Clock drift on the trading computer
Computer clocks drift. A machine that is a second or two off is common, and several seconds happens if time synchronisation is turned off or fails.
For most tasks that is harmless. For a scalper who trades the seconds around news releases, it matters. A trader who thinks it is 8:29:55 may actually be at 8:30:01, with the data already out.
Drift also muddles analysis. Trade timestamps in the log do not line up with broker timestamps, making it hard to measure slippage or study fills around events.
The fix costs nothing. Enable automatic time sync, use a reliable time server, and check against an atomic clock website now and then. On a VPS, the host usually handles it, but verification is still wise.
Some traders keep a small clock on the screen, synchronised to a trusted source, and use that rather than the system tray.
It takes about a minute to verify: compare the system clock with a time server, note the offset, and write it down. If it exceeds a second, fix the sync settings before the next session.
How far off is the clock on the trading machine right now?
Clock drift on the trading computer
Clock drift on the trading computer
It’s Fairman 
-
FXS Prop Desk
- Posts: 1689
- Joined: Wed Oct 07, 2026 9:52 am
Re: Clock drift on the trading computer
The 8:29:55 example is exactly the danger. A simple check before news trading is to compare the computer clock with an internet time source. A few seconds of drift is enough to place orders at the wrong moment.
FXS Official Team | The Idea Lab & Strategy
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Clock drift on the trading computer
It's one of those settings nobody thinks about until a news trade goes wrong. On Windows the default sync interval can be quite long, and the built-in time service isn't particularly precise, so a machine can wander well over a second between syncs. Shortening the poll interval or running a dedicated NTP client fixes most of that, and it's worth confirming the service actually starts on boot rather than assuming it does.
The other half of the picture is the broker's server clock, which is what really matters for your fills. Your platform's server time display and the timestamps on ticks give you a way to compare your local clock with theirs. If the two disagree by a consistent amount, that offset matters more for planning entries around releases than how far you are from an atomic clock.
For analysis, logging both your local time and the server time on every order makes slippage studies much cleaner, since you can line up your send moment with the broker's execution moment. And on a VPS, check after any host maintenance or migration, because a rebuilt instance sometimes comes back with sync disabled.
The other half of the picture is the broker's server clock, which is what really matters for your fills. Your platform's server time display and the timestamps on ticks give you a way to compare your local clock with theirs. If the two disagree by a consistent amount, that offset matters more for planning entries around releases than how far you are from an atomic clock.
For analysis, logging both your local time and the server time on every order makes slippage studies much cleaner, since you can line up your send moment with the broker's execution moment. And on a VPS, check after any host maintenance or migration, because a rebuilt instance sometimes comes back with sync disabled.