How much the Nikkei 225 moves overnight: 18 hours, measured
Source: Hyperliquid public API / Last updated: / Perpetual-futures prices are reference values.
The cash session at the Tokyo Stock Exchange runs from 9:00 to 15:30. Out of the 24 hours in a day, the Nikkei 225 has only six "official" hours.
What happens to the price over the other 18 hours? A 24-hour perpetual-futures reference price makes that question measurable in figures.
This article is a measurement covering April to August 2026. It counts the 77 trading days, starting on a Monday through a Thursday, on which the overnight window is complete.
What was measured, and how
The data source and what it covers
The subject is the hourly candles of the Hyperliquid perpetual future xyz:JP225. They were retrieved on 25 August 2026.
This price is not the Nikkei 225 itself. It is a reference value of a contract that references it. The same distinction, and how the solid and dashed lines are drawn, is set out in what a perpetual future is.
The period holds 3,325 hourly candles, with 0 missing.
What was measured is a perpetual-futures reference value. No candles were missing over the period.
The two windows
The day was split into two windows. The labels correspond directly to the calculation.
| Window | Definition | Length |
|---|---|---|
| Cash-session window | Close of the hourly candle immediately before 09:00 JST → close of the hourly candle immediately before 15:30 JST | 6 hours |
| Overnight window | Close of the hourly candle immediately before 15:30 JST → close of the hourly candle immediately before 09:00 JST on the next trading day | 18 hours |
"The close of the hourly candle immediately before" is the close of the last hourly candle that completed before the target time. A Friday night runs into the weekend, so the overnight window counts only the 77 days that start on a Monday through a Thursday.
Public holidays are not excluded. Perpetual futures move on holidays as well, so the only exclusion rule is whether the window has its anchor candles.
The cash session is 6 hours and the overnight window is 18. Friday nights and weekends are covered in a separate article.
The result — the moves were larger overnight
The central figures
| Item | Cash-session window (Mon–Thu, N=77) | Overnight window (N=77) |
|---|---|---|
| Median absolute change | 0.68% | 1.06% |
| Days beyond ±0.5% | 64.94% (50/77) | 76.62% (59/77) |
| Days beyond ±1% | 38.96% (30/77) | 51.95% (40/77) |
| Largest rise | +2.89% (2026-06-11) | +3.83% (night of 2026-07-30) |
| Largest fall | -3.35% (2026-06-23) | -2.30% (night of 2026-06-10) |
The overnight absolute change exceeded the cash session on the same day on 45 of the 77 days, or 58.44%.
The cash session has large days too. The largest fall, -3.35%, was on the cash-session side. Even so, on the median and on the counts, the overnight window comes out higher.
"Night of 2026-07-30" in the table is the value of the overnight window that starts on 30 July. The window runs from the close immediately before 15:30 JST to the close immediately before 9:00 JST on the next trading day.
On the median, 0.68% for the cash session and 1.06% overnight. The overnight window was higher on 58.44% of days.
The night is, however, three times as long
The overnight window is 18 hours, three times the six hours of the cash-session window. The longer the window, the more change accumulates in it.
So this result is not "the night is wilder". The accurate reading is "the night is long, and the price keeps moving through it".
Read the figures with the threefold difference in window length in mind.
83% of the day's range was formed overnight
The other way to measure it is the range. Of the gap between the 24-hour high and the 24-hour low, the share made by the overnight candles alone was calculated.
The formula is (overnight high − overnight low) ÷ (24-hour high − 24-hour low) × 100. The median of that share was 83.28% and the mean was 80.41%.
Seen through the day's high-to-low range, most of it was formed while the Tokyo exchange was shut.
The overnight share of the day's range had a median of 83.28%.
The days with the largest overnight change
From the 77 measured days, the top five by overnight absolute change.
| Date (window start, JST) | Day | Overnight change | Cash-session change on the same day |
|---|---|---|---|
| 2026-07-30 | Thu | +3.83% | +0.63% |
| 2026-06-11 | Thu | +3.77% | +2.89% |
| 2026-06-24 | Wed | +2.67% | -0.25% |
| 2026-05-05 | Tue | +2.54% | +0.41% |
| 2026-08-04 | Tue | +2.47% | -0.32% |
Even on days when the cash session barely moved, there are several cases of more than 2% overnight. The opposite combination — the cash session moving while the night was quiet — also occurs.
Even on quiet cash-session days, several nights moved by more than 2%.
The monthly breakdown, and what to keep in mind
| Month | Days counted (Mon–Thu) | Cash-session median | Overnight median | Median overnight share of range |
|---|---|---|---|---|
| 2026-04 | 12 | 0.69% | 0.72% | 87.92% |
| 2026-05 | 16 | 0.96% | 1.10% | 83.68% |
| 2026-06 | 18 | 0.64% | 1.41% | 82.21% |
| 2026-07 | 18 | 0.73% | 1.28% | 69.51% |
| 2026-08 | 13 | 0.39% | 1.14% | 76.97% |
There are two things to keep in mind. First, trading activity in this symbol grew a great deal over the period. The median trade count in the cash-session window moves from 97 in April to 839.5 in August.
Second, the period is only about four and a half months. This is not a "tendency". It is a record that says "this is what the measurement showed over this period".
Trading was thin early on and the period is short. Read it as a record of the period, not a tendency.
Common questions
Q. Is this the movement of the Nikkei 225 itself?
A. No. It is the reference value of the Hyperliquid perpetual future xyz:JP225. It can differ from the official value of the Nikkei 225.
Q. Are public holidays and non-trading days excluded?
A. They are not. The only exclusion rule is whether the window has its anchor candles. Two days were excluded: the first and the last day of the period.
Q. Why does the period cover only four and a half months?
A. The source has a limit on how far back hourly candles can be retrieved. Coverage of this symbol begins on 9 April 2026. Only the period that could be retrieved is used, with no interpolation.
Q. Will the night keep moving more from here?
A. That is not written here. This article is a record of the measured period, not a forecast of future price moves.
Summary
- Across 77 Monday-to-Thursday days measured, the median absolute change was 0.68% in the cash session and 1.06% overnight.
- The overnight window exceeded the cash session on 58.44% of days.
- The overnight share of the day's range had a median of 83.28%.
- The overnight window is three times the length of the cash-session window. The figures include that difference in length.
- All of it is a record of April to August 2026, on Hyperliquid perpetual-futures reference values.
Measurement note: the subject is 3,325 hourly candles of xyz:JP225, with 0 missing. Retrieval ends at 12:00 UTC on 25 August 2026. The statistics were checked against an independent recalculation and matched.
The overnight moves can be seen as they are on the 24-hour tab and on the Nikkei 225 symbol page. For the weekend measurement, see how much world share prices move at the weekend.
This article explains a mechanism and a measurement. It does not recommend any particular trade.