Lift — $n=36$ at aspect 10

Same Newton $Q=I/D$ and mass-opt cut $Q>10/11$ as q10. The q10 bottleneck was the compact $P_{\max}$ tax at $n=35$.

Raising the bin count to $36$, target $0.9117$: SLSQP $\varphi=0.911881$, $P$ error $0.003385$, certified $\gamma=0.908315$, $1/\gamma=1.100940$, printed $1.1010$. Cut $10/11>\gamma$. Faces $68{,}719{,}476{,}735$, copositive, $16296$ residual skips, $\min m^\top Mm>6\cdot 10^{-4}$, $\min\varphi=0.911885$.

$R\le 9$ with the mass-opt cut cannot beat $1.1013$. The withdrawn $1.1168$ stays withdrawn. Finite $Z$ and $N_0(Z)-Z$ stay leftover. A higher-target probe on the $n=34$ or $n=35$ matrix is not an $n=36$ dent.

Replay: ../run_all.sh. Certificate: ../certs/lift.json.