ConcertoNotes Fitts's Law and pointer acceleration
The mouse and trackpad are not "raw" inputs — operating systems apply acceleration curves that translate physical movement into cursor displacement. Fast physical motion produces disproportionately fast cursor motion; slow physical motion produces precise, slow cursor motion. This curve interacts with Fitts's Law in ways that affect how target size and distance are perceived.
What pointer acceleration does
A non-accelerated cursor maps physical motion to screen motion 1:1. To cross 1000 pixels, the user must move the mouse 1000 mouse-equivalent units. Fast or slow physical motion makes no difference to the resulting distance.
An accelerated cursor varies the multiplier based on physical speed:
- Slow physical motion → low multiplier (e.g., 0.5×). Used for precise positioning over small targets.
- Fast physical motion → high multiplier (e.g., 3–8×). Used for crossing large screen distances quickly.
The acceleration curve is what lets a single mouse on a 4K display traverse the full screen without lifting and re-anchoring (a problem 1990s mice without acceleration had).
For Fitts's Law, this means:
- Distance is partly free. Long traversals are accelerated, so movement time grows much less than the linear formula predicts.
- Target size still matters. Once the user is near the target, they slow down for fine positioning, and target size dominates again.
- Trackpads and mice have different curves. Same Fitts's task, different time outcomes.
Implications for UI design
1. Distance matters less than naive Fitts's Law suggests on accelerated pointers
A toolbar at the top of a 1440px-wide window vs. one halfway down: the difference in actual cursor travel time is smaller than the geometric distance suggests. Don't sacrifice layout sense to "place everything close together" — acceleration absorbs much of the cost.
2. Target size matters more than naive Fitts's Law suggests near the target
When the cursor decelerates for fine positioning, small targets become disproportionately hard. A 16×16 button is much harder than a 24×24 button — more than the 50% size increase would suggest, because deceleration phase amplifies precision cost.
The takeaway: when in doubt, prefer bigger targets to closer targets. Acceleration handles distance; nothing handles fine-positioning except size.
3. Drag operations are sensitive to acceleration
A drag-and-drop interaction asks the user to move while pressing the button. Acceleration is still active, so a small physical motion can become a large cursor motion. This causes:
- Overshoot on rapid drags. The user releases past the target.
- Jitter on fine drags. Slow drags feel "sticky" or imprecise as the curve transitions.
Mitigations:
- Provide visible drop zones that are larger than the dragged item's footprint. The user has slop room.
- Snap-to-grid or snap-to-target for positioning interactions. Reduces precision burden.
- Hold-modifier for fine control. Hold
Shift(or similar) to disable acceleration during drag. Common in drawing and layout tools.
4. Sliders and scrubbers benefit from gain control
A volume slider, a video scrubber, a date-range picker: the user is dragging to a precise value. With raw acceleration, large values are easy to skip past.
Patterns:
- Variable gain on slider drag. Drag near the slider track for fast scrubbing; drag away from the track (off-axis) for fine scrubbing. iOS scrubbers use this.
- Snap-to-tick for slider values that should land on integers or other discrete values.
- Display the current value during drag so the user can correct without lifting.
5. Different input devices, different curves
The same web app is used with:
- Desktop mouse — moderate acceleration, medium precision.
- Trackpad — high acceleration, lower precision (small physical surface), gestures.
- Stylus / pen — low/no acceleration, high precision.
- Touchscreen — no acceleration (1:1), but finger contact width adds imprecision.
- Eye-tracking — low precision, no fine positioning, dwell-based selection.
A UI that works on a 27" monitor with a precise mouse may be unusable with a trackpad on a 13" laptop. Test both.
For accessibility:
- Switch devices (single-button accessibility input) — every interactive element must be reachable through scanning, not pointer acquisition.
- Sip-and-puff and other adaptive devices — same.
Worked examples
Example 1: a slider with click-to-position
<input type="range" min="0" max="100" value="50" />
Native <input type="range"> works across input devices. The user can click anywhere on the track to jump to that value (no drag required); drag the thumb for fine adjustment; tap arrow keys for precise stepping. All of this is built in.
For custom sliders:
- Make the track full-width clickable (not just the thumb).
- Make the thumb hit area ≥ 24×24 (often the visible thumb is smaller, but the hit area extends).
- Support arrow keys for keyboard increment.
Example 2: a draggable list item
<li draggable="true" class="task">
<span class="drag-handle" aria-label="Drag to reorder">⋮⋮</span>
Task name
</li>
The drag handle is visually small but should have a 32×32+ hit area. During drag:
- Show a visible placeholder where the item will drop (large drop target).
- Display drop indicators between rows.
- Provide a keyboard alternative (e.g., arrow keys to reorder) for users who can't drag.
Example 3: a color picker
A 256×256 color square asks for very fine positioning (each pixel = a different color). Pointer acceleration makes this hard.
Mitigations:
- Provide numeric inputs (RGB / HSL / HEX) for users who need precision.
- Provide preset palette swatches for common selections.
- Allow
Shift+click(or another modifier) for fine-position mode.
Example 4: pinch-and-zoom (touch)
Touch has no acceleration, but two-finger gestures benefit from velocity-aware behavior:
- Slow pinch = precise zoom (small zoom factor change per pixel).
- Fast pinch = rapid zoom (larger zoom factor change per pixel).
Implementations: track pinch velocity; multiply zoom delta by a function of velocity.
Anti-patterns
- Custom sliders without keyboard support. A drag-only slider is unusable for keyboard, switch-device, and many assistive-tech users. Always support arrow-key increment.
- Tiny drag handles. A 16×16 drag handle on a list row. The user can't reliably grab it; reordering becomes painful.
- Drag-only reordering. Reordering that requires drag-and-drop with no keyboard alternative. Inaccessible by default.
- No snap on positioning. A drag-to-arrange canvas with no grid, no snap-to-edge, no align-with-neighbor. Every placement requires fine positioning that pointer acceleration sabotages.
- Hover-precision interactions. A hover-only menu where the dropdown is below the trigger and the user must traverse a wedge-shaped path to reach the items. The trackpad user, with high acceleration, overshoots into adjacent menu items. Either use click-to-open or design wider safe traversal areas.
Heuristics
- Test on a trackpad. If your interaction requires a mouse to be usable, it's broken for ~half your users.
- Test with arrow keys. Every interactive element should respond meaningfully to keyboard alternatives. (This is also accessibility.)
- The fine-positioning audit. Walk through every interaction that asks for precision. Is there a coarser alternative (typed value, snap, increment buttons)?
Related sub-skills
fitts-law(parent).fitts-law-touch-targets— touch has no acceleration but its own precision constraints.affordance— drag handles need visible affordance.accessibility-operable(process plugin) — pointer-precision interactions must have non-pointer alternatives.mapping(cognition) — sliders and scrubbers map control motion to value change; mapping must be intuitive.