Keyboard ghosting explained

Ghosting describes an unexpected key reported when it was not pressed. A missing key in a combination is a different observation, often called blocking. Both can involve keyboard scanning, but software can also change what reaches a browser. Rollover describes simultaneous-key reporting capability; a web page observes only the events delivered to it.

Build a small, repeatable combination

1. Open Keyboard Ghosting Test with a physical keyboard, activate the workspace, and release all keys.
2. Hold W alone. Expect W to appear as held. Add A while keeping W down; expect both to remain visible.
3. Add Shift, then Space, one key at a time. Note the first step where a held key disappears or an untouched key appears.
4. Release every key. The held-key display should clear. Reset before repeating the sequence slowly.
5. Reverse the press order and repeat. Write down the exact keys, order, and observed discrepancy rather than just the maximum count.

These are expected observations for a comparison procedure, not a claim that this specific combination has been tested on every keyboard. An unexpected result is evidence to investigate, not certification of a fault.

Separate browser behavior from a hardware claim

When the maximum looks right but a key feels wrong

Two keyboards can both reach a maximum of three while reporting different keys. On the ghosting test, keep your fingers on W and A while tapping D. Compare Held keys with your fingers at each step. The tool displays the physical codes as KEYW, KEYA and KEYD.

Action	Current held	Maximum held since reset
Reset, then hold W and A	2 (W, A)	2
Add D	3 (W, A, D)	3
Release only D	2 (W, A)	3
Release W and A	0	3

If the display reaches three but names S instead of D, record both observations: D was missing and S was unexpected. If it stays at two and still lists W and A, record just the missing D. Reset and try W+D and A+D separately. If both pairs work, repeat the three-key sequence with D first. Note which press changes the list; that is more informative than reporting only a peak of three.

If the page loses focus or a shortcut opens another interface, that run cannot establish rollover behavior. Refocus and start again. Remapping utilities, extensions, the operating system, firmware, and the connection mode can all affect the event path. MDN KeyboardEvent documents what the browser event exposes; it does not reveal the keyboard’s circuit or diode layout.

If a discrepancy repeats with the page focused, compare the same combination in the application where it matters. If possible, try another keyboard on the same computer, then the original keyboard on another computer. Keep the key order and connection mode in your notes. A reproducible difference narrows the cause; it still does not identify a failed circuit.

What anti-ghosting and NKRO do not prove here

Anti-ghosting claims may cover selected combinations rather than every possible chord. NKRO describes independent reporting of simultaneous key presses, but seeing a large held-key count in the N-key rollover test cannot certify that capability across all combinations or connection modes. A small count also does not prove a hardware limit if the browser never received the other events.

Use the manufacturer’s model-specific specification alongside your observations. For a key that fails even on its own, follow unresponsive-key troubleshooting. This site cannot repair a matrix limitation or measure physical keyboard latency from these observations.