Ping and jitter test

• The page sends 25 back-to-back requests to an address of ours that replies empty, and times every round trip.
• One at a time, never in parallel: simultaneous requests compete for the connection, and the number would rise for a reason that has nothing to do with the quality of your internet.
• The first measurement is discarded. It carries the connection setup — DNS, TCP, TLS — and is always far larger than the rest.
• The headline number is the median, not the average: a single hitch should not change what you read as "your ping". The peak is shown separately, because it matters too.
• Jitter is the average difference between consecutive measurements. It is the classic reading, and it beats standard deviation here: it tells a connection that drifts slowly apart from one that saw-tooths on every packet, and the saw-toothing is what hurts in game.

• An online game does not need the packet to arrive fast; it needs it to arrive when expected. Every game compensates for fixed latency — that is what client-side movement prediction is for.
• What no compensation can fix is the unpredictable. When the gap between packets varies, the game receives a position that contradicts the one it already drew, and corrects it by force: that is the character teleporting, the shot that lands but does not count, the hit on someone who had already passed.
• That is why a stable 60 ms plays better than a fluctuating 25 ms. People who switch internet plans because of "lag" and still have the problem almost always had jitter, not high ping.
• Jitter usually comes from Wi-Fi, someone downloading in the same house, a cheap router with a full queue, or unstable power to the modem. An Ethernet cable solves a good share of cases.
• High and constant ping, on the other hand, is distance and routing: a server on another continent cannot be fixed from your side.

• Ping under 30 ms: excellent, including for competitive play.
• Ping from 30 to 60 ms: good. Most games are comfortable in this range.
• Ping from 60 to 100 ms: acceptable, but you feel it in shooters.
• Ping above 100 ms: gets in the way of any reaction-based game.
• Jitter under 5 ms: stable connection. 5 to 15 is normal; from 15 to 30 you start noticing; above 30 is what makes your character rubber-band.

• This is the time to our hosting, not to your game server. No web page can measure that: the browser does not do ICMP and does not open connections to arbitrary IPs. Anyone promising "your Valorant ping" straight from the browser is estimating.
• What counts here is your end of the connection — and it is the same for every game. High jitter here is high jitter there.
• The measurement includes the time the browser takes to process the response. It is a few milliseconds, but on a very low ping they weigh proportionally more.
• A background tab is throttled by the browser. Keep this page visible while measuring.
• Corporate networks, VPNs and blocking extensions can intercept the requests and distort the number — or stop the measurement from completing.

Look at the jitter before anything else. It is the most common case: an acceptable average ping with heavy fluctuation produces exactly the feeling people call lag. If jitter is above 15 ms, start with an Ethernet cable instead of Wi-Fi and check who else is using the internet.

Sometimes, and less than you would think. Ping is distance plus route: if the game server is in your city, the floor is already low and no plan improves it much. Switching ISPs helps when the current route is bad or the provider’s network is congested — and that shows up as jitter, not as constant high ping.

It fixes one specific case: when the default route to the server is bad and the service has a better one. It does not fix unstable Wi-Fi, a router with a full queue, or someone downloading in the same house — which is where most jitter comes from. Measure before you subscribe.

Some variation is normal, and it is exactly what jitter quantifies. If the median changes a lot from one run to the next, that is already a sign: your connection is not stable. Measure three times, a minute apart, before drawing a conclusion.