A touchscreen has no cursor, so endpoint error cannot be corrected mid-flight

What it is

With a mouse, a cursor shows the current position before clicking, allowing users to watch it approach and correct it. In direct touch, the system decides the true endpoint only at contact, while the finger can cover the target. This cursorless touch pointing means many errors cannot be seen and corrected during the movement in the same way.

Why it happens

Mouse input separates control position from display position: the hand moves on the desk while a cursor provides continuous visual feedback. Touch unifies them; there is no independently visible proxy during approach, and the finger occludes the local display. Users must plan from proprioception, prior experience, and coarse visual estimation. At contact, an error has already become a miss or wrong activation and is usually recoverable only afterwards.

Studying it

Compare mouse, touch, and pen input with hover preview using the same targets, distances, and task; measure endpoint bias, time, misses, and correction count. Video or eye tracking can reveal whether people lose visibility of a target before contact. Separate post-contact software correction from pre-contact motor control, since snapping, magnification, and delayed confirmation alter how errors appear.

Where it stops holding

Not every touch system lacks pre-feedback: hovering pens, pressure previews, magnifiers, and drag proxies restore some correction opportunity. Continuous dragging also permits some online adjustment, but has a different error structure from a tap. The issue matters most for precise, adjacent targets with distinct consequences; large, sparse, low-risk controls are less affected.

Applying it

• Do not make a single precise tap carry all responsibility for small or adjacent actions; provide magnification, snapping, drag preview, or reversibility.
• Show the selected target and outcome immediately after contact, especially as the finger lifts, so an accidental activation can be noticed and recovered.
• In testing, record whether users can tell where they will land before tapping and what recovery follows a miss; when recovery is costly, add space or use a two-stage selection.

Related

• Same group: B1.02.1 Finger size sets a lower bound on effective width · B1.02.2 Once a target is smaller than the fingertip, gains from enlargement diminish
• Nearby: B1.01 Fitts's law · B1.03 Infinite width at screen edges
• Search terms: touch pointing · feedforward · finger occlusion · error recovery