๐Ÿ•นStick drift test

โœ“ Tested & verified Updated: How the test works

Stick drift test

Left stick

Right stick

Drift score

โ€”/ 100

Take your thumbs off the sticks and leave the pad flat. The score is the biggest reading the sticks sent while nobody was touching them: under 20 is normal, 21 to 50 is mild drift, over 50 is a stick that will move your character on its own.

  • Left stickโ€”
  • Right stickโ€”
  • Samplesโ€”
  • Time leftโ€”

Put the controller down and let go completely โ€” a thumb resting on a stick invents drift that is not there.

One gear check a month

Monthly

One thing a month worth measuring before it fails properly โ€” a mic that reset itself, a stick that drifts, a switch sending two clicks. Two sentences and the test.

Free, no account. One click unsubscribes and deletes your address.

Try also

What counts as stick drift?

The Devicedial stick drift test watches both analog sticks for ten seconds while you do not touch them and returns a drift score from 0 to 100. The score is the resting displacement of the worse stick multiplied by 100: a score of 12 means the stick sits 0.12 of the way toward the edge while nothing is holding it there.

That number matters because games ignore small offsets on purpose. Most titles apply a deadzone somewhere between 0.15 and 0.25, so a score under 20 disappears before the game ever sees it. Between 21 and 50 the offset is past most deadzones and shows up as a camera that creeps or a menu cursor that walks. Above 50 the stick is effectively being held in a direction.

The reading is the raw axis value, taken before any game applies its own deadzone โ€” which is why a controller can look fine in one title and unplayable in another.

The score is on the same 0 to 100 scale as the deadzone sliders in most games, so it can be read against them directly. This is what each band means in play.

Drift scoreResting offsetWhat you notice in a gameWhat to do
0โ€“5under 0.05nothing at allhealthy stick
6โ€“200.06 to 0.20hidden by a standard deadzoneretest in a month
21โ€“350.21 to 0.35slow camera creep, menu cursor walksraise the in-game deadzone above your score
36โ€“500.36 to 0.50constant creep, aim pulls one wayclean or replace the module
over 50over 0.50the game reads the stick as heldreplace the module

One caveat worth knowing: a score of exactly 0 on a stick you are certain drifts means the offset is smaller than the controller’s own firmware deadzone, not that it has gone away.

For the rest of the pad โ€” every button, both triggers and all four axes at once โ€” use the gamepad tester. The same failure mode in a mouse switch is measured by the double click test, and the mouse tester covers buttons and scroll wheel.

Frequently asked questions

What drift score is still OK?

Under 20 is fine: the offset is smaller than the deadzone most games apply, so nothing in the game ever sees it. Between 21 and 50 you are past typical deadzones โ€” expect the camera to creep or a menu cursor to walk when you let go. Over 50 the stick is being held in a direction as far as the game is concerned, and no deadzone setting will hide that.

How do I test drift when the game has its own deadzone?

Test here first, because this page reads the raw axis before any game filters it. Then compare: if a title exposes a deadzone slider from 0 to 100, your score is on the same scale, so a score of 18 needs a deadzone above 18 to vanish. Raising it is a workaround, not a repair โ€” a larger deadzone also means the stick must travel further before slow movements register at all.

My score is zero but the stick still drifts in game.

Some controllers apply their own deadzone in firmware and report exactly 0.00 until the offset clears it, so a small drift is invisible here and to everything else. Drift also grows as the pad warms up, and it often only appears after the stick has been pushed and released rather than left alone. Run the test again after a session instead of before one.

Is the drift in the controller or in the game?

This page answers that. It reads the controller directly, with no game engine, sensitivity curve or acceleration in between, so a near-zero score here plus drift in one title points at that title’s configuration. A high score points at the hardware. Test a second game before you open the controller โ€” response curves differ enormously between engines.

Can stick drift be repaired?

Sometimes. On many pads the fault is contamination in the potentiometer, and electrical contact cleaner worked in around the stick base clears it for a while. The lasting fix is replacing the stick module, ideally with a Hall effect module that reads a magnet instead of a wiper and so has no contact surface to wear. Check the warranty first: an opened controller has none.

Why does the score change between runs?

Because the stick does not return to the same place twice. It settles wherever the spring leaves it, so a run after the pad has been idle and a run straight after pushing both sticks to the edge can differ by several points. Take the worse of the two โ€” that is the offset a game will actually be handed mid-session.

How the test works

The Gamepad API reports each axis as a float normalized to the range -1 to +1 โ€” a proportion of full deflection, not an angle in degrees and not a physical distance. For each stick, the page samples both axes once per animation frame for the full window and computes the displacement from center as the length of the vector, the square root of x squared plus y squared.

The Devicedial drift score is our own definition, and it is deliberately simple: the median displacement of the worse of the two sticks over the whole sample, multiplied by 100 and rounded. The median rather than the peak, so one knock against the desk cannot decide the result; the worse stick rather than the average, because a controller is as good as its bad stick. A score of 12 therefore means a resting offset of 0.12 on a scale where 100 is a stick pinned to its edge. Nothing is uploaded โ€” the sampling happens in your browser and ends when the page closes.

Worked example

With axis values while nobody is touching the stick 0.09, โˆ’0.04, 0.02, โˆ’0.011, this page works out a drift score of 9 out of 100. Under 20 is what a healthy stick looks like: a resting stick is never exactly zero, and every game ignores this much through its own dead zone.

  • Largest resting deflection0.09 of full travel
  • Drift score9 / 100
  • Service threshold20

Sources:

Drift, deadzone and the number in between

Drift is a hardware fault. Deadzone is a software setting. The stick reports a position it should not, and the game throws away small positions before acting on them. Neither knows about the other, which is why the same controller can be fine in one title and unusable in the next.

A drift score is only useful next to a deadzone. That is the whole reason this page reports a number on a 0 to 100 scale rather than a verdict: deadzone sliders in games use the same scale, so the two are directly comparable and you can see exactly how much headroom is left.

Raising the deadzone is a trade, not a fix. It buys back a stable center at the cost of fine control, because the first fraction of the stick’s travel now does nothing at all. On a score above 50 there is not enough travel left to trade, and the module has to be replaced.