Two weeks a year when the overlap shifts
Posted: Mon Nov 02, 2026 8:09 am
Two weeks a year when the overlap shifts
The US and Europe change their clocks on different weekends. For a short stretch in March and again in October or November, the gap between London and New York is one hour different from normal, so the famous overlap is shorter or longer.
Anyone running time-based rules sees the damage. A New York open breakout setup set for 8:30am New York time based on a London reference suddenly fires at the wrong moment. Session indicators drawn in server time end up out of step with the real flow.
The market itself does not care about the clocks. Banks open their desks at the same local time. News releases come out at fixed local times. Only the charts and scripts are confused.
A good habit is writing down the dates each year and double checking any automated rule that depends on time of day. Another is defining everything in terms of actual events, like the release of US data or the London open, instead of clock hours.
A quick audit: list every time-based rule in the plan and note which clock it uses. A session filter written in server time will drift by an hour relative to the real flows, while one written in New York time will stay put. Writing the rules around events, like the London open or the US data release, avoids the problem altogether.
Has a clock change ever shifted a strategy's behaviour for a week or two?
The US and Europe change their clocks on different weekends. For a short stretch in March and again in October or November, the gap between London and New York is one hour different from normal, so the famous overlap is shorter or longer.
Anyone running time-based rules sees the damage. A New York open breakout setup set for 8:30am New York time based on a London reference suddenly fires at the wrong moment. Session indicators drawn in server time end up out of step with the real flow.
The market itself does not care about the clocks. Banks open their desks at the same local time. News releases come out at fixed local times. Only the charts and scripts are confused.
A good habit is writing down the dates each year and double checking any automated rule that depends on time of day. Another is defining everything in terms of actual events, like the release of US data or the London open, instead of clock hours.
A quick audit: list every time-based rule in the plan and note which clock it uses. A session filter written in server time will drift by an hour relative to the real flows, while one written in New York time will stay put. Writing the rules around events, like the London open or the US data release, avoids the problem altogether.
Has a clock change ever shifted a strategy's behaviour for a week or two?