Trend judgement depends on line-segment angle

What it is

"How fast is it rising?" — readers do not compute this, they read it off the angle: a rise feels "steep" because its angle against the horizontal is large; which of two segments is steeper is a comparison of two angles. Obvious-sounding, but every corollary is practical: the angle is picture geometry (data slope × aspect ratio — previous leaf), not the data itself; angle comparison has a precision profile — differences near 45° resolve best, differences near flat or near vertical barely at all; so "can readers see the trend difference" depends on where the angles fall, and where they fall is tunable through aspect ratio (and axis range). The reading mechanism dictates the design duty.

Why it happens

Orientation discrimination is not uniformly sensitive: the visual system resolves angle differences best around 40°–50° from the horizontal, while the same one-degree difference near 0° or 90° requires multiples of actual angular separation — two nearly flat segments hide a twofold slope difference, and two nearly vertical ones compress it into an unresolvable band. Trend reading is also direction-anchored: the horizontal serves as the "no change" reference, so near-flat segments read as "stable" even at half the slope of their neighbours. Stack these properties and the transmission efficiency of slope information is set by the angle distribution, which aspect ratio and value range jointly determine — tuning the key comparison segments into the sensitive band is a hidden duty of line-chart layout.

Where it stops holding

Angle reading covers only slope questions (faster, accelerating, turning); level questions (absolute height) go through position reading, unrelated to angle, and flat charts serve them well. "Everything at 45°" is unattainable across many segments (first and last naturally differ); tune for the comparison that matters most, and interactive zoom can bring the currently attended segment into the sensitive band. Vertical lines have infinite slope and horizontal ones zero; comparing across those two classes is asymmetrical by nature — be slow to conclude "which is faster" when the chart mixes them.

Applying it

• Scan the orientation distribution before finalizing (tools measure segment angles directly) and confirm the main comparison segments sit within 30°–60°; otherwise adjust the aspect ratio first, redesign second.
• Give key comparison segments a reference line (same-period-last-year flat line, target line), replacing the default horizontal anchor with an explicit benchmark.
• Verification: ask readers which of two nominated segments is steeper; repeat on well- and poorly-distributed versions — the accuracy gap is this leaf demonstrated.

Related

• Same group: U3.07.1 Aspect ratio changes the visual impression of slope · U3.07.3 Extreme aspect ratios amplify or flatten change
• Nearby: U3.06.1 Uneven sampling plotted at even spacing distorts rates of change · U2.03.1 Line charts suit trends over ordered variables
• Search terms: slope comparison · orientation discrimination · trend reading · line chart angle