How to test Wi-Fi against Ethernet
Use one computer in one spot. Run the full test three times on a cable plugged into your router, then three times on Wi-Fi, and compare the middle result of each set. If the cable is clearly better, your Wi-Fi is the problem and you can fix it at home. If both are equally bad, the problem is your router, your line or your provider.
Why compare
Every number on the test can be hurt by Wi-Fi or by the line itself, and the fixes are completely different. A cable takes Wi-Fi out of the path, so testing both ways on the same device tells you which one to fix. Providers use the same logic: Starlink's own troubleshooting asks you to test with a computer connected by Ethernet to get accurate results.
What you need
- A laptop or desktop with an Ethernet port, or a USB Ethernet adapter.
- An Ethernet cable long enough to reach the router from where you'll test.
- About 15 minutes when nobody else is using the internet heavily.
A phone or tablet can't do the cable half, so use a computer.
The method
- Pick the spot. Put the computer where you'll test both ways: near enough to the router for the cable to reach. If you mostly use Wi-Fi in another room, you can add a third set there later.
- Quiet the network. Pause downloads, cloud backups and streaming on other devices, and close other programs on the computer.
- Plug in the cable to one of the router's LAN ports and turn off Wi-Fi on the computer, so you know the test uses the cable.
- Run the full test three times. Write down ping, jitter, download, upload, packet loss and bufferbloat grade for each run, or use Copy result and paste each into a note.
- Unplug the cable and turn Wi-Fi back on. Stay in the same spot.
- Run the full test three times again and note the same numbers.
- Optional: move to where you usually sit and run three more on Wi-Fi.
Use the middle result of each set of three, not the best or the worst. It's less affected by a one-off spike.
A worked example
An example of what a comparison can look like:
| Cable (middle of 3) | Wi-Fi, same spot (middle of 3) | |
|---|---|---|
| Ping | 14 ms | 22 ms |
| Jitter | 1 ms | 9 ms |
| Download | 480 Mbps | 210 Mbps |
| Upload | 22 Mbps | 21 Mbps |
| Packet loss | 0% | 0% |
| Bufferbloat grade | D | D |
Reading it: Wi-Fi adds 8 ms of ping and 8 ms of jitter and halves the download, so Wi-Fi costs something, but not enough to fail the competitive-gaming limits on its own (ping 50 ms, jitter 10 ms). The bufferbloat grade is D both ways, though, and that's the real problem: it comes from the router and line, not the Wi-Fi. The fix to start with is bufferbloat, and the cable's ping is a bonus.
How to read your result
| What you see | What it means | Next step |
|---|---|---|
| Cable good, Wi-Fi much worse ping or jitter | Wi-Fi is adding the delay | High ping and jitter on Wi-Fi |
| Cable 0% loss, Wi-Fi shows loss | Wi-Fi interference or weak signal | High ping and jitter on Wi-Fi |
| Loss on both | The router, the line or the provider | Fix packet loss |
| Same bufferbloat grade C or worse on both | The router's queue, not Wi-Fi | Fix bufferbloat |
| Cable download far below your plan | The line, modem or router | Restart the modem and router, then contact your provider with your results |
| Cable fast, Wi-Fi much slower download, good ping | Normal for Wi-Fi to some degree | Only worth fixing if it fails what you do (4K streaming needs 15 Mbps) |
| Both good | Your connection isn't the problem | If a game still lags, see lag in online games |
Some differences are normal. Wi-Fi almost always shows a little more ping and jitter than a cable, and a lower download on fast plans. What matters is whether the Wi-Fi numbers still pass the limits for what you do: for competitive games, ping 50 ms or less, jitter 10 ms or less, loss 1% or less and grade B or better; for video calls, 150 ms, 30 ms, 2% and grade C.
Common mistakes
- Wi-Fi left on during the cable runs. With both on, you can't be sure which one the test used. Turn Wi-Fi off.
- Testing on different devices. A phone on Wi-Fi against a desktop on a cable compares two devices, not two connections.
- Moving between sets. Test Wi-Fi from the same spot first; that isolates the radio from the distance.
- Something else downloading. A game update in the next room changes the bufferbloat grade and the speeds. Run the sets close together, with the network quiet.
- Comparing with another speed test. Different tests measure to different servers in different ways. Compare our results with each other; why results differ explains the rest.
Test again
After any fix, repeat the method: three runs on a cable, three on Wi-Fi, same spot, and compare with your first results. That tells you whether the fix worked and how much it changed.
Sources
Checked November 2026.
- Starlink: Slow speeds: troubleshooting
- Test method, thresholds and grades: how it works