The Azimuthal-Average Test in the New Orientation: A Registered Prediction for the CMB Axis with the Axis Hot and the Cold Spot Cold

Registration, frozen before any fit under it is run, of the azimuthal-average test of the cosmic microwave background axis (l, b) = (-100, 60) degrees rerun in the opposite orientation to the founding registration (DOI 10.5281/zenodo.24102852, whose prediction failed on execution, DOI 10.5281/zenodo.24115737): the axis hot and the Cold Spot cold. The registration shows, with the four sign combinations tabulated, that no sign flip of the founding profile realises this orientation -- the founding extremum condition fixes a unique ratio (b/a = +2.1376) on which the axis pole and the Cold Spot latitude always carry the same sign -- and derives the new template from the wider {l = 2, 3} family by a frozen rule (orientation conditions bounding the ratio to (-1, -0.34111], depth maximisation under the founding no-cancellation cap, and an axis-dominance condition), giving dT(theta') = C [P2(cos theta') - 0.6776 P3(cos theta')] with extrema at 0, 33.80, 103.93 and 180 degrees (hot, hot, cold, hot), the Cold Spot latitude at ratio -1.0000 to the pole height, the 90-degree witness cold, and the cap C <= 21.69 microK. Numbered predictions P1-P5 with refutation thresholds are frozen in REGISTRATION.md, with the data (Planck PR3, four component-separated maps), the statistic (5-degree annuli, single-amplitude fit, 10,000-axis rotated null in all four maps, seed 20261112), and the boundary paper's bearing stated in advance (its allowed region permits C in [-28.86, +23.97] microK along this ray; the cap point lies inside the region, so the boundary does not pre-exclude this template). AI use disclosure: this work was produced with Dave, an AI research system; the derivation was executed and the text drafted by Dave under the author's direction, and every number is reproducible from the deposited code.