V overshoot — 6 translateY variants

Why the V floats

The V SVG path is M10 14 L31 50 Q32 52 33 50 L54 14 L42 14 L32 32 L22 14 Z. ViewBox 10 14 44 38. The quadratic apex hits y=52 — exactly the viewBox bottom. With preserveAspectRatio="xMidYMax meet" + vertical-align: baseline, the SVG element's bottom edge sits on the line baseline, putting the V's geometric tip on the baseline.

But the V's visible mass is concentrated at the top. A flat-bottomed letter like "e" has horizontal mass at the baseline; the V has a single point. Visually the V reads as floating above where "ecreal" is grounded.

The fix is the same pattern we used for the dot in rev-4: transform: translateY(N×em) on the .bk-v element. Em-based so the descent scales proportionally with font-size. The V tip ends up sitting slightly below the baseline (visible in the comparison via the clay baseline ruler in each specimen stage), and the visual mass of the V is properly grounded with "ecreal".

Pointed shapes may want more overshoot than round shapes. The dot (round) got 0.025em. The V (pointed) may want the same, or more. Variants below range from 0.015em (less) to 0.060em (more).