Stopwatch

What this tool is used for

• Timing a manual process to find out how long it really takes.
• Recording lap times for repeated attempts at the same task.
• Timing a build or a deploy when the tooling does not report it.
• Measuring an interval during a demo or a test.
• Comparing two approaches by timing both.

Frequently Asked Questions

How accurate is a browser stopwatch?

Good to a few milliseconds, using performance.now() — a monotonic clock that is immune to system time changes. Date.now() is wrong for timing because an NTP correction or a DST change can make it jump backwards.

Why do timers drift in a background tab?

Browsers throttle setInterval in inactive tabs — often to once per second or less — to save battery. Counting ticks therefore loses time; computing elapsed time from a start timestamp does not, and stays correct across throttling.

What is the difference between lap and split times?

A split is the total elapsed at that point; a lap is the time since the previous mark. Four 60-second laps give splits of 1:00, 2:00, 3:00, 4:00. Confusing them is why reported race times sometimes look impossible.

Why is timer resolution deliberately reduced?

Because high-resolution timers enabled Spectre-class side-channel attacks. Browsers now coarsen performance.now() and add jitter, which is why sub-millisecond precision is not available and should not be relied on.

Does it keep time if I close the tab?

No — state lives in the page. A refresh or close ends it unless the elapsed time was persisted. This is a genuine limitation of browser tools compared with a native app that can run in the background.

Why measure elapsed time from a timestamp rather than by counting?

Because a timer that fires late loses the difference permanently, and the errors accumulate. Recording a start time and subtracting means a delayed frame shows the correct elapsed time on the next one.

Which clock should be used?

A monotonic one such as `performance.now()`, because the wall clock can jump backwards when the system syncs time or crosses a daylight-saving boundary — producing a negative elapsed time on an otherwise normal run.

How precise is the displayed value really?

Browsers deliberately coarsen high-resolution timestamps to limit timing side-channel attacks, so hundredths are honest and finer digits are not. Displaying milliseconds implies a precision the platform does not offer.

Common errors and gotchas

• Trusting millisecond precision for a human-started measurement, where reaction time dominates the error.
• Leaving the page in a background tab, where timers may be throttled by the browser.
• Losing the recorded laps by navigating away, since they are held in the page.
• Using it to benchmark code, where a proper profiler measures what actually matters.
• Comparing timings taken under different conditions as if they were controlled.