Three measurements, three different questions
- Ping or latency
- How long a probe takes to reach a selected target and return, usually reported in milliseconds.
- Jitter
- How much the delay changes across samples; stable timing can matter as much as the average.
- Packet loss
- The percentage of probes for which the test receives no reply.
Suppose five round trips take 25, 26, 24, 25, and 26 milliseconds. The average delay and variation are both low. If they take 20, 80, 25, 140, and 22 milliseconds, the average alone hides unstable timing. If one probe has no reply, the test may record loss—but it still needs context to distinguish congestion from filtering or rate limiting.
How the numbers connect to common symptoms
| Observation | What you may notice | What to investigate first |
|---|---|---|
| Consistently high latency | Slow interaction or delayed game response | Distance to target, route, VPN exit, or congestion |
| High jitter | Uneven voice, video, or game timing | Wi-Fi radio conditions, changing load, queueing |
| Repeated packet loss | Pauses, retransmissions, reduced media quality | Weak link, congestion, unstable route, or filtering |
| One target alone fails | That test or service looks broken | Target policy, rate limiting, or route to that target |
These associations are clues, not diagnoses. Applications buffer, retransmit, and adapt differently. A streaming service may hide short variations that a live voice call exposes immediately.
Continue with How to check a VPN using IP and DNS, compare it with Learn IP and network basics, or use IP Finder for Android for the next practical step.
Why the target changes the answer
A measurement is always to somewhere. A nearby provider edge and a server on another continent have different physical distance and network paths. A game server, video-call relay, and generic test endpoint can therefore produce different results while all measurements are correct.
Choose a target that represents the actual problem when possible. If calls are unstable, a distant unrelated target may not reproduce the same route. If every target becomes worse at once, the local link or provider path becomes a stronger suspect.
The protocol details in this guide follow RFC 3393 — IP Packet Delay Variation Metric and RFC 7680 — One-Way Loss Metric for IPPM; use those primary specifications when you need exact terminology or edge-case behavior.
Wi-Fi, mobile data, VPN, and load
On Wi-Fi, radio interference, weak signal, busy airtime, and movement can change delay and loss before traffic reaches the router. Mobile performance varies with coverage, radio conditions, tower load, and carrier routing. A VPN adds a tunnel and exit location, which can change both distance and route.
Background uploads and downloads can build queues and make interactive traffic wait. When investigating, pause large transfers and keep the phone in the same position. Then compare one network or VPN condition at a time.
A repeatable comparison
- Describe the symptom
Record what fails, when it happens, and which application is affected.
- Fix the test conditions
Use the same target, method, network type, and similar device position.
- Collect several runs
Keep the individual samples or repeated summaries; do not rely on the best or worst single run.
- Change one variable
Try closer to the access point, another network, without the VPN, or at another time.
- Compare patterns
Look for a consistent change across runs and targets before assigning a cause.
Why there is no universal “good” number
The acceptable result depends on the application, target, geography, protocol, and user expectation. A remote service cannot have the same minimum physical delay as a local one. A voice call may care more about stable timing than a file transfer, while a file transfer can recover missing data through retransmission at the cost of speed.
Probe traffic may also be deprioritized, rate-limited, or blocked. If a target does not reply but its application works, test another appropriate target and compare application behavior before declaring packet loss across the connection.
What to save when asking for help
- Date, time, and timezone.
- Wi-Fi or mobile data, plus VPN state.
- Target name or address and the test method.
- Several latency, jitter, and loss results rather than one screenshot.
- The real application symptom and whether another device has it.
- What changed when you moved, changed network, or disabled the VPN.