LiquidityLevelsLondon takes out the Asia range 9 days out of 10. It tells you very little about the US session.
A study of 458 overnight-to-cash sessions in ES and NQ futures (2024-05-15 to 2026-10-06): how often the London session breaks the Asia high or low, how often the US session follows, and what the numbers do and do not support.
Descriptive statistics on historical data. Not trading advice.
Key findings
- The Asia range rarely holds. London traded beyond at least one side of the Asia range on 93.4% of ES sessions and 91.7% of NQ sessions. Both sides were taken on 22.3% (ES) and 19.4% (NQ) of days. Only 6.6% of ES sessions stayed fully inside.
- Upside breaks were more common than downside breaks. London traded above the Asia high on 62.9% of ES days versus 52.8% below the Asia low, in a sample that includes a strong uptrend.
- Direction of the US session was close to a coin flip in every condition. The US session closed above its open on 50.7% of all ES days. Conditional on what London did, the figures ranged from 45.7% to 56.7%, and every 95% confidence interval includes 50%.
- The Asia range is small. The median Asia range was 18.25 points in ES against a prior regular-session range of 56.75 points, so ordinary London trading overlaps it by construction.
The practical reading: the Asia high and low are useful reference points for describing the overnight session, but a break of them is the normal case, not a signal. That is consistent with how the Academy treats the idea: what matters is whether price is accepted or rejected beyond a level, not whether a level was touched.
How often London breaks the Asia range
| Share of sessions (N = 458) | ES | NQ |
|---|---|---|
| London above the Asia high | 62.9%288 of 458 · 95% CI 58.4–67.2% | 60.3%276 of 458 · 95% CI 55.7–64.6% |
| London below the Asia low | 52.8%242 of 458 · 95% CI 48.3–57.4% | 50.9%233 of 458 · 95% CI 46.3–55.4% |
| At least one side taken | 93.4%428 of 458 · 95% CI 90.8–95.4% | 91.7%420 of 458 · 95% CI 88.8–93.9% |
| Both sides taken | 22.3%102 of 458 · 95% CI 18.7–26.3% | 19.4%89 of 458 · 95% CI 16.1–23.3% |
| Neither side taken | 6.6%30 of 458 · 95% CI 4.6–9.2% | 8.3%38 of 458 · 95% CI 6.1–11.2% |
By weekday, the share of sessions where London took out at least one side of the Asia range was similar every day (Monday is excluded because its Asia window starts after the weekend gap):
| Weekday | ES | NQ |
|---|---|---|
| Tue | 93.8% (105/112) | 91.1% (102/112) |
| Wed | 93.3% (112/120) | 92.5% (111/120) |
| Thu | 92.2% (106/115) | 92.2% (106/115) |
| Fri | 94.4% (102/108) | 90.7% (98/108) |
What happened in the US session
For sessions where London traded above the Asia high, we split by how London finished: back inside the range at 8:00 AM ET (a sweep that returned) or still above the Asia high (acceptance beyond it). We did the same for the downside. The table shows how often the US session (9:00 AM to 4:00 PM ET hourly bars) closed above its open.
| US session closed above its open | ES | NQ |
|---|---|---|
| All sessions | 50.7%232 of 458 · 95% CI 46.1–55.2% | 52.8%242 of 458 · 95% CI 48.3–57.4% |
| After a London sweep above the Asia high that returned inside | 56.7%76 of 134 · 95% CI 48.3–64.8% | 53.2%66 of 124 · 95% CI 44.5–61.8% |
| After London accepted above the Asia high | 48.7%75 of 154 · 95% CI 40.9–56.5% | 51.3%78 of 152 · 95% CI 43.4–59.1% |
| After a London sweep below the Asia low that returned inside | 52.7%78 of 148 · 95% CI 44.7–60.6% | 49.0%70 of 143 · 95% CI 40.9–57.1% |
| After London accepted below the Asia low | 45.7%43 of 94 · 95% CI 36–55.8% | 58.9%53 of 90 · 95% CI 48.6–68.5% |
The largest gap from the all-days figure is under ten percentage points, and each subgroup has between 90 and 154 observations. With samples that size the 95% intervals are roughly 16 to 20 points wide and all contain 50%. Nothing here supports treating a London sweep, in either direction, as a directional signal for the US session. An effect could exist and be too small for this sample to detect, and the sample covers one market regime.
Overnight versus the prior day range
The overnight session (Asia plus London) traded above the prior regular-session high on 35.4% of ES sessions and 39.7% of NQ sessions, and below the prior regular-session low on 27.5% and 27.3%. Overnight breaks of the prior regular-session high or low are common, which is why overnight levels, the prior-day high and low and the overnight high and low are tracked together in LiquidityLevels briefings.
Methodology and limitations
- Data: hourly bars for the continuous front-month ES and NQ futures contracts (Yahoo Finance symbols ES=F and NQ=F), converted to America/New_York time. Sessions run 2024-05-15 to 2026-10-06. Days with incomplete data were dropped.
- Sessions: Asia = hourly bars starting 8:00 PM through 1:00 AM ET (prior evening into the morning). London = bars starting 2:00 AM through 7:00 AM ET. US = bars starting 9:00 AM through 3:00 PM ET. Prior regular session uses the same 9:00 AM to 3:00 PM bars of the previous trading day.
- Days: Tuesday through Friday sessions (N = 458). Monday sessions are excluded because their Asia window begins after the weekend.
- Definitions: a London break means a London bar high above the Asia high (or a low below the Asia low). A returned sweep means London broke the level but its last bar closed back inside the Asia range; acceptance means it closed beyond it. US direction is the close of the last US bar versus the open of the first.
- Confidence intervals: 95% Wilson score intervals for proportions.
- Limitations: hourly bars hide intrabar order (a bar can contain both a break and a return); the 9:00 AM bar includes the half hour before the cash open; continuous futures series contain contract-roll gaps on a few days per year; the sample is one market regime; Yahoo Finance data is unaudited; costs, slippage and execution are ignored. This is not a backtest of a strategy.
- Reproduce it: the script that generated every number here is overnight-session-study.py (Python 3, standard library only; pass it the two Yahoo hourly JSON files).
Cite this study
You are welcome to quote and cite these figures. Suggested citation:
LiquidityLevels (2026). ES and NQ Overnight Session Study, 2024-05-15 to 2026-10-06. https://liquiditylevels.com/research/overnight-session-study/