Mass bends the light that passes it. A star behind the Sun shifts outward from the limb; a source dead-behind a lens smears into a ring. The engine computes the deflection α = 4GM/bc² from first constants and rebuilds the 1919 eclipse number — 1.75″, exactly twice what a Newtonian "photon-bullet" would give. Rendered, not quoted.
source Dyson, Eddington & Davidson (1920), A Determination of the Deflection of Light by the Sun's Gravitational Field…, Phil. Trans. R. Soc. A 220, 291–333 · Einstein (1915) light deflection — doi:10.1098/rsta.1920.0009
Light follows a null geodesic through spacetime curved by mass M. For a ray with impact parameter b (closest approach), General Relativity gives a deflection:
α = 4GM / (b·c²)Grazing the Sun's limb (b = R☉) this is 1.75″. The Einstein-ring radius for a source dead-behind the lens:
θE = √(4GM·Dls / (c²·Dl·Ds))Both scale with mass: α∝M, θE∝√M. Inverse in b — closer grazing bends more.
Mass bends light — Einstein 1915, confirmed by Eddington 1919. The deflection α=4GM/bc² is twice the Newtonian value: half from time-curvature, half from space-curvature. This is the light-geodesic through curved spacetime, the same object that in the-schwarzschild-metric defines the well the ray is falling through; align source/lens/observer and the arc closes into an Einstein ring.
Re-runs the full selfcheck live and reports. Flips red the instant the engine (or a tamper in window 6) changes a proven value.
Checks: 1.75″ solar deflection · factor-of-2 over Newton · inverse-in-b · ring ∝√M.
Fixed constants (SI) and the geometry of a grazing ray:
Live: a background star's true ray (dashed) vs. the bent ray (solid) grazing the Sun, and the Einstein ring for a perfectly aligned source.
Proven at boot:
wall "Just use Newton — a photon is a fast particle falling in a 1/r² field." That prediction (Soldner 1801) gives 0.875″ — real, but half the measured value. Newtonian gravity bends the trajectory but leaves space flat; GR curves space too, doubling the bend. The 1919 result did not merely confirm deflection — it killed the half-value. The engine here reproduces the whole.
"Light has no mass, so gravity cannot touch it."
→ Gravity acts on energy/momentum and on the path itself; massless light still follows the curved geodesic.
"The 1919 error bars were too big to prove anything."
→ Sobral+Príncipe favoured 1.75″ over 0.875″ and over zero; later radio/VLBI pinned it to <0.02%.
"A lens magnifies but conserves surface brightness, so it adds no light."
→ True for brightness, but it enlarges solid angle — total flux is magnified; that is how lensing reveals faint high-z sources.
Swap the GR factor 4 for the Newtonian 2 in the deflection law. The solar bend collapses to 0.875″ — the Witness (7) catches it.