Free field constants leave at most 4.13% of additive magnitude dipole information in 1,623 DES events with fixed Dovekie covariance

We test how constant shifts between observing fields affect an additive magnitude dipole in the public DES-Dovekie distances. We use all 1,623 DES events at a fixed release commit, exact event positions, and the marginal statistical and systematic covariance. The baseline allows a global magnitude constant and a local matter-density term about flat Lambda CDM with Omega_m=0.3. With free constants for all 15 exact field labels, the three generalized dipole information fractions are 0.0006252, 0.0015615, and 0.0412780. Four free sky-region constants give 0.0014992, 0.0045675, and 0.1274713. Thus most baseline additive dipole information depends on constraints between fields. If each field-label shift lies between -0.010 and +0.010 mag, the maximum change in the baseline fitted dipole vector is 0.435414 mag. Cube enumeration and an independent convex-hull calculation agree. This is a chosen sensitivity scale, not a measured calibration error. Statistical-only, fixed-density, and redshift controls are reported. Independent SVD, joint-fit, precision-matrix, and direct-perturbation checks pass. The primary residual dipole has nominal Gaussian p=0.558; the free-field fit has p=0.282. These results do not establish physical anisotropy or assess scalar dark-energy constraints. We preserve code, input hashes, row matches, transformed data, response matrices, figures, and reproduction instructions.

DES-Dovekie · supernova distances · additive magnitude dipole · field constants · calibration sensitivity · Fisher information · public data · reproducible research