The safety envelope that squared a 15-point gap

The same data on a cream background, kicked “The same eight numbers · Euro NCAP 2025” and headlined “A 15-point gap, drawn as a 4.4× one”. A subhead reads “Every percentage from the radar, on one zero-based scale — one length per number, nothing squared, no order to argue about.” A dot plot on a zero-based axis running 0 to 100 shows four rows, each with a blue dot for the MG 3 Hybrid+ and a gold dot for the smart #5 joined by a grey connector: adult occupant 74% and 88%, child occupant 73% and 93%, vulnerable road users 81% and 84%, safety assist 69% and 92%. A note beside the vulnerable-road-users row reads “Three points apart — the pillar where the two cars nearly meet.” A note beside the safety-assist row reads “The widest pillar: 69 to 92 is 1.33×. Measured from the radar’s 60% floor it becomes 9 against 32 — 3.6× — before the area squares it.” A bracket under the axis marks the four-pillar averages, 74.25% and 89.25% — 15 points, or 1.20×. A card headed “Where the 4.4× came from” explains that the radar’s rings start at 60%, so each spoke is the score minus 60, turning a 1.20× gap in the averages into a 2.05× gap in spoke length, and that area then finishes the job because a polygon grows as the square of its spokes: the two shapes enclose 385 and 1,690 square units, or 4.39×, where a zero-based radar would enclose 1.45×, which is just 1.20 squared. A second card, “And the spoke order moves it too”, notes that with four spokes at right angles the enclosed area is exactly half of (top + bottom) × (left + right), so it depends only on which pillars sit opposite each other — swap two labels around the circle and the MG 3’s polygon travels between 385 and 405 square units, 5% of itself, without a single number changing. A footnote credits Euro NCAP’s published category scores for the MG 3 Hybrid+ 1.5 HEV, tested 2025, four stars, and the smart #5 Pro, tested 2025, five stars, and says Euro NCAP publishes the four percentages, not a radar chart — both drawings are our own demonstration. — Showing the misleading chart

Be the first to star this exhibit

Two 2025 Euro NCAP results drawn as a four-spoke radar. Every percentage is the official published figure, and the smart #5 really does lead the MG 3 Hybrid+ on all four pillars. But the gap on the page is 4.4× where the gap in the numbers is 15 points — because the rings start at 60% instead of zero, and a radar polygon grows as the square of its spokes.

Euro NCAP 2025 test results — published category scores for the MG 3 (four stars) and the smart #5 (five stars), 2025Added August 7, 2026

01The claim

One of these cars has a much bigger safety envelope. All four Euro NCAP assessment pillars sit on a single profile, drawn from the official 2025 published percentages: the smart #5 leads on adult occupant protection (88% to 74%), child occupant protection (93% to 73%), vulnerable road users (84% to 81%) and safety assist (92% to 69%), and its shape covers 4.4× more of the chart. You don’t have to read the numbers — one car fills its envelope and the other never gets clear of the inner rings.

02The trick

Every figure on the radar is correct, and the smart #5 genuinely does lead all four pillars. The trouble is the headline statistic, because “4.4× the area” is not a fact about the cars — it is a fact about the drawing, and it is built in two steps that are both geometry rather than data. The first step is the radial axis. The innermost ring is 60%, not zero, so each spoke is drawn as the score minus 60. That turns the MG 3’s 69% on safety assist into a spoke 9 units long and the smart #5’s 92% into one 32 units long: 3.6× the length for a 1.33× difference in score. Across all four pillars the averages are 74.25% and 89.25%, a gap of 1.20×, and the 60% floor stretches that to 2.05× before a single pixel is shaded. The second step is that a radar chart asks you to read area, and area grows as the square of the spokes. Squaring 2.05 gets you to roughly 4.2; measured exactly, the two polygons enclose 385 and 1,690 square units — 4.39×, rounded up to the 4.4 on the stat card. A third knob is hiding in the same picture, and it is the strangest of the three. With four spokes at right angles the enclosed area works out to exactly half of (top + bottom) × (left + right), so it depends only on which pillars happen to sit opposite each other. Reorder the labels around the circle — a decision with no meaning at all — and the MG 3’s polygon travels between 385 and 405 square units, 5% of its own area, with no number touched. (This exhibit is our own demonstration, drawn in the style of a comparison slide from the published scores; Euro NCAP publishes the four percentages, not a radar.)

03The fix

Put the four pillars on one zero-based scale and each number gets exactly one length — nothing squared, no ordering to argue about. Read that way the story is perfectly legible and much smaller than the radar made it look: the smart #5 leads by 14 points on adult occupant protection, 20 on child occupant, 23 on safety assist, and by three points on vulnerable road users, 84% against 81%. That last pillar is the one worth noticing, because it is where the two cars nearly meet and it is exactly what the squared ink flattens out of sight. Averaged across the four, 89.25% against 74.25%: 15 points, or 1.20×. Both cars cleared their tests — the smart #5 took five stars and Euro NCAP’s Best in Class award for large SUVs tested in 2025, the MG 3 four stars — so the honest chart is a comparison, not a verdict. If the shape of a profile genuinely matters, say for spotting a car that is strong everywhere except one pillar, small multiples of bars compare shapes without inventing an area to measure. And if the radar stays, start its rings at zero: the same eight numbers then enclose 1.45× rather than 4.4×, and 1.45 is simply 1.20 squared — the amount of exaggeration a radar hands you for free, before anyone touches the axis.