Lift — tighter compact γ at aspect 10

Same Newton $Q=I/D$ as q3. The q3 bottleneck was the compact $\gamma$ at $R=12$ ( $\varphi=0.9055$ , $P$ error $\approx 0.00597$ ), not the cut $12/13$ .

Raising $\varphi$ and $n$ on the mid-radius Rayleigh:

$R=12$ , $n=22$ , $\varphi=0.90685$ already gives $1/\gamma\approx 1.1100$ .
$R=10$ , $n=26$ , $\varphi=0.9091$ gives $\gamma=0.904414$ , $1/\gamma=1.105688<1.1057$ .
Cut $10/11\approx 0.90909>\gamma$ , so the mass-opt dichotomy lifts the class bound.

$67{,}108{,}863$ faces, copositive, $23$ residual skips, $\min m^\top Mm>4\cdot 10^{-4}$ . Stdlib rebuild of $A$ matches the stored matrix to $10^{-15}$ .

Same §7 chain. Printed

N_c<1.1057Z+3.946\,Z^{1/3}\qquad(Z\ge 4).

Not the withdrawn $1.1168$ . The aspect- $\le 4$ number $1.1087$ is not quoted as class-only; the unrestricted leading sits below it.

Finite $Z$ and $N_0(Z)-Z$ stay residue.