Volatility isn't spread evenly across the clock. Tick data across the major pairs often shows small spikes in activity at the top of the hour and at the half hour, even without any scheduled news.
Several factors contribute. Algorithms are scheduled by clock time. Orders are placed to execute at round times. Human traders glance at the chart and act when a new hourly or half-hourly bar forms. Fixings and benchmarks cluster there.
The effect is small, a few percent more volatility in the first seconds of the minute. But for a strategy with tiny targets, it can matter.
Plotting the average absolute price change by second within the minute gives a picture. Peaks at second 0, 1, and sometimes 30 stand out against a flatter background.
Entering exactly on the minute, as many scripts do, collides with this. The first second of the minute can have wider spreads and faster moves, so entries made then get worse fills.
Offsetting entry times by a few seconds, such as waiting until second 3 or 5 of the minute, sometimes improves average fill quality.
Does the top of the minute really stand out in the second-by-second volatility profile?