Methodology

12 straight hues and 57 exact two-color descriptions. 741 GIA-attributed historical sold observations (509 Included; 232 Qualified), plus three excluded legacy rows. Latest update adds 300 transactions, A445–A744, and corrects A29/A50 and enriches A18/A36 with original lot evidence. Selected sample, not a census.

Exact lifetime GIA counts were not located for any of the 12 straight hues. Column B states this explicitly. Column C reports the source’s historical counts or percentages, with the period alongside. Do not interpret study totals as lifetime grading or mining populations.

No secondary hue in the grade. Fancy, Fancy Light, Fancy Intense, Fancy Vivid, Fancy Deep and Fancy Dark describe tone/saturation, not extra colors. Faint/Very Light/Light are included in some GIA populations.

Interprets the user’s “dual modifiers” as two hue terms: one modifying color plus the dominant color, e.g., orangy yellow. The last color is dominant. Three-color grades such as grayish bluish violet are excluded.

Orangy yellow, orange-yellow, yellow-orange and yellowish orange remain separate. Pink and red wording and gray/grey spelling are handled explicitly; gray/grey are normalized to gray. A study’s broad group is not an exact grade census.

GIA hue-circle names plus supported pink, brown and gray variants; includes four added pink–brown directions/strengths. This is research coverage, not a claim that every theoretical permutation exists or has a published frequency.

Published evidence keeps direct counts, bounds and rounded shares separate. Inferred point inputs appear only on Extrapolation with their assumptions; they are not published GIA counts or strict population floors.

The 2025 red paper reports a broad submission share and, separately, a 399-stone Fancy Red FTIR dataset (156 known Argyle; 3 reportedly Brazilian). It also identifies 24 public-record unmodified Fancy Red stones >1 ct. These populations have different scopes and may overlap; none is a complete lifetime census.

Studies have different periods, inclusion criteria and denominators. Never add overlapping cohorts or rank colors by their percentages across studies. GIA submissions are selected laboratory intake, not all mined diamonds; unique-stone deduplication is not established.

Broad green-component count and chart shares are not reconciled; blue prose count exceeds a naive rounded-denominator product. Direct count statements take precedence. GIA 2025 orange-mechanism and red-appendix reviews did not yield a new exact per-hue census.

The latest 300 transactions come from Christie’s (257), Phillips (34) and Heritage (9). Public catalogues establish GIA attribution, natural color, exact grade and target weight; completed sale results are linked per row. Wider archive searches were screened without counting dealer retail asking prices, estimates, current bids, unsold lots or owner repurchases.

Total realised auction lot price divided by the target colored diamond carat weight (sum for a matched pair). It includes setting, side stones, branding and provenance. It is not a separately appraised loose-diamond value.

Price basis is recorded per row. Published results for the 300 additions include buyer premium. The full lot price is divided by the GIA target carats, with setting and accent value not deducted. Lots with substantial other stones or provenance/design value are Qualified. Earlier Qualified result chains and price-basis limitations remain disclosed in their original rows.

Original realized price and currency are preserved. Reported USD is used when available; otherwise USD is calculated with the sale-date FRED rate, or the preceding observation for non-reporting days. Nine Phillips HKD sales on 29–30 September 2026 use the latest published 25 September rate, 7.8439 HKD/USD, because sale-day data was not yet published; all nine are Qualified. Their local sold prices are verified.

House currency equivalents can be rounded or inconsistent. Preserve the selected published USD amount and document conflicting figures. Do not silently substitute today’s exchange rate.

All 300 additions have documentary GIA natural-color and completed-sale evidence: 225 Included and 75 Qualified. Qualified records carry whole-lot comparability, repeat-sale, report-conflict, provenance/design or provisional FX limitations and remain separate from Included medians. Older Qualified records retain their earlier disclosed limitations. Restored A50 is Included; corrected A29 remains Qualified. A36 is moved to Qualified because royal provenance and JAR craftsmanship materially affect its lot price.

Low/high/median are sample statistics, not market limits or current appraisals. Intensity Comparisons separates exact grades and also gives Included observations dated 2020 onward. No market price is filled for an unobserved grade. Uneven sample sizes, weight, clarity, settings and provenance still limit comparisons.

Repeated identifiable gems are not automatically independent market supply. Earlier known repeat transactions are Qualified where a later usable comparable is retained. Matched pairs/suites with the same exact grade use combined target weight and count as one transaction. Mixed target grades are excluded when the price cannot be allocated.

Verification means documentary verification against public auction catalogues/report descriptions and completed results. It does not mean independent authentication through GIA Report Check. Report numbers are published for 299 of the 300 additions. Christie’s lot 1886104 cites a GIA certificate dated 19 May 2000 with natural Fancy Deep Brown-Yellow and 14.89 ct, but publishes no report number; this limitation is explicit in its row.

Straight Colors and Two-Color Grades summarize the expanded prices. Intensity Comparisons provides all 621 hue/intensity coverage rows, with price evidence and supply columns Q–W. Intensity Supply holds dated counts, percentages, denominator definitions, limitations and primary sources. GIA Evidence, Argyle Evidence and Extrapolation retain their prior corrections.

Revisit original result pages and GIA publications before using this research for a transaction. This is a dated evidence workbook, not an appraisal, complete auction census or access to GIA’s private database.

Brown, gray, Fancy Black and Fancy White were absent from the original requested seven-color list, as was straight Pink. Colorless D–F is not a fancy-color hue. Champagne/cognac are marketing descriptions for brown-related grades; chameleon describes color change, not a separate straight hue. GIA terminology source: G12.

Keep all-origin laboratory populations, selected tender stones and rough production separate. Argyle stones can already occur in GIA cohorts. Never add overlapping populations or directly convert carats to stones. Extrapolation applies conditional ratios only; those ratios are not evidence of actual grading intake.

Argyle and GIA use different color scales. Argyle RED can correspond to GIA Fancy Deep Pink (Kalina), and Argyle 2P to GIA Fancy Vivid Purplish Pink (Spectre). Classify by the stated GIA grade for this workbook; see AR13–AR16.

Rio Tinto lists 12 GIA Fancy Violet diamonds in its final 2025 tender; individual reports were not audited. The 2018 study separately reports 0.1% straight Violet within >15,000 blue/gray/violet stones. A 15-stone proxy appears only as an explicit Extrapolation input. No public tender prices were located.

GIA 2014 dates the inaugural tender to 1985; Argyle’s current retrospective says 1984–2021. Period labels are preserved as reported. 2023–2025 Beyond Rare tenders are separate from the Signature series, and can include reoffered old stones.

300 net-new transactions appended as A445–A744: 132 Fancy, 93 Fancy Intense, 39 Fancy Deep and 36 Fancy Vivid. No new pale, Light, Fancy Light or Fancy Dark observations. A matching-grade pair is one transaction with summed target carats. Repeated gemstones in distinct sales are flagged. Four rediscovered listings were matched to existing A18/A29/A36/A50 and corrected there, not counted as additions. Sale-room notices govern amended weight, clarity and grading.

The 16 September revision removed misleading count proxies but missed useful published evidence. The yellow ~23,435 inference was not a cumulative grading count; the original color coverage also omitted Pink and other categories. The subsequent source audit was incomplete.

Retained ~50 unmodified purple, >100,000 historical broad-yellow reports and >500,000 fancy-color study submissions. Corrected public-record Fancy Red >1 ct from <24 to exactly 24, clarified <0.07% as a combined share bound, and added the separate 399-stone analytical dataset with origin subsets. No claim of an exhaustive error-free audit is made.

Corrected all 2006–2025 world production inputs from KP annual summaries and sourced Argyle production from Rio Tinto reports. Fixed scenario envelopes, linked index assumptions, included both yellow population cases, removed the purple Argyle proxy and unsupported claims. These conditional scenarios remain separate from published counts.

Nine labels are checked for every hue. This is a research coverage grid, not proof that all 621 combinations are GIA grades. Red-dominant colors and Fancy Black/White use Fancy only. Faint/Very Light/Light unmodified Yellow and Brown are handled on the D–Z scale. Other empty cells are not declared impossible.

These counts describe historical submitted/analyzed diamonds or selected tender stones. They do not measure current sale inventory, annual mine production or lifetime GIA counts. Separate cohorts may overlap and must not be added. Auction transaction counts do not measure supply.

GIA Table 3 supplies exact Fancy/Intense/Deep/Vivid counts for Pink, Purple-pink, Purplish pink, Orangy pink, Brownish pink and Brown-pink. The period is not dated precisely. Its table percentages are hue shares within intensity, so this workbook instead calculates each intensity’s share of that exact hue’s FOUR listed stronger grades. Pale-grade denominators are not known by hue.

2002 Table 2 describes all 1,490 pink-dominant study stones. The 1998/2003 yellow distributions include modifiers. Argyle 1995–2014 intensity shares combine selected pink-to-red tender hues. These related distributions appear in Intensity Supply but are never assigned as an exact single-hue distribution.

Blue 2018 individual percentages are visual readings of unlabelled bars, rounded to the nearest percentage point; they total 99% and are not normalized. The caption’s 54% combined pale range remains separate. Pink 2018 per-bin bars were not digitized because they do not reconcile cleanly with the published combined share. No intensity count is generated from a rounded share times a population bound.

A zero count in the 2002 exact table is a true zero only in that historical study. Blank count/share cells mean unavailable. A zero auction n means no qualifying observation in this workbook, not no diamonds or no historical sales. Purple’s 42% Intense-or-Vivid and Pink’s 54% Faint-through-Light cannot be split into separate bins.

My earlier work confused dated yellow cohorts with cumulative GIA grading, omitted Pink and other hues, and relied on too few high-intensity prices. I also retained IGI and unverified-lab rows. Three unsupported legacy records remain excluded from statistics. A50 was previously excluded for missing attribution; its original GIA catalogue has now been recovered. Missing supply evidence remains explicit instead of unsupported totals.

USD/ct, summaries and the linked supply distribution calculations are formulas. The cohort selection and source-to-row mapping are documented research decisions, not a live database. Adding new rows requires updating the summary references and duplicate/qualification review.

A59 White and A61 Black are IGI, not GIA; A06 Yellow lacks target-specific GIA evidence. These three remain Excluded. A50 is restored after finding original Christie’s 2005 lot 4503848 with GIA 13780794 and natural Fancy Vivid Greenish Blue. A29 now uses exact CHF 7,781,000 instead of rounded CHF/USD figures. A17 remains supported by GIA 2205845362 in the 2026 New York catalogue.

Latest additions cover Fancy, Fancy Intense, Fancy Vivid and Fancy Deep only. The earlier nine-grade grid is preserved for historical context; filter column B to these four grades. White and Black use Fancy in the retained GIA framework, so Intense/Vivid/Deep cells for these hues are not filled artificially.

The 300 additions include 176 straight-hue transactions: Yellow 122, Blue 33, Pink 8, Green 5, Gray 4, Black 2 and Orange 2. The other 124 have exact two-color grades, including one Gray-Violet and two Pink-Purple. No further pure Purple, pure Violet, Fancy White or straight Brown sale passed the screen. Modifier sales remain separate from straight hues; gaps do not establish that no such sales have ever occurred.

Auction counts describe sales with GIA attribution; they are never substituted for supply. Intensity Supply retains GIA study cohorts and Argyle tender evidence only with their stated grading scope. Argyle production or Argyle-only color grades are not counted as GIA-graded stones. Exact lifetime GIA totals and most hue-by-intensity supply cells remain unpublished.