What is a controller deadzone?

The deadzone is the small area around the centre of an analog stick where movement is ignored and the stick still reports zero. It exists to stop a slightly worn or off-centre stick from making your character creep on its own. Too small and you get unwanted drift; too large and small precise aims feel unresponsive.

What is stick circularity and why does it matter?

Circularity describes whether rolling the stick around its rim traces a true circle. On a perfect stick, full deflection reports the same magnitude at every angle. In practice most sticks trace something between a circle and a square, so diagonal input can be stronger or weaker than cardinal input. That inconsistency is what makes diagonal aim feel different from horizontal aim.

What is a good average circularity error?

Under about 5 percent is excellent and typical of a healthy modern pad. Between 5 and 10 percent is good and normal. Between 10 and 18 percent is common on older or budget controllers. Above roughly 18 percent the stick is noticeably inconsistent, which usually means wear, a squared-off output range, or reduced travel in some directions.

How do I know if my controller has stick drift?

Release both sticks and run the rest measurement. A healthy stick sits at or very near zero. If it reports a steady non-zero offset while untouched, that is stick drift, and in a game it shows up as slow unwanted movement. A large deadzone can mask drift without fixing the underlying wear.

Does this measure the raw hardware or the value after the driver?

It measures the value after the operating system and driver, which is exactly what games receive through the same browser Gamepad API. That makes it a useful real-world reading, but it means a deadzone applied by the driver or by vendor software appears as part of the result rather than being separated out.

Why does my stick report more than 100 percent at the diagonals?

Some controllers expose a square output range, where the horizontal and vertical axes can both reach full scale at the same time. Combined, that gives a magnitude above 1 on the diagonals, up to about 1.41. It is not a fault, but games that expect a circular range may treat those diagonals as faster than intended.

Why does the test require three full laps before showing a result?

Because the circularity figure comes from the maximum reach recorded in each sector, and a maximum only converges if you pass through that sector more than once. On a single lap you inevitably fail to reach the true rim everywhere, so every shortfall is counted as error and the result reads worse than the controller actually is. Requiring three passes per sector removes most of that bias. The shaded area on the plot shows the ground you have already covered.

Which controllers and browsers work with this test?

Any controller the browser recognises through the Gamepad API, including PS5 DualSense, PS4 DualShock, Xbox controllers and most generic USB or Bluetooth gamepads. It works in Chrome, Edge and other Chromium browsers; Firefox and Safari support varies. There is no install, no account and no limit on how long you test.

Controller Deadzone & Stick Circularity Test (Free, No Install)

The short version

Rest offset is stick drift. A healthy stick reads essentially zero when you let go.
Deadzone is how far you must push before anything registers. Big deadzones hide drift but cost you fine aim.
Average circularity error is how far a full roll deviates from a true circle. Under 5% is excellent, 5–10% is normal, above ~18% is noticeably inconsistent.
The result needs 3 full laps per stick — one lap always under-reads the maximum, which fakes a worse score.
Everything here is measured after the driver — the same values a game receives.

How the test works

Three quick measurements, all from the same live stream of stick positions:

Rest position. With both sticks released, the tool averages the reported position over a second. Anything meaningfully above zero is drift.
Deadzone. As you push a stick slowly out from centre, the reading stays at exactly zero until the deadzone is cleared. The smallest non-zero deflection seen is the effective deadzone.
Circularity. Rolling the stick around its rim, the tool records the maximum magnitude reached in each of 72 five-degree sectors. Comparing those maxima to a perfect circle of radius 1 gives the average and maximum error. The plot shades the area you have swept, from centre out to your furthest reach at each angle, so you can see at a glance which directions you have not covered yet.

Three laps, slowly, against the rim

The result stays locked until every one of the 72 sectors has been passed three times. That is deliberate: a maximum only settles if you revisit it, and a single lap always falls short somewhere, which shows up as error the controller does not really have. Roll slowly and firmly — the progress bar counts laps, not time.

Deadzone and stick drift, explained

An analog stick never returns to exactly the same spot. As the potentiometers or Hall sensors wear, the centre creeps. A deadzone is the software's answer: ignore everything inside a small radius so a slightly off-centre stick doesn't make your character walk on its own.

That fix has a cost. Every millimetre inside the deadzone is travel you can't use for slow, precise movement — the thing that matters most for tracking a target or lining up a shot. A pad with heavy wear and a large compensating deadzone feels simultaneously drifty and unresponsive.

This is why the two numbers belong together. A large deadzone with a clean rest position is a design choice. A large deadzone hiding a large rest offset is a controller on its way out.

Circularity and average error, explained

Push a stick fully right and it reports 100%. Push it fully up-right and — ideally — it still reports 100%, just split across both axes. Trace the whole rim and you should get a circle.

Real sticks don't. Two common shapes:

Squared output. Both axes can hit full scale at once, so diagonals read up to about 141%. Games expecting a circular range read those diagonals as faster than intended.
Clipped diagonals. The stick physically can't travel as far diagonally, so diagonals fall short — often 85–95%. Diagonal movement is quietly slower than cardinal movement.

The average circularity error collapses this into one figure: the mean distance between your traced shape and the ideal circle, as a percentage. It's the number worth comparing between two controllers, or between the same controller before and after a repair.

How to read your results

Reading	Healthy	Worth attention
Rest offset	Under 2%	Above ~5% is real drift
Deadzone	5–12% is typical	Above ~20% costs noticeable precision
Avg circularity error	Under 10%	Above ~18% is inconsistent
Max reach	~100%, or ~141% on square-output pads	Below ~90% means lost travel

What you can actually do about it

Recalibrate first. DualSense and many PC pads expose a calibration routine (through the console settings, Steam's controller settings, or vendor software). It costs nothing and often clears a small rest offset.
Clean around the stick base. Dust and skin oil under the gate cause both sticky travel and clipped diagonals. Compressed air and isopropyl alcohol, sparingly.
Lower the in-game deadzone only once the rest offset is genuinely near zero — otherwise you're trading responsiveness for drift.
Replace the module if the rest offset stays high after calibration. Drift is mechanical wear; software can mask it but not undo it. Hall-effect replacement sticks don't wear the same way.
If your overlay looks wrong rather than the stick feeling wrong, that's a different problem — see fixing controller overlay lag.

What this test can't tell you

The browser reads your stick after the operating system and driver have processed it, through the same Gamepad API a web game would use. That's genuinely useful — it's what software actually receives — but it has limits worth stating plainly:

A deadzone applied by the driver or vendor software is indistinguishable here from one baked into the controller.
Values are typically quantised to 8 bits per axis, so very small differences aren't resolvable.
It measures the stick's reported geometry, not the mechanism. It can tell you a stick is inconsistent; it can't tell you which component is worn.
Reaching a true rim position at every angle is down to your hand. Low coverage means a less trustworthy average — the tool shows coverage for exactly that reason.

Sticks checked out? Put them on stream.

Turn your controller into a live, transparent input overlay for OBS or Streamlabs in a couple of minutes.

Build my overlay →

Controller deadzone & stick circularity test

Left stick —

Right stick —