Shape judgment needs adequate mark size; tiny scatter dots defeat it

What it is

Size-dependent shape judgment means that corners, notches, and internal structure needed for category identification may disappear as a mark is rendered and viewed smaller. The title names a failure risk, not a universal pixel cutoff. Effective size depends on visual angle, the symbol set, contrast, crowding, device-pixel ratio, zoom, viewing distance, and antialiasing.

Why it happens

Scaling, rasterization, and antialiasing alter contour sampling, while crowding and occlusion reduce access to local features. Different symbols depend on different features, so hollow versus filled and elongated versus compact marks have no fixed failure order across conditions. Higher device density can help only after CSS size, zoom, and viewing distance jointly determine the rendered visual angle.

Studying it

Render candidate symbols through the target pipeline while varying visual angle, size, density, background, and zoom. Measure category identification, pairwise confusion, and response time. Record CSS and physical pixels, device-pixel ratio, viewing distance, and export method. Retest with users who did not select the palette; enlarged design previews are not evidence for deployed readability.

Where it stops holding

Direct labels, faceting, or interactive enlargement can transfer the identification task but do not prove that the original marks are legible. Low vision, peripheral viewing, and projection may worsen loss; familiar symbols and sparse layouts may help. Color also degrades on small marks, so switching channels is not an automatic remedy.

Applying it

• Validate candidates in the target medium and smallest layout rather than accepting a design-file size or fixed pixel number.
• When density forces tiny marks, aggregate, facet, sample, label directly, or provide on-demand enlargement instead of retaining an unreadable shape dimension.
• Test hollow, filled, outlined, and internal-feature symbols separately and choose from observed confusion rather than assuming one geometry always wins.
• Provide category names, filters, and structured data for nonvisual access.

Related

• Same group: U1.07.1 Shape is categorical only, and reliably distinguishable shapes are few · U1.07.3 Texture fills are unstable under scaling and printing · U1.07.4 Orientation encodes a few ordered values but rotational symmetry confuses it · U1.07.5 Shape and texture differences survive greyscale output and colour-vision deficiency
• Nearby: U5.06 Density limits of a chart · U1.05.3 Smaller colour patches need larger hue differences to stay discriminable
• Search terms: mark size · minimum size · scatterplot glyph · visual acuity