FRIENDS, RIVALS, AND LONE WOLVES: HOW SEVEN OPTICAL METRICS RELATE ACROSS 92,862 GEMSTONES

Every gem cutter and jewelry lover has wondered about the hidden trade-offs inside a faceted crystal. If you cut a gemstone for bold dark-and-bright Contrast, do you have to give up white Brightness? Does chasing explosive Sparkle unlock more rainbow Fire, or do they fight each other? And why do some minerals throw far wider rainbow flashes than you would expect from their refractive index alone?

To find out without any scoring distortion, we looked at all 92,862 gemstone designs in the OffTheDop catalog using their raw, un-squashed physical measurements across all nineteen tilt angles (without any Hearts-and-Arrows baseline ratios or diminishing-returns squashes). Out of all the numbers, three clear lessons stand out.

1. Windowing Ruins Everything, While Contrast Unlocks Free Rainbow Fire#

When you tilt a gemstone in your hand and light leaks out the back pavilion instead of bouncing back to your eye, gem cutters call that see-through patch a window. Across all 92,862 designs, Windowing is the universal enemy of gemstone optics:

As the left panel shows, whenever a design leaks light out the back across tilt, every other metric collapses with it: Sparkle (-0.75), Rainbow Fire (-0.74), Tilt Stability (-0.69), Contrast (-0.56), and Brightness (-0.40) all drop together. A ray that escapes through the pavilion cannot return white light, cannot blink on and off as you rock the stone, and cannot split into a rainbow prism.

Once you seal the pavilion against windowing, look at the surprise on the right panel: our optimizer placed 0% weight on Rainbow Fire, yet Fire correlates at +0.75 with Sparkle and +0.59 with Contrast! Why do you get rainbow fire for free when you cut for contrast and sparkle? Because a flat table facet acts like a windowpane where red and blue rays exit parallel and recombine into white light. To carve crisp dark-and-bright contrast and rapid sparkle, the optimizer builds multi-tier sloped crown facets over a deep, window-free pavilion, and those angled crown facets happen to be the exact prisms needed to fan white light into rainbow fire.

2. Simpson's Paradox In Brightness Vs. Contrast (+0.28 Overall Vs. -0.69 Sliced)#

If you correlate raw Brightness and raw Contrast across the entire catalog at once, you get a mildly positive number (+0.28). At first glance, that seems backwards: aren't Brightness and Contrast supposed to be opposites? After all, a gemstone's face is a fixed 100% pie of surface area. Every facet zone that turns dark to create crisp contrast borders cannot simultaneously be bright white to maximize light return!

What we are looking at is a textbook case of Simpson's Paradox (where an entire data cloud slopes up (/) while every single slice inside the cloud slopes down (\)), reinforced by Berkson's Paradox from our catalog score cutoffs. Watch how the correlation flips in three steps:

1. Step 1 - Lumped All Together (+0.28, slopes upward): When you mix all eight minerals and all quality tiers into one giant cloud, leaky low-RI stones sit near the bottom-left (because windowing hurts both Brightness and Contrast at once, and dim facets cannot clear the minimum brightness jump needed to form contrast edges), while sealed high-RI stones sit near the top-right. That makes the combined cloud tilt upward (+0.28).
2. Step 2 - Grouped by Mineral (+0.03 in Cubic Zirconia, +0.04 in Diamond and Moissanite, goes flat): The moment you stop mixing Quartz with Cubic Zirconia and Diamond and look inside a single window-free mineral, almost all of that positive correlation vanishes (+0.28 drops to +0.03).
3. Step 3 - Sliced by Quality Score Tier (-0.69, every slice slopes downward): Within any mineral, our optimizer values Contrast four times more than Brightness (42.6% versus 10.6%) and caps Brightness rewards five times harder once a stone reaches decent light return (around 38%), while our catalog cleanup purges any design below the mineral's minimum score cutoff (0.76 in Quartz up to 0.90 in Moissanite). As a result, every quality tier forms a downward-sloping diagonal slice where Brightness and Contrast trade off at -0.69 across the catalog (-0.38 in Quartz, -0.64 in Sapphire, -0.70 in Cubic Zirconia, -0.60 in Diamond, and -0.63 in Moissanite). To stay in a top score band, any stone with lower Brightness only survives if its Contrast is sky-high, and vice versa! If you want maximum white light return when designing in ProFacet Studio, raise the Brightness weight slider explicitly so the optimizer does not trade your surplus light away for extra contrast.

3. The Diamond Fire Surprise: Why Cubic Zirconia Beats Diamond By Nearly 50%#

As you step up the mineral ladder from Quartz (RI 1.54) to Moissanite (RI 2.67), windowing plunges sixfold (from 21.5% down to 3.4%) and raw Sparkle triples. You might expect Rainbow Fire to climb smoothly with refractive index too. Instead, raw Fire leaps to 5.21% in Cubic Zirconia (RI 2.15), drops by a third to 3.49% in Diamond (RI 2.42), and then climbs back to 5.56% in Moissanite (RI 2.67):

Why does Diamond show a third less rainbow fire than Cubic Zirconia even though Diamond has a much higher refractive index? Three laws of crystal optics explain the drop:

1. Higher Refractive Index does not automatically mean higher Dispersion: Refractive index measures how hard a crystal bends light overall; dispersion measures how wide it splits red light from blue light. Real Cubic Zirconia has a material dispersion of 0.060, whereas Diamond's dispersion is only 0.044 (about 27% lower than Cubic Zirconia, while Moissanite reaches 0.104). You can see the exact same effect on a smaller scale between Spinel (RI 1.72, dispersion 0.020, fire 2.51%) and Sapphire (RI 1.76, dispersion 0.018, fire 2.37%).
2. Snell's Law squeezes the rainbow fan narrower at higher RI: Why is Cubic Zirconia's ray-traced fire 49% higher than Diamond's (5.21% versus 3.49%) when its raw dispersion is only 36% higher? Because when light enters a crown facet, Diamond's higher refractive index (2.42) bends all colors closer to the facet normal, compressing the angle between red and blue rays inside the stone. A crystal's effective prism spread scales with dispersion divided by refractive index: Cubic Zirconia's relative spread is 0.060 / 2.15 = 2.79%, compared to 0.044 / 2.42 = 1.82% in Diamond. That gives Cubic Zirconia a 53% wider rainbow fan, matching our 49% GPU ray-tracing jump almost to the percent.
3. Higher surface glare and tighter critical-angle trapping: As refractive index climbs from Cubic Zirconia (2.15) to Diamond (2.42) and Moissanite (2.67), outer surface glare rises from 13% to 17% and 21%, while the shrinking critical-angle exit cone (from 27.7 degrees down to 24.4 and 22.0 degrees) traps more steeply angled rainbow rays inside the stone instead of letting them exit cleanly on the first pass.