Twenty small requests, timed. Median and jitter, not a made-up score.
Unlike every other check here, this one does make network requests — 20 of them, for a one-byte file on this domain. It has to: latency is a property of the path between you and a server, and there is no way to measure it without using that path.
Each probe, in order
Twenty bars appear here — one per probe. Even bars mean a steady line; a ragged row is jitter, and jitter is what breaks up calls.
Jitter. The average may look acceptable while conversation still breaks up, because speech cannot tolerate arrival times that keep changing.
Next step: Move closer to the router or plug in a cable, then measure again. If the row becomes even, Wi-Fi was the cause and no other change will help as much.
Distance or a slow first hop. A steady large number usually means the path is simply long — satellite, mobile, or a VPN exit in another country.
Next step: Turn off any VPN and measure again. If it drops sharply, you were measuring the detour rather than your line.
Bufferbloat: your router queues traffic instead of dropping it, and everything else waits in that queue.
Next step: Look for “smart queue”, “SQM” or “bufferbloat” in your router settings. This one setting fixes more call quality complaints than upgrading the plan.
| Approach | Best for | Trade-off |
|---|---|---|
| The ping command in a terminal | Anyone comfortable with a command line: it uses ICMP and is the most accurate option | Measures a different protocol than your browser uses, and many networks throttle or block ICMP, which makes results look worse than reality |
| The ping figure inside a game | Gamers, because it measures the exact server that matters to them | Tells you nothing when the game is not running, and each game defines the number slightly differently |
| A full speed test | Checking whether you get the bandwidth you pay for | Downloads tens of megabytes, is unusable on a metered connection, and reports latency only as an afterthought |
| This page | Finding out whether a wobbly video call is your connection, using the same protocol the call uses | Measures the path to this site only — a server on another continent will give a different number |
Where this stops being the right tool — and what to use instead.
Measuring bandwidth means pushing tens of megabytes to every visitor. That is a real cost that grows with every person who visits, and it is a field already owned by dedicated services.
Next step: Use a dedicated speed test for bandwidth. Come back here when the question is why calls stutter, because that is jitter, not speed.
A browser cannot address an arbitrary host: it can only reach servers that agree to answer it. The number here is the path to this site.
Next step: Use the in-game network display, or the ping command against that server’s address.
Ordinary web requests retry silently, so loss hides inside a slower response instead of showing up as a missing packet.
Next step: A high jitter figure here is the visible symptom. For the real measurement you need a command-line tool.
Hotel and airport networks intercept requests and answer them themselves, so you may be measuring the distance to the portal rather than to the internet.
Next step: Complete the portal login first, then measure again — the difference between the two runs is itself informative.
