Brilliance puts you at a lapidary’s bench with a rough stone, a set of proportions to choose, and an honest judge: light itself. Every angle you change — pavilion depth, crown height, table width, girdle thickness — re-traces thousands of rays through the stone’s real geometry, and the maps a working jeweller would use tell you what your choices did. It lives on the Engagement Rings library because it answers the question that library keeps circling: what does “cut quality” physically mean?
How to play
Everything works with a mouse or a fingertip, with keyboard shortcuts if you want them. Start with a round brilliant and drag the proportion sliders: the stone re-renders as you move, and the bench shows you three readings at once — the traced light return, an ASET map (which colours each part of the face-up view by where its light actually comes from), and an Idealscope view (which shows light leakage as washed-out pale zones). A plain-English verdict translates the maps: where you are leaking light, where contrast comes from, what a grader would likely say.
There are five stone shapes to work, a cut-grade estimate for each attempt, and an optimizer that searches proportion space for the best light return so you can compare your hand-cut result against what the search finds. When you are ready to leave the guardrails behind, open the chisel: it lets you cut entirely new facets — each cut is one more plane through the stone — and see what the light does about it. Nothing about the game is timed. It is a bench, not a race.
What the simulation actually models
The optics are the game. Rays enter the stone in three dimensions, refract at each facet according to diamond’s refractive index, reflect internally when they strike a facet past the critical angle, and split by wavelength for fire — the coloured flashes that dispersion produces. Fresnel’s equations decide how much light each surface returns versus transmits. The ASET and Idealscope maps are not textures; they are read out of the traced rays, the same way the physical tools work when a grader holds a real stone under a structured light hemisphere.
That is why the classic lessons of cut theory simply emerge from play. Cut the pavilion too shallow and you get a fish-eye; too deep and the centre goes dark as a nailhead — not because the game has rules that say so, but because that is where the traced rays actually go. The cut-grade estimate is likewise traced instead of asserted: it is computed from your stone’s light behaviour, not looked up from a table of “ideal” numbers.
How it was built
Brilliance is hand-written WebGL and JavaScript — no engine — with the ray tracer running in a web worker so the bench stays responsive while thousands of rays are in flight. The optics live in their own modules (tracing, Fresnel, dispersion, illumination) and are covered by unit tests that check the physics against known results, from refraction angles to grading behaviour. The chisel tool has a physical ancestor too: the chisel you swing was modelled in Blender and baked into a mesh module, and each facet you cut is implemented as one more half-space intersected with the stone — the same constructive geometry a cut designer’s software uses.
The honest confession of the commit history is that the first version traced a flat slice through a round stone; the current bench traces the full three-dimensional geometry and then grades it. If the maps disagree with your intuition, trust the maps — that is rather the point.