Rollover Time: The Five Minutes You Should Not Trade
At 5 p.m. New York time, the forex trading day officially rolls over. For a few minutes around then, conditions are unusually poor.
Liquidity is low, because banks are settling positions and many participants have finished for the day. Spreads widen, sometimes dramatically, and quotes can flicker. Brokers apply swap charges at this time, and some adjust their systems, which can cause brief delays.
For traders, the dangers are practical. Stops can be triggered by spread spikes, even though the underlying price barely moved. Market orders placed at rollover often fill at poor prices. Pending orders can also trigger unexpectedly.
The effect varies by broker and pair. Major pairs usually recover within minutes, while exotics may stay wide for longer.
Simple habits help. Avoid entering trades within about fifteen minutes before and after rollover. Check that stops are not placed so tightly that a spread widening will hit them. And if you are holding positions and your strategy is sensitive, consider stops that account for it.
Remember that daylight saving changes shift the rollover time in terms of your local clock, because New York and your location may change clocks on different dates.
Also note that triple swap on Wednesday gets applied at this same moment.
Practical step: convert 5 p.m. New York time to your local time for both summer and winter, and put reminders in your calendar.
Rollover Time: The Five Minutes You Should Not Trade
Rollover Time: The Five Minutes You Should Not Trade
It’s Fairman 
Re: Rollover Time: The Five Minutes You Should Not Trade
For anyone in West Africa, 5pm New York is 10pm in Lagos most of the year and 11pm when the US is on standard time. That one-hour shift catches people every November and March.
It’s Fairman 
Re: Rollover Time: The Five Minutes You Should Not Trade
Pending orders are the sneaky part. A limit you set during the day can fill at rollover on a spread spike, and you wake up in a trade you never really chose.
It’s Fairman 
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Rollover Time: The Five Minutes You Should Not Trade
On the pending order point, it helps to separate the order types, because a wider spread does not threaten them equally. Buy stops trigger on the ask and sell stops on the bid, so when the spread opens in both directions, breakout entries on either side are the ones most likely to fire.
Limits are usually safer. A buy limit needs the ask to come down to it and a sell limit needs the bid to rise, which only happens at rollover if the quote spikes lopsidedly, so a stray fill is possible but rarer than with stop entries.
The swap side is the other trap for scalpers. Hold a trade from 21:58 to 22:03 Lagos and it pays a full night of swap, and on Wednesday that becomes three nights for a five minute position.
Say the long swap on a pair is minus 7 dollars per lot. A one lot scalp that crosses rollover on a Wednesday pays 21 dollars, which on a 5 pip target is more than 40 percent of the gross profit.
Fairman, do you pull breakout stop orders before 22:00 and leave limits in place, or clear the book entirely? An expiry time on the order itself is one less thing to remember on a tired evening.
Limits are usually safer. A buy limit needs the ask to come down to it and a sell limit needs the bid to rise, which only happens at rollover if the quote spikes lopsidedly, so a stray fill is possible but rarer than with stop entries.
The swap side is the other trap for scalpers. Hold a trade from 21:58 to 22:03 Lagos and it pays a full night of swap, and on Wednesday that becomes three nights for a five minute position.
Say the long swap on a pair is minus 7 dollars per lot. A one lot scalp that crosses rollover on a Wednesday pays 21 dollars, which on a 5 pip target is more than 40 percent of the gross profit.
Fairman, do you pull breakout stop orders before 22:00 and leave limits in place, or clear the book entirely? An expiry time on the order itself is one less thing to remember on a tired evening.
-
FXS Prop Desk
- Posts: 1690
- Joined: Wed Oct 07, 2026 9:52 am
Re: Rollover Time: The Five Minutes You Should Not Trade
For EA users, a practical step is to add a time filter that blocks new entries from a few minutes before to about fifteen minutes after the server's rollover time, and to widen or temporarily suspend trailing stops during that window. Many unexplained stop-outs in EA backtests and live logs turn out to be spread spikes at exactly that moment.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Rollover Time: The Five Minutes You Should Not Trade
On the EA filter FXS Prop Desk describes, it matters which clock the code reads. Most brokers set server time so that midnight on the server lands on the New York rollover, which means a filter written in server time follows the US clock change by itself.
A filter written in local Lagos time does not. Hard coded as 21:50 to 22:20, it is right until 1 November, then guards the wrong half hour for months while rollover sits at 23:00 Lagos.
The window length also depends on the pair. Majors usually settle within a few minutes, but crosses like GBPNZD and exotics like USDZAR can stay wide for 30 minutes or more, so one fifteen minute block is too short for them.
Gold is its own case, since many brokers pause it for a short break around the same time. A pending order on gold placed just before that pause can fill on the first quote afterwards, which may be several dollars away.
Fairman, do you keep one window for every pair or set them per symbol? Per symbol is more work, but a single filter tends to be either too tight for exotics or too wide for EURUSD.
A filter written in local Lagos time does not. Hard coded as 21:50 to 22:20, it is right until 1 November, then guards the wrong half hour for months while rollover sits at 23:00 Lagos.
The window length also depends on the pair. Majors usually settle within a few minutes, but crosses like GBPNZD and exotics like USDZAR can stay wide for 30 minutes or more, so one fifteen minute block is too short for them.
Gold is its own case, since many brokers pause it for a short break around the same time. A pending order on gold placed just before that pause can fill on the first quote afterwards, which may be several dollars away.
Fairman, do you keep one window for every pair or set them per symbol? Per symbol is more work, but a single filter tends to be either too tight for exotics or too wide for EURUSD.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Rollover Time: The Five Minutes You Should Not Trade
The 5:00 pm New York time rollover marks the end of the standard forex trading day and the point where many brokers apply swap charges. In the minutes surrounding that moment, liquidity providers adjust pricing and many reduce quoted depth.
The most visible result is spread widening. On some accounts, spreads on majors can expand several times beyond normal, and exotic pairs may become barely tradable. Stop losses and pending orders can be triggered by the temporary quote rather than by any real market movement, and a position that appeared safe may be closed at an unfavorable level.
Rollover also determines the swap date. Positions held through the cutoff are charged or credited interest, and on Wednesday the charge typically covers three days to account for weekend settlement. Scalpers who normally close within minutes can still be affected if a trade happens to overlap the window.
Practical routines vary. Some platforms allow trading to continue through the cutoff, others pause briefly. Checking historical spread charts for the rollover hour on a specific account shows how a given broker behaves, and many scalpers simply flatten positions beforehand and wait for quotes to normalize.
The most visible result is spread widening. On some accounts, spreads on majors can expand several times beyond normal, and exotic pairs may become barely tradable. Stop losses and pending orders can be triggered by the temporary quote rather than by any real market movement, and a position that appeared safe may be closed at an unfavorable level.
Rollover also determines the swap date. Positions held through the cutoff are charged or credited interest, and on Wednesday the charge typically covers three days to account for weekend settlement. Scalpers who normally close within minutes can still be affected if a trade happens to overlap the window.
Practical routines vary. Some platforms allow trading to continue through the cutoff, others pause briefly. Checking historical spread charts for the rollover hour on a specific account shows how a given broker behaves, and many scalpers simply flatten positions beforehand and wait for quotes to normalize.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Rollover Time: The Five Minutes You Should Not Trade
A monthly profit figure says nothing about how it was earned. Two accounts can finish the same month at the same level, one through steady contributions from dozens of trades and the other through a single outsized winner that rescued an otherwise flat stretch.
A simple concentration check ranks trades by result and computes the share of total profit contributed by the top five trades, or the top ten percent. It then recalculates the month with those trades removed. If the remainder is clearly negative, the month reflected luck or a rare event more than a repeatable routine.
The same logic applies to losses. A single trade that ignored a stop can dominate the monthly loss, and removing it may show that the strategy itself performed as designed while discipline failed. Those two situations call for different responses.
Concentration also affects expectations. A system that relies on rare large wins will often look poor for months and then recover quickly, while one built on frequent small gains tends to show a smoother path but is more exposed to rising costs. Neither is better in itself, but a trader should know which one is being run, because drawdown tolerance and sizing should match the profile.
A simple concentration check ranks trades by result and computes the share of total profit contributed by the top five trades, or the top ten percent. It then recalculates the month with those trades removed. If the remainder is clearly negative, the month reflected luck or a rare event more than a repeatable routine.
The same logic applies to losses. A single trade that ignored a stop can dominate the monthly loss, and removing it may show that the strategy itself performed as designed while discipline failed. Those two situations call for different responses.
Concentration also affects expectations. A system that relies on rare large wins will often look poor for months and then recover quickly, while one built on frequent small gains tends to show a smoother path but is more exposed to rising costs. Neither is better in itself, but a trader should know which one is being run, because drawdown tolerance and sizing should match the profile.