| Rank | ID | Fam | Class | n (work) | n_orig | GR | tailDom | κ_conn | iso | m_local | F=GR−D | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 ★ | QMI-spec | QMI | FULL | 1,684 | 1,684 | 1.0000000 | 0.9978294 | 271,999 | 0 | 0.0021706 | +0.0021706 | Canon QMI |
| 2 | QMI-pre | QMI | FULL | 3,365 | 3,365 | 1.0000000 | 0.9980771 | 10⁶ | 0 | 0.0019229 | +0.0019229 | |
| 2 | QMI-gap | QMI | FULL | 2,523 | 2,523 | 1.0000000 | 0.9980771 | 10⁶ | 0 | 0.0019229 | +0.0019229 | |
| 1 ★ | ZEE-trip | ZEE | GIANT | 127,335 | 254,669 | 0.9978168 | 0.9995552 | — | 36 | 0.0478168 | −0.0017384 | Canon ZEE † |
| 2 | ZEE-sing | ZEE | GIANT | 17,689 | 17,689 | 0.9977952 | 0.9992884 | — | 30 | 0.0477952 | −0.0014932 | |
| 3 | ZEE-full | ZEE | GIANT | 169,144 | 338,288 | 0.9975346 | 0.9996097 | — | 48 | 0.0475346 | −0.0020751 | approx † |
m_local = G₁ = 1 − tailDom. Higher m_local = lower tailDom = farther from FULL/GIANT boundary. All three are FULL → Theorem 7.2 holds.
| Rank | ID | tailDom | m_local | κ_conn | Class |
|---|---|---|---|---|---|
| 1 ★ | QMI-spec | 0.9978294 | 0.0021706 | 271,999 | FULL |
| 2 | QMI-pre | 0.9980771 | 0.0019229 | 10⁶ | FULL |
| 2 | QMI-gap | 0.9980771 | 0.0019229 | 10⁶ | FULL |
m_local = G₂ = GR − 0.95. Higher m_local = higher GR = farther from HARD boundary. All three are GIANT → Theorem 7.2 holds. ZEE-trip margin gap = 0.000022 from ZEE-sing (tight).
| Rank | ID | GR | m_local | iso | Note |
|---|---|---|---|---|---|
| 1 ★ | ZEE-trip | 0.9978168 | 0.0478168 | 36 | approx † |
| 2 | ZEE-sing | 0.9977952 | 0.0477952 | 30 | exact |
| 3 | ZEE-full | 0.9975346 | 0.0475346 | 48 | approx † |
All three Zeeman encodings transition from TAIL (Phase Mapping, n≈2000) to GIANT (direct runs, n=17,689–338,288). The transition is driven by tailDom(n) dropping below 1.000 as magnetic sub-levels are resolved: at low n, outlier gaps dominate (tailDom=1.000 → TAIL condition); at full n, bulk sub-levels dilute the outlier fraction (tailDom<1.000 → GIANT). A logarithmic interpolation estimates n_crit ≈ 2,000–3,000 for singlet Zeeman. This is a parametric geometric path in the (tailDom, GR) chart: a trajectory tailDom(n) crossing the TAIL threshold as n increases — the first dynamic geometric result in the corpus.
All three Zeeman encodings simultaneously satisfy G₁ = 1 − tailDom ∈ [0.00039, 0.00071] (near tailDom=1.0 boundary) and |G₃| = |GR−1| ∈ [0.0022, 0.0025] (past GR=1 threshold). Both boundaries are approached at the same scale (ratio G₁/|G₃| ∈ [0.16, 0.32]). This is the corner at (tailDom=1, GR=1) — the Zeeman family approaches it from the GIANT side with both decisive coordinates near their thresholds. The corner is a property of the physical Zeeman fine-structure gap distribution, not an artefact of any single encoding.