The second sheet

The deeper account

The index opened at 7,717.81, seventy-nine hundredths of a point under Friday's close, and that first print was the high of the day — the third session in a row to open at its extreme, after Thursday's open at the low and Friday's at the high. The first ten minutes took twenty-one points; ten o'clock printed 7,680. And then, for four and a half hours, almost nothing: from twenty to eleven until half past two every ten-minute close sits between 7,684 and 7,696, the afternoon high a 7,699.32 at ten past eleven, the sort of session the wrap services call quiet and narrow because it was. At twenty to three a bar printed 7,673.05 and broke the floor; the next hour held between 7,679 and 7,686; and the last ten minutes opened at 7,680.42, printed 7,666.99, the low of the day, and closed the session at 7,673.52, down 45.08. The last ten minutes were the day's low on wire reports out of the Gulf. The official version — Central Command saying it had destroyed five Iranian crude carriers after the Revolutionary Guard fired ballistic missiles at an American warship twice in two days, the warship untouched, the crews put off first — was posted at two minutes to five, an hour after the bell; the archive does not fix the minute the reports reached the tape, because the record does not. Around the index the day had the shape of the week before Friday's price report: crude for October settled 93.03, up 1.69 percent and the highest in six weeks, and Brent 97.92, after Houthi missiles and drones set fires at Saudi facilities in the south and wounded seventy-three; Canada's counter-tariffs — fifteen to fifty percent on some twenty billion dollars of steel, dairy, appliances, farm machinery, paper, wood and electronics — took effect just after midnight, dollar for dollar; the two-year note rose a basis point and a half to 4.394 and the ten-year one to 4.796; the price of a rise at next week's meeting stayed about sixty; gold fell 1.73 on the feed's fund; the VIX rose 8.19 percent to 15.72. Energy led the sectors at +1.11 and utilities followed; technology rose 0.32 because Intel, AMD and Broadcom outweighed Nvidia and Apple. Health care fell 2.52, a point worse than the next sector, and the Dow fell 1.18 percent, 628 points, twice the S&P and nearly four times the Nasdaq. Those two numbers have one cause.

At half past four on Friday afternoon, after the close and with a holiday weekend in front of it, Novartis reported that the largest trial yet of a drug for lipoprotein(a) had failed. Lipoprotein(a) is a particle in the blood, largely set by inheritance, that has been associated for forty years with heart attacks and strokes; the industry's bet, made by several companies at once, was that a drug that removed it would remove some of the events. Pelacarsen removed it. In 8,323 patients with high levels and established heart disease it lowered lipoprotein(a) and did not, against placebo, reduce the composite of cardiovascular death, heart attack, stroke and urgent procedure it was built to reduce. Novartis's American shares fell 13.93 percent on Tuesday, the first session that could trade on it, on ten million shares against Friday's one — the worst day in the feed's record of them. That was the company whose trial it was. Amgen's trial is a different one. Its drug, olpasiran, is a different kind of molecule that shuts the particle off a different way, and its outcomes study, OCEAN(a), has enrolled its patients and has not reported; nothing was announced by Amgen on Friday or Tuesday, and the only thing that changed about its drug over the weekend was the answer to the question underneath it. The industry calls what happened next a read-through, and means it plainly: one company's stock repriced on another company's data. Amgen opened down six and a half percent, was down ten by half past ten, and closed at 393.17, down 44.06, or 10.08 percent — its worst session since the twenty-seventh of October, 2000, which the archive checked against every session in the feed's history from September 1999, on four times Friday's volume. A bank downgraded it that morning and cited the year's thirty-four percent run, and the run was real: on Thursday, three sessions earlier, Amgen had closed at 444.12, its highest close of the past twelve months. And because the Dow is thirty share prices added up and divided by a number, a forty-four-dollar fall is the same number of points whichever company loses it: on the feed the thirty components' dollar changes sum to a little under 106, Amgen's 44 is two fifths of that, and by the divisor the published close implies its share of the 628 was about 262. Take it out and the oldest average fell about seven tenths of a percent, which is roughly what the S&P did. Elsewhere Intel rose 9.05 percent, its third straight gain and fourteen percent since Wednesday, on a trade-press report the company has not confirmed that it will raise PC-processor prices a tenth from October, and on an upgrade; Qualcomm opened up 6.91 on a custom-chip deal with Amazon and closed up 3.17; Corning rose 7.56 on a fiber contract; Howmet fell 10.70 for the second leg of the turbine-blade story; the airlines fell on the oil.

The page is a page from a duplicate book. The pad has a white sheet on top and a pink one underneath with carbon between them, and the instruction printed on the cover is the only one it needs: press firmly, you are writing on two sheets. On Friday evening Novartis pressed firmly. The words it wrote were about its own drug, in its own trial, on its own sheet, and they were true: the number the drug was built to lower came down, and the number that mattered did not. The sheet underneath was headed with a different name. Nobody wrote on it. Amgen has a different drug for the same particle in a trial that has not spoken, and on Tuesday morning the market held the pink sheet up to the light and read, in carbon, what had been pressed through from above — and charged it to that sheet at ten percent, the most in twenty-six years, on a company that had closed at a twelve-month high the Thursday before. Whether the carbon is right is the one thing nobody on Tuesday knew. The particle may be the wrong target, in which case the copy is a fair copy and the market did in one morning what the second trial will take years to confirm; or it may be a target that the first drug did not reach hard enough, or in the right patients, in which case the second sheet will one day carry its own entry in its own ink, and the carbon will have been a stain. The archive records the mechanics, not the verdict: one result, written once, on Friday, after the close; two companies repriced on Tuesday, the one that wrote it and the one it came through onto; and, at the bottom of the second sheet, the small figure that made the Dow's day twice as bad as the S&P's — about 262 points of 628 — written in nobody's hand.