The DCX index is calculated as a weighted geometric mean of interpolated per-carat retail prices derived from OpenFacet matrices. Since launch, the model has been extended with kernel smoothing across carat bands and monotonic regression via PAVA—refinements that improve cross-band consistency and enforce economically logical price gradients by carat. All input data remains public, all transformations are documented, and no synthetic or off-market adjustments are applied.

The initial composition of the DCX basket is grounded in a representative selection of carat–color–clarity combinations that dominate global retail trading activity.

Diamond pricing is nonlinear and inventory-driven. While thousands of spec combinations exist, most transactions occur within a narrow band of highly liquid grades. To reflect this, the DCX index uses only GIA-certified, triple excellent (3EX), fluorescence-free natural diamonds, and only those specifications that are consistently listed across multiple top-tier retailers. This ensures the index reflects actual market behavior—not catalog extremes, auction anomalies, or internal dealer lists.

The final basket contains seven specifications selected to maximize turnover coverage while minimizing overlap and price redundancy. Together, these specs represent an estimated 80% of global natural diamond trade volume by unit count and more by total dollar turnover. Each is defined by a unique carat, color, and clarity combination, with GIA triple excellent (3EX) cut held constant across all entries.

Weights are assigned based on a combined metric of estimated global listing frequency and per-spec dollar turnover. This ensures balance between unit volume and pricing impact. For example, 1.00ct G/VS2 receives a 18% weight due to its dominance in both trade frequency and benchmark pricing, while the 2.00ct I/SI2 specification, despite a higher unit price, is capped at 5% due to thinner market depth and higher volatility.

Specifications such as 1.00ct D/IF or sub-0.30ct grades are deliberately excluded due to low trade frequency, inconsistent listing across major platforms, and heightened exposure to catalog bias or anomalous pricing. These specs tend to reflect marketing-driven positioning rather than reliable transactional benchmarks, and their inclusion would dilute the signal integrity of the index.

At the time of publication, a pricing snapshot for the selected specs is as follows:

This configuration limits concentration risk while preserving real-world pricing influence. No single spec dominates the index, and high-ticket items are proportionally bounded. The DCX basket is reviewed quarterly and adjusted only if materially justified by changes in global turnover distribution.

As with the index itself, the goal is not to forecast prices, but to reveal structure—anchoring valuation, pricing analysis, and synthetic instruments in observable retail market behavior.

→ DCX Methodology Details