The overall X-ray polarization of the neutron star Cen X-3 looks energy-independent. Flat. Featureless. Nothing to see.
Note (2026-11-15; read against the paper's full text): the essay says a disk wind scattering off pulse-phase-modulated material creates a phantom polarization component and that the components cancel into apparent uniformity when averaged. The paper's wording is more tentative: phase-averaged polarization shows marginal (not zero) energy dependence; phase-resolved analysis shows a strongly phase-dependent polarization-angle energy dependence in a few phase intervals; it models this with a Rotating Vector Model pulsed component plus an additional phase-dependent component, notes that significant deviations from the model persist in a few phase bins even then, and says its spectroscopy (phase-modulated column density and covering fraction) suggests the wind properties are modulated with pulse phase, so that phase-dependent scattering in the disk wind 'may' significantly alter the observed polarization. I found no statement that the components cancel on averaging.
Phase-resolved analysis tells a different story. The polarization angle has strong energy-dependent variations hidden inside specific pulse intervals. A disk wind scattering off pulse-phase-modulated material creates a phantom polarization component that scrambles the underlying signal.
The mechanism: the neutron star's rotation creates a time-varying polarization pattern. Scattered light from the disk wind adds a second polarization component with different energy dependence. Average over the full rotation, and the two components cancel into apparent uniformity. Resolve by pulse phase, and the structure appears.
What looked like simple behavior was an averaging artifact. The “boring” measurement contained exactly the information needed to identify a complex scattering geometry — it was just encoded in a way that time-averaged observation couldn't extract.
This is a general pattern: systems that appear uniform at aggregate resolution often contain structure that only becomes visible when you resolve along the right axis. The absence of a signal is itself a signal — it tells you that something is canceling, which constrains the geometry as tightly as a detection would.