Every scheduled US release this site measures, ranked by how far the Nasdaq 100 actually travelled in the thirty minutes after the print. The figure is a median across the whole record, not a headline picked from one day.
| # | release | releases measured | median swing | biggest | up / down |
|---|---|---|---|---|---|
| 1 | CPI | 36 | 0.69% | 2.27% | 26 / 10 |
| 2 | Payrolls | 34 | 0.69% | 1.58% | 21 / 13 |
| 3 | FOMC | 24 | 0.51% | 1.36% | 10 / 14 |
| 4 | JOLTS | 37 | 0.48% | 1.53% | 16 / 21 |
| 5 | PPI | 31 | 0.44% | 1.02% | 15 / 16 |
| 6 | Core PCE | 26 | 0.39% | 0.93% | 16 / 10 |
| 7 | GDP | 24 | 0.27% | 0.84% | 13 / 11 |
| 8 | Retail sales | 29 | 0.26% | 1.18% | 17 / 12 |
CPI sits at the top with a median of 0.69% and retail sales at the bottom with 0.26% - a factor of 2.6 between the loudest release on the calendar and the quietest. Releases that printed in the same minute as another report here are excluded, since the move belongs to both and to neither.
Each row is one scheduled US release, and the number beside it is the median of every half hour this site has measured after that release: the highest price minus the lowest price in the thirty minutes after the print, as a percentage of the Nasdaq 100 level at the moment of release. The ranking is by that median, not by the biggest day, because a ranking by biggest day would be a ranking of accidents. One violent afternoon in 2024 would decide it.
The biggest column is there so the median has something to sit against. Retail sales is the clearest case: a median of 0.26% and a widest day of 1.18%, which is the shape of a release that is background noise most months and then, once in a while, is not. GDP sits at the bottom with 0.27%, and it only sits there because the days it shares a minute with Core PCE are left out. Counted together, GDP looked like a mid-table release. None of that was GDP.
The top of the table is one class of event and the bottom is another, and I stopped treating the eight of them as points on a line. The jobs report and CPI are within a hair of each other on the median, at 0.69% and 0.69%, and on a normal month either one is worth more half-hour travel than the two quietest releases put together. FOMC sits third at 0.51%, and it is the one row where the half hour is the wrong window entirely: the press conference starts thirty minutes after the statement, and the two-hour figure on the persistence page is more than double this one.
The up / down column is not travel. It is a count of where the index closed the half hour relative to the release price, and it is here because a release can be loud and still lean one way. CPI finished higher 26 times out of 36 on the current record; FOMC finished lower more often than higher. Neither count is a forecast of the next one, and the page on forecast surprises splits them further by what the figure printed.
I used to size every release position the same way and adjust by feel. Now the row decides. A stop that sits inside the median swing of the release I am holding through is a stop that an ordinary day will hit, not an unusual one, so the median is the smallest distance I am willing to give a position across the print, and the biggest column is the distance I have to be able to survive. For the two quietest releases that is a small number and I mostly ignore them. For the top three it is the whole trade.
It does not tell you direction, and it does not tell you that the next print will be near the median. It is measured on an index, not on the futures contract, so the percentages read across to NQ and ES but the points do not. The samples are between 24 and 37 releases per row, which is enough to rank them and not enough to trust the second decimal. Releases that printed in the same minute as another report on this site are excluded from every row, and the page on shared minutes shows why.
Releases: Core PCE Price Index Consumer Price Index FOMC Rate Decision GDP JOLTS Job Openings Nonfarm Payrolls Producer Price Index Retail Sales