Futures Candle Size: Are ES & NQ 1-Minute Candles Expanding or Contracting?
The average 1-minute candle on ES and NQ futures, measured as its high-to-low range in percent of price, averaged per session and published for the cash session and the overnight separately. The 20-session average is banded against every prior reading, from quiet to extreme, and each band carries what the index did next since 2020. It is realized volatility sampled every minute, so a regime change shows in days rather than weeks.
Today's reading
As of the September 10, 2026 close, the average NQ 1-minute candle over the last 20 cash sessions spanned 0.048% of price, the 11th percentile of every prior 20-session reading: the quiet band, contracting, for 10 sessions. The overnight candle averaged 0.028%. ES sits in the quiet band at 0.030%. Since 2020, sessions in the quiet band averaged +1.12% over the next 21 sessions against a +1.26% baseline (519 sessions).
Sources, methodology & freshnessLast updated 2026-09-10 · Open ↓Close ↑
Since 2020, quiet-band sessions on NQ averaged +1.12% over the next 21 sessions (66% positive) against a +1.26% baseline, 519 sessions; candle size twenty sessions later averaged 1.14× today's. That is context rather than a forecast.
Cash-session candle size — NQ
Mean 1-minute range per regular session, in percent of price, with the 5- and 20-session averages. The 20-session line is coloured by the band it sat in that day, ranked against every prior reading; the shaded zones mark today's quiet, elevated and extreme thresholds.
Cash session vs overnight — NQ
The two halves of the trading day, each as a 20-session average. The overnight is every candle outside 8:30–3:00 Central and only counts minutes that traded, so its level carries overnight liquidity as well as volatility.
Red vs green candles, and NQ vs ES
Above 1× the average red candle is bigger than the average green one.
Rising means the regime is tech-led; falling toward 1× means the two indices are moving together.
What followed each band — NQ
Average forward futures returns from every session in each band since 2020, beside the all-sessions baseline, and how much of the regime was still there twenty sessions later. Today's band, quiet, ran below the baseline over 21 sessions.
| Regime | % of days | Next 5 sessions | Next 10 sessions | Next 21 sessions | Size in 20s | N |
|---|---|---|---|---|---|---|
| Quiet (below 25th pct)now | 37% | +0.32% · 61%↑ | +0.53% · 61%↑ | +1.12% · 66%↑ | 1.14× | 519 |
| Normal (25th–75th) | 48% | +0.14% · 55%↑ | +0.41% · 56%↑ | +0.72% · 58%↑ | 0.99× | 666 |
| Elevated (75th–90th) | 10% | +0.17% · 53%↑ | -0.03% · 52%↑ | +1.50% · 63%↑ | 0.97× | 145 |
| Extreme (90th and above) | 5% | +0.73% · 57%↑ | +1.49% · 59%↑ | +1.51% · 47%↑ | 0.76× | 70 |
| All sessions (baseline) | 100% | +0.30% · 58%↑ | +0.59% · 61%↑ | +1.26% · 64%↑ | 1.00× | 1,669 |
| Regime | % of days | Next 5 sessions | Next 10 sessions | Next 21 sessions | Size in 20s | N |
|---|---|---|---|---|---|---|
| Expanding (5s above 20s) | 44% | +0.42% · 60%↑ | +0.66% · 60%↑ | +1.08% · 63%↑ | 1.10× | 621 |
| Contracting (5s below 20s)now | 56% | +0.09% · 55%↑ | +0.30% · 56%↑ | +0.92% · 60%↑ | 0.98× | 779 |
| All sessions (baseline) | 100% | +0.30% · 58%↑ | +0.59% · 61%↑ | +1.26% · 64%↑ | 1.00× | 1,669 |
Forward returns on the continuous futures close at the cash close. Consecutive sessions share bands and forward windows overlap, so treat the averages as descriptive tendencies. The percentile is expanding and point-in-time: each session is ranked only against sessions before it, so the first 250 sessions of the record carry no band. Today's direction, contracting, averaged +0.92% over 21 sessions.
Since 2020, quiet-band sessions on NQ averaged +1.12% over the next 21 sessions (66% positive) against a +1.26% baseline, 519 sessions; candle size twenty sessions later averaged 1.14× today's. That is context rather than a forecast.
How Futures Candle Size Works
- 1Measure every 1-minute candle in percentFor each 1-minute bar of ES and NQ, 23 hours a day, we take the range (high minus low) and divide it by the prior session's front-month close. Percent rather than points: NQ has traded from under 4,000 to near 30,000, and a 10-point candle means something different at each level.
- 2Use the real price as the denominatorThe continuous contract we take the candles from is back-adjusted, so its historical level is fake: the accumulated roll gaps put March 2009 near 4,650 when NQ actually traded near 1,240. Point moves are taken from the adjusted series, which has no roll seams; the level comes from the un-adjusted front month.
- 3Split the cash session from the overnightRegular trading hours (8:30–3:00 Central) are one series; everything else in the CME trading day is the other. TradeStation only reports minutes that traded, so the overnight candle count is a liquidity measure as well as a volatility one — the two halves are published separately rather than blended.
- 4Average by session, then band itThe daily reading is the mean candle range for that session. The 5- and 20-session averages smooth it. The 20-session average is ranked against every prior 20-session value, an expanding point-in-time percentile, and banded: quiet below the 25th percentile, normal to the 75th, elevated to the 90th, extreme above. Direction is the 5-session average against the 20.
- 5Attach what happened nextFor every session in each band, and for expanding versus contracting, we report the futures' average return over the next 5, 10 and 21 sessions beside the unconditional baseline, plus how much of the regime was still there twenty sessions later.