What Sixty Per Cent Means

Four days after the Pollock sold for $181 million and three days after Bob Mnuchin's eleven pictures were dispatched in a white glove, Christie's returned to Rockefeller Center on the evening of Thursday, 22 May with the considerably less glamorous proposition of a forty-five-lot broadly mixed sale — the kind of evening that earns its own name not because any single lot defines it, but because the room, collectively, refused to stop buying things for less than the catalogue had suggested they were worth.

By the time the auctioneer took the gavel off the rostrum, twenty-seven of the forty-five lots had closed above their published high estimate. The total hammer was three hundred and ninety-one point nine million dollars, not one lot had been bought in, and the sell-through was, as it had been for every major evening sale that week, one hundred per cent. It was, by most measures, an unremarkable evening. It was, by one very specific measure, quite remarkable indeed.

The number that wants an explanation

Sixty per cent. Twenty-seven of forty-five lots above the high estimate. This is the number that sits in the back of a sale analyst's mind for weeks afterward, because it doesn't feel like a coincidence and it resists the easy explanations.

The case for coincidence is weak. Forty-five lots is not a small sample. If you model a sale where each lot has a twenty per cent chance of exceeding the published high — which is roughly the long-run empirical rate across Christie's and Sotheby's major evening sales from 2020 to 2025 — the probability of seeing twenty-seven or more over-highs is, to use the technical term, vanishingly small. The exact figure, for anyone keeping score, is less than one in a thousand. The market, in other words, did not happen to feel generous on a Thursday.

The week context

The Thursday sale did not arrive in a vacuum. It arrived at the end of a week in which Sotheby's had produced a hundred per cent sell-through for Mnuchin at fifty-five per cent over-high, and Christie's had produced a hundred per cent sell-through for Newhouse at — headline Pollock aside — a much more conservative over-high rate of twenty per cent. Three sales, all sold, a combined total north of a billion dollars, and rates of market-over-estimate that accelerated as the week went on rather than fading.

The Newhouse outlier, for anyone inclined to argue that a single Pollock distorts the picture, is worth setting aside entirely and re-running the numbers. Strip the Pollock, and the Newhouse over-high rate falls to three out of fourteen lots — twenty-one per cent, which is to say exactly normal. The Pollock was a one-night event in its own category; the behaviour of the other thirteen lots was textbook. What the Thursday evening sale produced was not the textbook.

What sold and for how much

The top lot was a Basquiat — Warrior, from a prominent private collection, estimated at sixty-to-eighty million dollars — which hammered at $85.4 million, or roughly six and three-quarters per cent above the high. This is, notably, a comparatively modest result: a Basquiat that performs six per cent over high on an eighty-million-dollar estimate is a Basquiat the room wanted but did not fight over. The evening's real excitement — if excitement is the right word for something that can be expressed as a percentage — was in the bottom third of the card.

A Keith Haring — a mid-period work on canvas estimated at three-to-five million — hammered at $7.4 million, forty-eight per cent above the high. Across the card, lots estimated below ten million outperformed their high estimates at a rate of approximately seventy per cent; lots estimated above twenty million outperformed at a rate closer to forty per cent. The over-high phenomenon, to put it plainly, was not distributed evenly. It was concentrated at the lower-cap end of the card, which is exactly where estimate-setting is hardest and where the houses' models are most likely to have drifted from current market appetite.

The two readings

There are two standard explanations for a sixty-per-cent over-high rate, and they are not mutually exclusive.

The first is that the market for twentieth-century art has moved north of where it was when the estimates were set, which was roughly three months before the sale. If the bid level for a five-million-dollar Haring has risen by forty or fifty per cent in a quarter, then the estimate range is simply wrong — it reflects a market that no longer exists. This reading is comfortable for the houses (they cannot be expected to re-estimate on the week of the sale) and uncomfortable for anyone trying to model where the market actually is.

The second reading is that the houses set conservative estimates deliberately, because conservative estimates generate urgency and urgency generates competing bids and competing bids generate exactly the kind of above-high results that fill the catalogue for next May's equivalent sale. This reading is uncomfortable for the houses (it implies the catalogue is a managed document) and entirely consistent with how auctions have worked for the last forty years.

Pick your reading. Probably both are true. The interesting thing, the thing that will be studied by people who model this market for a living, is that sixty per cent over-high on forty-five lots in a single evening is not something that happens when both forces are absent. Something moved. Whether it was the market or the estimates is, in the end, the same question differently worded.

The three under-lows

Three lots closed beneath the published low estimate. One was a Lichtenstein print — the smallest piece on the card by estimated value, priced at $1.2–1.8 million, which hammered at $0.9 million. The other two were works on paper by secondary market names at the lower end of the card. None of the three were in the top fifteen lots by estimate. None of them were, in any obvious sense, the room failing; they were the room being slightly less interested in three specific works than the catalogue had hoped.

In a normal sale, three under-lows out of forty-five lots would be unremarkable. In a sale where twenty-seven lots went over high, three under-lows are almost conspicuous by their scarcity — evidence, if you want it, that the evening's appetite was genuine rather than auction-room theatre.

The math, briefly

Forty-five lots sold, $391.9 million, sell-through one hundred per cent. Over-high rate sixty per cent — or, if you want the week in aggregate: Mnuchin at fifty-five, the Newhouse ex-Pollock at twenty-one, and the Thursday evening at sixty. Three sales. Mean over-high rate across the week, weighted by lot count: forty-eight per cent. Long-run expected value: twenty per cent. The week of 14–22 May 2026, taken as a whole, produced a market that ran approximately two and a half times the historical rate of above-estimate results. That is not a coincidence. That is a signal.

What it means

The Thursday Christie's sale will not be remembered the way the Monday Newhouse sale will be remembered — there is no single lot to put in the press release, no one number that rewrites the record books, no canonical photograph of the auctioneer's expression in the moment the Pollock bid touched one hundred and eighty million. What it will be remembered for, by the people whose job is to remember these things, is the sixty per cent. The week that produced three consecutive hundred-per-cent sell-throughs, a Pollock at 3.6 times the high, and then, quietly, a forty-five-lot evening sale in which sixty per cent of the card went above the number the catalogue had assigned it.

The market spoke three times in one week. The Thursday sale was the third time, and it was not the loudest, but it may have been the most consistent.