Up or down - which way is the travel?

Splitting the swing

A swing is distance without direction. This page takes the same half hour and splits it: the furthest the Nasdaq 100 reached above the release price at any point in the thirty minutes, and the furthest it reached below, each a median across the record. The ratio is up divided by down. A ratio near one is a release whose half hours have been symmetric on average; a ratio well above or below it is a release whose travel has leaned one way, whatever the close did.

Which releases lean, and which way

Payrolls leans hardest upward on the current record, 1.46 times as much travel above the release price as below it, and CPI leans the same way at 1.30 (a median best point of 0.31% against a median worst of 0.24% for CPI). PPI leans hardest the other way, 0.51. FOMC leans down too, 0.57, with a worst point of 0.22% against a best of 0.13%, and so do JOLTS and retail sales. GDP and Core PCE are close to even.

This is the page most tied to the period it was measured in. The record runs from 2023, through a stretch in which inflation prints were more often at or below forecast than above it, and a soft inflation print is the kind of half hour that runs up. A different stretch of the cycle, with the same measurement, would lean the other way. So the reading is not that CPI is an upward release. It is that the last 37 CPI half hours have leaned up, and the number is there so you can see how much.

What it is useful for

It tells me which side of the release price the overshoot tends to be on, which is the side where a stop gets hunted. On FOMC the worst point is deeper than the best point, so a long that survives the first thirty minutes has usually had to survive a deeper dip than a short would have. Paired with the closing counts on the ranking page, it also separates two things that look alike: a release that closes higher because it went straight up, and one that closes higher after a deep round trip. The second kind costs more to hold, and only this page shows it.

What it cannot tell you

Direction. A release can travel further upward on the median and still close lower more often than not, which is why the close count lives on another page. Best and worst are medians of extremes, which is a coarse thing to take a median of, and the samples are the same few dozen as everywhere else. Most of all, the lean is a property of the last three years of prints and not of the release, and I would expect this page to look different before the ranking page does.