Three or four reference lines at once start competing with the real data for attention

What it is

Confusion risk rises with reference-line count: one labelled line is crisp, but three or four coexisting lines start competing with the data for classification — readers miscount series, misread trajectories, and the chart degenerates from "data plus benchmark" into "a heap of lines." A reference line's value comes from scarcity and clarity, and count is its first enemy.

Why it happens

The mechanism is scene-model capacity: on entering, readers build a "there are n data series here" model, and every additional reference line must squeeze into that model and be continuously discriminated — the discrimination load grows linearly with line count. The trouble compounds because horizontal reference lines resemble flat data segments in shape; once count rises, stylistic differences (dashed, grey) no longer support real-time separation, and readers must identify each line by its label — precisely the most expensive way to read a chart. Past roughly four lines, most readers' actual behaviour is to abandon discrimination entirely, and the chart's information structure fails with it.

Where it stops holding

Coexistence is not absolutely banned: two or three lines with clear semantics, uniform styling, and full labels remain readable — especially when they jointly form a band (an upper and a lower limit). The real red lines are "count approaching the number of data series" and "meaning the reader does not need": when the chart's job is trend, carrying budget, SLA, and historical mean simultaneously is overload. The standard cure for multi-benchmark demands is splitting the chart or moving to a table, giving each benchmark to the question it serves.

Applying it

• Before drawing a third reference line, prune the requirement: which judgement does this benchmark serve? No concrete judgement, no line.
• When several must coexist, uniform their styling (same colour, same dash; separated only by labels) and keep the layer clearly distinct from data strokes.
• Verification: have readers count the data series and state each reference line's meaning; miscounts or failures mean overload — split the chart or cut lines.

Related

• Same group: U5.03.1 Reference lines mark external benchmarks such as thresholds, means, or targets · U5.03.2 Reference lines must be labelled or they read as data · U5.03.3 Reference-line visual weight belongs between data and gridlines · U5.03.5 Gridlines are necessary only when values must be read off the chart
• Nearby: U5.06.2 Past the density ceiling, aggregate, sample, or facet · U5.02.4 Too many legend entries degrade matching into a linear search
• Search terms: reference line overload · multiple benchmarks · chart simplification