Online Game Performance Test
Start the test and play a small driving game while it runs. It is a stand-in for a real match, but the lag is real: your measured ping delays your controls, a lost request rubber-bands your car, and a disconnect freezes the game, so you can feel what your connection would do mid-match.
Test duration
Short tests are useful for quick checks. Longer tests are better for finding intermittent problems.
Free, no signup. The advanced test also moves test data to measure latency under load: up to a few hundred MB on a fast connection, so switch it off on metered data.
Customize: goal, connection type
Connected via (optional — sharpens the diagnosis)
How your connection shows up in the game
Your real ping on your controls
The latency measured right now is added to your steering, so a higher ping means your inputs land later. On a good connection that delay is tiny, because that is the truth.
Rubber-banding
A lost request or a latency spike snaps the car back to where it was a moment ago, the way a real match corrects your position after a hiccup.
Dropped connection
A detected disconnect freezes the game with a "Connection lost" overlay, just like losing your match connection.
Lag is never your fault
No life is lost while the connection is misbehaving, and the HUD shows your real ping and jitter, exactly as measured on every other test page.
Online game performance metrics on your results
Estimated input lag
Your median latency, reframed as roughly how long an action takes to reach the server and back.
Rubber-banding events
A count of latency spikes — the moments that would visibly snap a character back.
Disconnect risk
Low/Medium/High based on how many times the connection actually dropped during the test.
Hit-registration confidence
Based on jitter: how consistent your timing was, which is what makes a shot feel like it should have landed.
Why online games care about jitter more than average ping
A steady 60ms connection often feels better than one that averages 30ms but swings between 10ms and 120ms. Games interpolate movement assuming roughly steady timing; when that timing varies (jitter), the game has to guess, and guesses that turn out wrong are rubber-banding.
This is why the test reports median latency and P95 jitter rather than just an average — and why the game reacts to spikes and drops, not to the average number.