Ping test

Press Start. The browser sends ten small requests to our server in Frankfurt and times each one; we report the minimum, median and 95th-percentile round-trip time, the jitter and any lost requests. Because this runs over HTTPS from the browser rather than ICMP from the system, the numbers are a few milliseconds above a classic ping to the same server — compare like with like.

Browsers cannot send ICMP, so a page cannot ping arbitrary hosts from your device. Pinging a server of yours from our location is possible and is planned for this page; until then, use your system's ping command (steps below) or the traceroute page for the path.

Round-trip time depends mostly on distance: about 1 ms per 100 km of fiber each way, plus every router and the last mile. To a server in your own country, under 20 ms is excellent and under 50 ms is normal; across an ocean, 80–150 ms is physics, not a fault.

The time for a packet to go there and back. Fixed by distance and the access technology: fiber adds almost nothing, cable a few milliseconds, DSL ten or more, 4G 30–50 ms, satellite hundreds (low-orbit constellations tens).

How much the latency varies between consecutive packets. Low jitter (a few milliseconds) matters more than low latency for calls and games; high jitter means a congested or interfered link, usually Wi-Fi.

Requests that never got an answer. Any sustained loss above 1 % breaks calls and games; it points at Wi-Fi interference, a failing cable or modem, or congestion on the provider's side.

Under 30 ms to the game server feels instant; 30–60 ms is fine; above 100 ms you notice. Game servers are regional, so your ping to them differs from the ping to Frankfurt — the relation between the two (yours should be similar for a server at a similar distance) is what matters.

Latency under 150 ms one-way is unnoticeable; jitter under 30 ms and loss under 1 % are what keep audio clean. Jitter and loss are nearly always a Wi-Fi problem fixable with a cable or a closer access point.

Wi-Fi adds 2–10 ms and most of the jitter; a cable adds nothing measurable. Run this test twice — on Wi-Fi and plugged in — and the difference is your Wi-Fi's contribution.

Plug in a cable, or move closer to the access point and switch to the 5 GHz band. Stop uploads and backups while you play — a full upload queue adds hundreds of milliseconds (bufferbloat; measure it with the connection quality test and fix it with SQM on the router). Pick servers closer to you. Switch from DSL or satellite if you can. A VPN almost always adds latency unless it routes around a bad path; test with and without.