How to Test Your Keyboard for Ghosting & Key Rollover

Most keyboard ghosting advice boils down to: mash a bunch of keys, see if one drops, and check whether the box says NKRO. That test tells you almost nothing useful, because the number of keys that ghost depends entirely on which keys you pressed, not how many. A keyboard can survive a fifteen-key mash test perfectly and still drop an input the first time you strafe, jump and reload at once, because that specific combination happens to sit on a bad part of the internal matrix.

Testing this properly means testing your actual combinations, understanding why some combinations fail and others do not, and checking whether your keyboard's rollover changes depending on how it is connected. That last part surprises almost everyone.

Why ghosting happens at all

Inside a keyboard, keys are not wired individually. They sit at the intersections of a grid of rows and columns, and the controller chip scans that grid rapidly, checking which intersections are connected to detect which keys are pressed. This is done for cost and complexity reasons: wiring every key individually would need far more pins than any controller chip has.

The problem shows up when three keys pressed together happen to form an L-shape across two rows and two columns of that grid. Electrically, current can flow through the third, unintended corner of that rectangle, and the controller cannot always tell the difference between a genuine fourth key press at that corner and current merely passing through it. Depending on the keyboard's design, this produces one of two outcomes: a phantom key gets reported that nobody actually pressed, which is true ghosting, or the controller plays it safe and simply ignores one of the keys you did press, which is called jamming or blocking. Most modern keyboards are designed to jam rather than ghost, because a missing input is a smaller nuisance than a false one, but from your perspective in a game both feel identical: a key that should have worked did not.

Anti-ghosting and NKRO are not the same claim

This is the distinction that most marketing blurs together, and it is worth being precise about because the two claims solve different problems.

Anti-ghosting means the keyboard has diodes installed, one per key, that force current to flow in only one direction through the matrix. This closes off the phantom path described above and prevents false ghost reports. It does not, by itself, guarantee that any particular number of keys can be read simultaneously. A keyboard can be genuinely anti-ghosted on every key and still only support six simultaneous key presses.

N-key rollover, or NKRO, is a separate claim about how many genuine simultaneous presses the keyboard can accurately report at once. A true NKRO keyboard can, in principle, report every key on the board pressed at the same time correctly.

A board can be one without the other. Diodes on every key are common on modern gaming keyboards specifically because they are a prerequisite for real NKRO, but a keyboard can have anti-ghosting diodes and still cap out at 6KRO if its firmware or USB report format only allows six simultaneous keys through.

The 6-key limit comes from USB itself, not just cheap keyboards

Here is the part almost nobody explains, and it is the most useful thing to know if a keyboard behaves inconsistently across setups.

The standard USB HID "boot protocol" for keyboards, the basic compatibility mode every USB keyboard supports so it works in a BIOS, a boot menu, or any device with no dedicated driver, uses a fixed 8-byte report: one byte for modifier keys like Shift and Ctrl, one reserved byte, and six byte slots for ordinary keys. That format can only describe six non-modifier keys held down at once, full stop, regardless of how good the keyboard's internal matrix is.

A genuine NKRO keyboard gets around this by switching, once your operating system loads its driver or its own software, to a custom report format that sends a full bitmap of every key rather than six fixed slots. This works while that driver or software is active. It does not work in situations where the boot protocol is what gets used instead: entering a BIOS or UEFI setup screen, a boot menu before an operating system has loaded, or occasionally in games and applications that specifically request boot-protocol compatibility for broad hardware support.

This is why some gaming keyboards support full rollover in a game with their software running, and drop back to a firm six-key limit for entering a multi-key BIOS shortcut. It is not a fault. It is the boot protocol doing exactly what it was designed to do.

Bluetooth and 2.4GHz can quietly reduce your rollover too

If your keyboard has more than one connection mode, do not assume rollover is the same across all of them.

Many wireless keyboards, including ones with genuine NKRO over a wired USB connection, implement their Bluetooth mode using the standard HID-over-GATT keyboard profile, which is built around the same boot-protocol-style limited report for maximum compatibility with phones, tablets and other devices that may not support a vendor-specific format. The practical result is a keyboard that is full NKRO over USB and six-key limited over Bluetooth, with nothing in the marketing telling you this changes between modes.

A proprietary 2.4GHz dongle is not automatically better here, since it depends entirely on what the manufacturer chose to implement on that specific receiver. Some carry full rollover through, others do not.

The only way to know for certain is to test each connection mode your keyboard offers separately, using the exact key combinations you actually rely on.

Test your actual combinations, not a random mash

This is the practical difference between a test that tells you something and one that does not.

Open a keyboard test and, instead of pressing keys at random, work through the specific combinations your games and workflow actually use:

• WASD plus Shift, for sprinting while strafing
• WASD plus Space, for strafe-jumping
• WASD plus Ctrl, for crouch movement
• Your full ability or hotkey row held alongside a movement key, if you play anything with active combat abilities
• Any three-key combination you use for a shortcut, since three keys on the wrong two rows and columns is exactly the geometry that causes trouble

Hold each combination down fully and confirm every key you pressed lights up as active on the test, with nothing missing and nothing extra appearing that you did not press. Then, if your keyboard supports more than one connection type, repeat the exact same combinations over each one. A combination that works perfectly wired and drops a key over Bluetooth has told you something a generic full-board mash test never would have.

If you find a specific combination that consistently fails, that is a genuine, repeatable matrix limitation at that intersection of keys, and it will fail the same way every time you use it in a game, not just under test conditions.

Switch bounce is a related but separate fault

Worth distinguishing from rollover, since it produces a different kind of unreliable input. Mechanical switches can occasionally register a single physical press as two or more rapid electrical signals, a phenomenon called bounce, before the keyboard's own debounce logic filters it out. A switch with worn or contaminated contacts can bounce more than it should, producing an occasional double input from one press.

This is the keyboard-side version of the same underlying issue that causes mouse double-clicking, and it is tested a similar way: press a single key repeatedly at a controlled, deliberate pace and check whether it ever registers more than once per press. Our guide on fixing mouse double-clicking covers the mechanism in more depth, and a click speed test is a reasonable stand-in for spotting an inconsistent switch, even though it was built with mouse buttons in mind.

What to do if you find a real limitation

If a specific combination genuinely fails and it is one you rely on, first check for a firmware update, since manufacturers occasionally revise matrix scanning logic. Then check whether your keyboard's software has an NKRO or gaming mode toggle that is not enabled by default, since some ship in a compatibility-first 6KRO mode out of the box specifically to avoid confusing devices that expect boot protocol behaviour.

If neither helps, the practical fix is remapping one of the conflicting keys in that specific combination to a key on a different part of the matrix, which most keyboard software allows. This is a workaround rather than a repair, but it is a permanent and reliable one, since the underlying matrix geometry is fixed in hardware and will not change.

If the keyboard is failing combinations it is specifically marketed as supporting, that is a legitimate case for a warranty claim, particularly if you can demonstrate the exact combination failing consistently on video.

Beyond the keyboard

If your inputs are clean but you still feel like you are losing races you should win, the keyboard is only one link in the chain. Our guide on eliminating input lag on PC covers the rest of it, and a reaction time test is worth running once you are confident your hardware is not the limiting factor, since it measures the part of the equation no amount of keyboard testing can improve.