Analyzed alone, the Planck 2018 temperature spectra prefer a spatially closed universe, with -0.095 < Ω_K < -0.007 at 99 percent confidence (Di Valentino, Melchiorri and Silk 2020). The pull comes from the same feature as the lensing-amplitude anomaly: the acoustic peaks are smoothed as if lensing were 10 to 20 percent stronger than the flat best fit predicts. Combining with BAO restores flatness, but Planck and BAO disagree at more than 3σ once curvature is free, so the cure mixes mutually inconsistent datasets.
The fit assumes every source of acoustic-peak smoothing is gravitational lensing by the catalogued matter distribution in a flat FLRW background. When the data carry more smoothing than that accounting allows, the model has nowhere to put the excess except the geometry itself, and the curvature parameter bends to absorb it.
SCT replaces the hot-dense-center with a superluminal collision, and with it the lone local matter field becomes an effective potential Φ_eff(r) = Φ_local(r) + Φ_mesh(r) (P51): the coherent superposition of comoving parent-frame contributions adds real lensing convergence along every line of sight (P50, P52, P53). From this single change the closed-universe preference becomes an artifact of omission. Planck's peaks really are smoothed more than catalogued matter can explain, and a fit with no mesh term absorbs the surplus into Ω_K < 0.
The quantitative chain already exists: the superposition boundary condition S(z₀) of about 4.4, integrated over the CMB lensing kernel, yields a smoothing excess of order A_lens of 1.18 (Paper 3, Section VII.3), which is exactly the signal the curvature parameter was bending to soak up. The geometry itself stays flat for an independent reason: SCT derives near-flatness mechanically from virialization of the collision remnant, 2K + U = 0 (P69), with no inflation and no tuning. The closed reading is a homogeneous template absorbing an inhomogeneous gravitational contribution, the same fitting artifact that turns environment-dependent Λ_eff into evolving dark energy.
This is the same mesh signature that appears as the lensing-amplitude anomaly itself, as the S₈ deficit, and as the growth-index excess: one unmodeled contribution read four ways by four parameterizations. With Φ_mesh in the model, Planck and BAO agree on flat geometry and the trilemma of closure, systematics, or broken concordance dissolves. There is no need to invoke positive curvature, exotic lensing physics, or dataset surgery.
The PR4/NPIPE reanalyses reduce the lensing excess to 1.04 to 1.10 and ACT DR6 measures A_lens consistent with unity, so the mesh-lensing amplitude SCT points to is under live observational pressure. The prediction stands or falls with CMB-S4.
Future CMB lensing measurements (CMB-S4, Simons Observatory) converging to A_lens = 1.000 +/- 0.005 would rule out any coherent mesh lensing contribution, removing SCT's account of the curvature pull entirely. Independently, a joint Planck plus BAO fit that includes a mesh-like lensing term and still demands Ω_K < 0 at high significance would show the curvature preference is not an absorption artifact.