Pragnell ring makes $475,000 at Heritage, 90% above the high

The house's own numbers. Everything above rests on one outlet's report of what Heritage's website published, and the house's own pages refused a direct fetch from here, which means the sale total, the lot count and the five prices have one carrier rather than two. A lot-by-lot results page would settle three things this article cannot: whether the $7.6 million is hammer or hammer plus premium, how many of the 425 lots failed to sell, and whether any lot above $112,500 sold outside the published top five. If the total turns out to include buyer's premium, the $17,882 average overstates what consignors received, and the estimate comparisons above, which are struck against estimates that never include premium, would be comparing two different bases.

An auction estimate is a range and the high end of it is the number a specialist expects a determined buyer to stop at. It is set before the sale, published in the catalogue, and it is not a reserve: the reserve, which is the price below which the lot will not sell, is confidential and conventionally sits at or under the low estimate. So a lot that clears its high estimate has done something specific. It found at least two bidders who disagreed with the specialist about the ceiling, and it kept finding them past the point the house had priced for. That is why this paper compares prices with high estimates rather than low ones, and why it prints the division rather than a verdict. Clearing a high estimate by 90% on one lot says the specialist mispriced that lot. Four of five lots clearing their highs, by 16.67% to 90.00%, says something closer to the house pricing a category conservatively, which is a different claim and a more useful one if you are deciding where to consign. The one lot in the five that sold below its high, at 93.75% of it, is the check on that reading: the pattern is not universal, and a D SI1 colourless stone of seven carats found its estimate roughly right.

One source carried this sale and that is stated in the article itself rather than left for a reader to infer. Rapaport's report was fetched direct with a browser user agent, 200, 466,304 bytes, and its own JSON-LD datePublished reads 2026-09-29T10:25:29+00:00, which was read off the page rather than assumed. Every figure printed here was grepped in that saved file before the sentence was written: 22.09 four hits, 475,000 three, 250,000 one, 7.6 million one, 425 lots one, 275,000 one, 212,500 one, 175,000 one, 112,500 one, 8.17 six, 22.06 three, 14.11 three, 7.02 eleven, 1.33 seventeen and September 28 one. The report attributes the prices to Heritage's own website. Heritage's site was then knocked directly to get a second carrier and refused: 403, 766 bytes, no content, so the figures above have one source and the article says so in its own closing line. The estimate leg does have a second, independent confirmation, and it is this paper's own: Carat Capital published $200,000 to $250,000 for the same 22.09-carat stone on 9 September, from Heritage's pre-sale listing, and the results report's $250,000 high matches it exactly. The Internally Flawless grade and the Kimberley documentation come from that earlier article and are attributed to it here, because the results report carries neither. All five percentages of the high estimate, the $17,882 average, the $1,250,000 and $840,000 aggregates and the 16.4% share are Carat Capital's arithmetic and are labelled wherever they appear. Archive check, read rather than counted: a search of the built site for Heritage and for Pragnell returns three files, every one opened. The nearest is 2026-09-09, which covers the PRE-SALE ESTIMATE for this exact stone ahead of the 28 September sale; this item covers the SALE RESULT of 28 September. Same stone, different period and a different event, so the verdict is ADVANCED, not repeat. The 30 July piece covers Heritage's first-half 2026 company results and is a different subject.

1. 22ct. Yellow Diamond Stars at Heritage Auction, Rapaport 29 September 2026