The last ten seconds of the minute carry a different tick pattern
Posted: Sat Oct 31, 2026 6:38 pm
Tick arrival is not uniform across the sixty seconds of a minute. On many feeds, the density of updates shifts around the minute boundary because of algorithmic activity keyed to bar closes, batch processing in bridges, and scheduled quote refreshes.
An easy way to see it is counting ticks by second-of-minute across a few weeks of data for one pair and session. The profile often shows bumps near seconds 0 to 2 and again around 28 to 32, with relative lulls in between. The exact shape varies by broker and pair.
That matters for strategies that sample or trigger on a clock. A signal computed at second 58 sees a different market from one computed at second 20. Spreads and fill speed vary along the same cycle.
Knowing the profile allows timing adjustments. Entering during the denser parts may offer faster fills but more noise. Entering in the lulls may give cleaner prices but slower reaction.
Collect the data before assuming anything. Each feed has its own signature.
What does the tick count by second-of-minute look like on the feed in use?
An easy way to see it is counting ticks by second-of-minute across a few weeks of data for one pair and session. The profile often shows bumps near seconds 0 to 2 and again around 28 to 32, with relative lulls in between. The exact shape varies by broker and pair.
That matters for strategies that sample or trigger on a clock. A signal computed at second 58 sees a different market from one computed at second 20. Spreads and fill speed vary along the same cycle.
Knowing the profile allows timing adjustments. Entering during the denser parts may offer faster fills but more noise. Entering in the lulls may give cleaner prices but slower reaction.
Collect the data before assuming anything. Each feed has its own signature.
What does the tick count by second-of-minute look like on the feed in use?