Physical size, not resolution, sets the reachable area

What it is

Where a thumb can tap is set by the screen’s millimetres and the grip, not by the pixel count. The same “a button at top right” layout falls outside the thumb arc on a 140 mm-tall body and may just be reachable on a 120 mm body. Swapping 1080p for 1440p does not lengthen bones, so the envelope does not change. The functional area of the thumb is the patch the pad can press stably in a one-handed grip without shifting the crotch of the hand. This entry is only about reach geometry. It is not about how crowded type and controls become on a small screen, and not about which device should be used for sign-off.

Why it happens

In a one-handed grip the crotch of the hand locks the long edge and the thumb sweeps an arc around the metacarpophalangeal joint. The far part of the arc is the lower screen and the grip side; the near-opposite top corner is the expensive region—reaching it means a grip change, the other hand, or sliding the phone down. A pixel grid only slices that physical surface finer; a finer slice does not move the arc. As phones grew taller, millimetres from the top to the grip increased and top controls left the arc first—sharpness is irrelevant. A right-handed grip makes lower-right cheap and upper-left dear; left-handed grips mirror. The same point coordinate in a layout file maps to different joint angles on different body heights. People patch with small grip shifts, but each shift risks a drop, so the actual strategy is not to shift: top actions are deferred or given to the other hand.

Studying it

Have people hold a real device one-handed and paint the screen with the thumb, no grip change, to get a reach heatmap. A discrete tap task also works: targets on a grid, recording hits, time, and grip shifts. Calibrate grip from video or a pressure sleeve.

Independent variables: diagonal and height in millimetres, holding hand, whether grip change is allowed, target inside versus outside the arc. Dependent variables: reachable-area share, time and error on far targets, grip-shift count, drops or other-hand assists.

Clicking a phone bezel on a desktop simulator does not measure reach—the pointer has no joint. The physical body of the target device is required. Do not treat pixel density as a covariate that explains hit rate; fix millimetres and grip first. Seated lab postures with a supported elbow inflate the envelope relative to walking.

Where it stops holding

A two-handed hold, or resting the bottom edge on the little finger, shifts the envelope up and the top is less remote. Stylus and mouse never take the thumb arc. Tablets are not one-handed devices by default. Left- and right-hand heatmaps are mirrors; “bottom-centre is easiest,” drawn for the right hand, fails for the left. System one-handed modes slide the whole picture down as a patch; they do not shorten the body. Joint range and habit vary widely; a group heatmap is a starting map, not a personal one.

Applying it

• Place actions that must be done one-handed inside the thumb arc: send, confirm, play—lower or grip-side, not top-bar right.
• Judge placement in millimetres, not pixels: on the tallest target body, measure the top button to the bottom edge and compare with your own thumb, rather than reading point coordinates on the artboard.
• Do not put a destructive action in the arc’s centre merely because it is easy to hit; the centre is for frequent, reversible actions.
• Verify on the tallest body in the support set, one-handed, no grip change, through every critical control on the main path. Record every point that needed a shift or the other hand—those were placed by resolution, not by millimetres.

Related

• Within the group: K1.02.2 The same layout becomes much denser on a small screen · K1.02.3 Validate on the smallest device you claim to support
• Adjacent: K1.10 One-handed Mode and Large-screen Reach · C2.11 Thumb Reach Zones · F1.04 Thumb Heat Zones and Placement of Key Actions
• Search terms: thumb reach · functional area of the thumb · one-handed grip