handles the first and completely misses the second. Any trade held through a scheduled binary event needs a size chosen from the gap scenario, not from the average range, and that thinking belongs in the same place as the rest of your risk rules.
Trading around political events carries a high risk of loss, and the outcome is genuinely unknowable in advance. Anyone telling you they knew what would happen that night is describing a memory, not a method.
"A binary event is not a trade with a wide stop. It is a coin flip you are paying spread to enter, and the size should say so."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- GBPUSD fell from around 1.50 to the low 1.30s across the night of 23 to 24 June 2016.
- The damage came from thin overnight liquidity, not from the size of the move alone.
- Widening a stop without cutting size increases risk on a binary event instead of reducing it.
- Trading the day after the result offered direction with normal liquidity and no forecast required.
Frequently Asked Questions
Why did GBP fall so far so fast on Brexit night?
The referendum had two far-apart outcomes and the market had priced the other one. As results came in, positions built on a Remain outcome were unwound during the thinnest hours of the trading day, when liquidity providers were quoting defensively.
Do wider stops protect against event gaps?
No. A stop only executes at a price someone is quoting. If the market relocates past your level without trading there, a wider stop simply books a larger loss at the fill price. Cutting position size is the control that works.
Do brokers raise margin before political events?
Most do. Raising margin on affected instruments in the days before a referendum or contested election is standard practice, and some firms also restrict the instrument to close-only orders during the event window.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.