This paper examines whether Opening Range Breakouts in the E-mini NASDAQ-100 futures market exhibit statistically meaningful follow-through properties. Using full market-by-order (MBO) tick data across 159 trading sessions (Sep 2025 – Apr 2026), we construct an intraday structural framework based on Volume Profile Value Areas, session-open order-flow delta, and a normalized deviation metric. Our central finding is that breakouts carry a strong base probability of follow-through (71.1% reaching 0.5× the filter range), with opening order-flow direction providing a consistent predictive signal across all deviation targets. The clearest actionable result is a warning pattern: when a breakout reverses to the VA midpoint without first reaching the 0.5× deviation target, continuation probability drops from 71.1% to just 22.7% — a 48.4pp gap. A notable new finding emerges with the expanded dataset: DOWN breakouts reach the 2.0× deviation target at 37.8%, nearly triple the rate of UP breakouts (14.1%), revealing a structural directional asymmetry at deep targets.
Introduction
The Opening Range Breakout (ORB) is one of the oldest intraday frameworks in active use. The premise is straightforward: markets establish a reference range during the early session, and when price escapes that range, it tends to follow through in the breakout direction for a meaningful distance. This study asks whether that premise holds empirically in modern NQ futures — and which observable conditions strengthen or weaken it.
Rather than defining the opening range as a simple high/low window, this study uses the Volume Profile Value Area — the price range containing 70% of opening volume. This provides a structurally grounded reference zone that reflects where genuine two-sided price acceptance occurred during the opening session.
Primary research questions:
- What proportion of breakouts follow through to normalized distance targets?
- Does opening order-flow delta (aggressive buy vs. sell imbalance) predict continuation?
- Does the structural context at 10:05 ET (Filter Case) influence follow-through depth?
- Does breakout candle volume or breakout candle delta add predictive signal?
- How does a second-episode breakout behave after an initial failed attempt?
Model & Methodology
2.1 Data
All data is sourced from Databento's CME Globex MBO (Market-By-Order) feed,
dataset GLBX.MDP3, instrument NQ.c.0
(E-mini NASDAQ-100, continuous front-month contract).
Tick-level trade records include price, size, and aggressor side as classified by
the exchange: side = B is an aggressive buy
(order lifting the ask); side = A is an aggressive sell
(order hitting the bid). All session times are Eastern Time.
Regular Trading Hours (RTH): 9:30 AM – 4:00 PM ET.
2.2 Value Area & Opening Window
The opening window spans 9:30–10:05 ET (first 35 minutes of RTH). All trades in this window construct a Volume Profile — a histogram of traded volume at each 0.25-point (one tick) price level. From this profile the Value Area is computed: the price range containing 70% of opening volume, expanded outward from the Point of Control (POC) using the standard single-tick expansion algorithm.
2.3 The 10:00 Candle & Filter Cases
The 5-minute candle opening at 10:00 ET (closing at 10:05) — the final bar of the opening window — determines the Filter Case. Where price closed relative to the Value Area sets the exact Filter Range used for all subsequent breakout detection. Three cases are defined:
(VAL ≤ close ≤ VAH) Filter HighVAH Filter LowVAL Interpretation Price accepted within opening value. Clean structural boundaries.
The Filter Mid = midpoint of the Filter Range (not VA_Mid when Cases B or C apply). It serves as the episode-reset level in the two-episode framework (§2.6).
2.4 Breakout Detection
Starting at 10:05 ET, each 5-minute bar is scanned for a closing break outside the Filter Range. The first such close is the breakout:
- Upside breakout: first 5-min close above
filter_high - Downside breakout: first 5-min close below
filter_low
Of 78 sessions, 77 (98.7%) produced a breakout before end of RTH, confirming that the market resolves the opening range on virtually every session.
2.5 Deviation Hit Rate (Sustainability Metric)
For each breakout, the maximum favorable excursion (MFE) is measured — the largest move in the breakout direction from the breakout price, using the high/low of all subsequent 5-minute bars through end of RTH.
MFE is normalized by the filter range size (filter_high − filter_low), creating a dimensionless deviation multiple:
A Deviation Hit at level σ means the market traveled at least σ × (filter range) beyond the breakout price. Four targets are tracked: 0.5×, 1.0×, 1.5×, 2.0×. Normalization ensures that a 50-point move on a 100-point filter-range day is treated equivalently to a 100-point move on a 200-point filter-range day — enabling fair comparison across different market volatility conditions.
2.6 Opening Delta%
Opening Delta% measures net directional order-flow imbalance during 9:30–10:05:
A positive Δ% indicates buyers were more aggressive; negative indicates sellers dominated. Days are classified as Positive or Negative based on the sign of Δ% (76 positive / 83 negative across all 159 days).
2.7 Two-Episode Framework
After Episode 1 (initial breakout), the session is monitored for a Filter Mid cross — a 5-minute close back through the Filter Mid in the direction opposite to the breakout. This signals the end of Episode 1.
If a cross occurs, the remaining session is scanned for
Episode 2 (bo2) — a fresh breakout through either filter boundary.
A key classification is whether Episode 1 reached the 0.5× deviation target
before the cross (bo1_hit_dev_before_cross).
As shown in §3.8, this is the strongest single predictor of follow-through identified.
(bo1)
Cross?
n=80 · 85.0% @ 0.5×
Before Cross?
n=35 · 60.0% @ 1.0×
n=44 · 22.7% @ 0.5×
Results
3.1 Sample Overview
The near-universal breakout rate (159 of 161 days) confirms that the ORB model generates an active signal on essentially every session. Breakout direction was moderately skewed: 85 upside vs. 74 downside over this period. Case A — price inside the Value Area at 10:05 — was the most common structural context (50.9% of days).
3.2 Overall Deviation Hit Rates
The chart below shows what percentage of breakout days reached each deviation target. All 159 breakout days are included.
The 71.1% hit rate at 0.5× establishes a strong baseline: after a Filter Range breakout, nearly three in four sessions extend at least half the filter range beyond the breakout level. The rate decays to 51.6% at 1×, 36.5% at 1.5×, and 25.2% at 2×. This geometric decay is consistent with a price process where each successive extension becomes progressively less probable.
3.3 Breakout Direction
Upside and downside breakouts show broadly similar follow-through at 0.5× and 1.0×. A notable divergence appears at the 2.0× level: downside breakouts reach 2× at 28.2% vs. only 15.8% for upside. This asymmetry is consistent with the historically sharper character of equity index selloffs relative to rallies — fear-driven moves tend to extend further and faster than greed-driven moves.
3.4 Opening Delta Analysis
Opening Delta% classifies whether buyers or sellers dominated the 9:30–10:05 window. The chart below shows how this classification relates to subsequent breakout follow-through across all four deviation targets.
Positive delta days outperform at every level: 75.0% vs. 67.5% at 0.5× (a 7.5 percentage-point gap), 52.6% vs. 50.6% at 1.0×, 36.8% vs. 36.1% at 1.5×, and 23.7% vs. 26.5% at 2.0×. The gap is modest and narrows with depth, suggesting opening order-flow direction has a mild but not dominant influence on continuation.
The strongest combination is Down breakout + Positive opening delta (83.3%). This configuration may reflect sessions where early-session buyers were overrun by larger institutional selling pressure — a capitulation dynamic that tends to accelerate. All four combinations are shown with sample sizes; note that smaller sub-groups (particularly DOWN + Positive, n=24) warrant further observation.
3.5 Filter Case Analysis
At 0.5× deviation, all three cases perform broadly similarly (63–81%). The divergence widens sharply at deeper targets: at 1.5×, Case A reaches 54.1% while Case C reaches only 10.5%. This is the most structurally informative dimension for setting follow-through expectations.
The intuition is straightforward. In Case A, price was inside the Value Area at 10:05 — the filter boundaries (VAH and VAL) represent genuine price acceptance zones. A close through these levels is a clean structural break. In Cases B and C, the filter was already probed or extended beyond VA during the opening window, creating noisier boundary conditions and greater probability of reversal at deeper extension targets.
Average filter range sizes differ substantially by case — a critical factor:
| Case | Condition | Avg Filter Range | n (days) | 1.5× Target |
|---|---|---|---|---|
| A | Close inside VA | 97.8 pts | 38 | 54.1% |
| B | Close above VAH | 140.6 pts | 21 | 23.8% |
| C | Close below VAL | 186.2 pts | 19 | 10.5% |
Reaching 1.5× of a 186-point range means traveling 279 additional points beyond the breakout — a significantly higher bar than 1.5× of a 98-point range (147 points). The normalized metric is fair, but the absolute implied move still scales with filter range size.
3.6 Breakout Candle Volume
The conventional intuition — high-volume breakout candles signal conviction — is not supported by the data. High-volume breakout candles show the worst follow-through at every target level: 65.4% at 0.5× vs. 80.8% for low-volume candles. The inverse relationship is consistent across all four deviation levels.
A possible explanation: extremely high-volume breakout candles may represent stop-run dynamics or news-driven spikes where liquidity from both sides is consumed simultaneously, followed by a rapid mean reversion. Low-volume breakouts may represent more orderly, one-sided expansion with less exhaustion. This finding warrants investigation with a larger sample.
3.7 Breakout Candle Delta%
The Delta% of the breakout candle itself — aggressive buy minus sell volume, normalized — provides a directional conviction measure for the specific bar that triggered the breakout. For downside breakouts we expect negative delta; for upside, positive.
For downside breakouts, the signal is clear and monotonic: the more negative the breakout candle delta, the higher the follow-through probability. Days with delta below −10% show a 100% hit rate at 0.5× (n=6) and 83.3% at 1.0×. For days where sellers hit the bid aggressively on the breakout candle, continuation is almost certain at the minimum deviation target.
For upside breakouts, the relationship is weaker. The 0–5% delta range (n=14) achieves the highest rate (78.6%), while the 10–20% range (n=12) achieves 66.7% — a modest inverse rather than the expected positive correlation. This may reflect that upside NQ breakouts encompass a more heterogeneous mix of trend continuation and short-squeeze dynamics. Small sub-group sizes for the extreme ranges limit confidence.
3.8 Two-Episode Framework
The Two-Episode Framework produces the largest single signal gap in this study. Three behaviorally distinct groups emerge:
Episode 2 (bo2) breakouts were found in 68 sessions (42.8% of breakout days). Notably, 47 of 68 bo2 events (69.1%) moved in the opposite direction from bo1, confirming that the two-episode framework captures genuine intraday directional reversals. bo2 achieved 76.5% at 0.5× — exceeding the overall base rate, suggesting that second-episode breakouts carry at least equivalent structural validity as first-episode breakouts.
Key Findings & Edge Assessment
Each finding below is assessed against the baseline (71.1% at 0.5×) to determine whether it represents an actionable edge, a neutral observation, or a counter-intuitive result requiring further investigation.
Strong Base Rate — Breakouts Follow Through
71.1% of all breakout days reach 0.5× deviation across 159 sessions. The market routinely commits after an opening range break. The base rate remained stable as the dataset doubled from 78 to 159 days (72.7% → 71.1%), confirming this is not a small-sample artifact. Systematic fade strategies are penalized by this probability.
✓ Confirmed Structural EdgeOpening Delta Predicts Follow-Through Consistently
Positive delta days outperform Negative by 7.5pp at 0.5× (75.0% vs. 67.5%). The gap narrows with more data but persists across all four deviation levels. The strongest single directional combination is Down breakout + Positive delta (83.3% at 0.5×, n=24). Opening delta is a consistent directional tilt but not a dominant predictor at this sample size.
~ Mild Edge — Gap Narrows With Larger SampleFilter Case Predicts Depth of Follow-Through
Case A is substantially better for deeper targets: 47.6% at 1.5× vs. Case C's 17.6%. At 0.5×, cases span from 61.8% (C) to 79.1% (B). Case B is best at 0.5× but drops sharply at depth. The edge is specific to predicting how far a breakout extends, not whether one occurs. Case C days are unlikely to reach even 1.0× deviation (29.4%).
~ Conditional Edge — Deeper Targets OnlyHigher Breakout Candle Volume Does Not Improve Follow-Through
With 159 days, the volume effect largely disappears: Low=69.8%, Mid=75.5%, High=67.9% at 0.5×. The initial study's apparent volume signal (80.8% vs. 65.4%) was likely a small-sample artifact. Volume does not predict follow-through in either direction. The "high volume = conviction" heuristic is not supported by the expanded data.
✗ No Edge — Volume Signal Fades With Larger SampleDownside Breakout Candle Delta Is a Useful Filter
With 159 days, DOWN breakout candle delta shows a flat pattern across all delta ranges (72.7%–74.5% at 0.5×). The initial "near-100%" result for strongly negative delta was a small-sample artifact (n=6 in the 78-day study). For UP breakouts, the 0–10% range (n=59) hits 71.2%, while higher delta slightly underperforms. No monotonic relationship found in either direction with the larger sample.
✗ Signal Faded — Initial n=6 Bucket Was NoiseEpisode Warning Signal: The "Cross + Miss" Pattern
When a breakout reverses through the Filter Mid without first reaching 0.5× deviation, continuation probability drops from 71.1% to 22.7% — a 48.4pp gap. This gap strengthened from 42.7pp in the initial 78-day study, confirming it is not noise. With 44 occurrences (n=44), this is now statistically reliable. The original breakout thesis is structurally invalidated when this pattern triggers.
✓ Strongest Signal — 48.4pp Below Baseline, Confirmed With 159 DaysEpisode 2 Breakouts Are a Real, Productive Phenomenon
68 of 159 sessions (42.8%) produced a second breakout after the Filter Mid cross. 69.1% of these moved opposite to bo1. bo2 achieved 76.5% at 0.5× — exceeding the overall base rate. The two-episode framework captures genuine intraday reversals, and bo2 now performs above baseline, strengthening the case for treating it as a distinct tradeable setup.
✓ Confirmed and Upgraded — bo2 Beats Base Rate With 68 OccurrencesNEW: Directional Asymmetry at Deep Deviation Targets
DOWN breakouts reach the 2.0× target at 37.8% (n=74), while UP breakouts reach it at only 14.1% (n=85) — a 23.7pp gap. The asymmetry is minimal at 0.5× (74.3% DOWN vs. 68.2% UP) but amplifies at depth. At 1.5×: DOWN=41.9% vs. UP=31.8%. At 2.0×: DOWN=37.8% vs. UP=14.1%. This was not visible in the 78-day dataset. With 159 days, the pattern is consistent enough to warrant directional target calibration: DOWN breakout days should be held for deeper targets than UP breakout days.
✓ New Finding — 23.7pp Asymmetry at 2.0×, Requires Out-of-Sample ValidationLimitations
- Sample size: 159 trading days provides substantially improved statistical grounding vs. the initial 78-day study. Key signals (Episode warning, base rate) are now reliable. Sub-group analyses at n<25 (e.g., DOWN × Positive delta, n=24; extreme volume buckets) still require caution. Findings should be treated as directional signals pending out-of-sample validation.
- Single instrument: All data is NQ continuous front-month futures only. Findings may not generalize to other equity indices (ES, YM, RTY), commodities, currencies, or different contract specifications.
- Single time period: January–April 2026 represents one specific market regime. ORB behavior may differ materially in higher/lower volatility periods, strongly trending environments, or different macro regimes.
- No transaction costs: Deviation hit rates measure raw price distance. Bid-ask spread, slippage, and commissions are not incorporated. Realized results would be lower, potentially materially so for tight targets.
- No entry logic: This is a structural analysis of follow-through probability, not a complete strategy. Entry price (market vs. limit), stop placement, and position sizing are unmodeled. Translating hit rates to expected value requires explicit assumptions about all three.
- Continuous contract roll: NQ.c.0 uses continuous front-month stitching. Roll periods may introduce small artificial price gaps that could affect a small number of sessions near quarterly expiration.
Conclusion
This study provides empirical evidence that the Opening Range Breakout in NQ futures, when defined using Volume Profile Value Area structure, exhibits meaningful and consistent follow-through properties. The 71.1% base rate at 0.5× deviation is a strong foundation — stable across doubling the dataset from 78 to 159 days — confirming this is a genuine structural property of NQ futures, not a small-sample artifact.
The expanded dataset resolves several ambiguities from the initial study. The Episode Outcome remains the single most actionable classifier: the "Cross + Miss" warning pattern drops continuation probability by 48.4pp (from 42.7pp in the initial study), now backed by 44 occurrences. The bo2 framework is upgraded — second-episode breakouts now exceed the base rate (76.5% at 0.5×), treating them as distinct setups is justified.
A critical new finding emerges with more data: a substantial directional asymmetry at deep targets. DOWN breakouts reach 2.0× deviation at 37.8% vs. 14.1% for UP breakouts — a 23.7pp structural gap that was invisible in the 78-day dataset. Practitioners should apply different target expectations by direction, not just by Filter Case.
Two initially promising signals fade with more data. The Opening Delta gap narrows from 16.6pp to 7.5pp. The BO candle delta effect for DOWN breakouts disappears entirely — what appeared as a near-100% signal in the initial study was a small-sample artifact (n=6). Volume likewise shows no edge. These corrections highlight the risk of over-fitting to sub-group patterns in small datasets.
The primary path forward is out-of-sample validation on 2024 data and cross-instrument testing (ES, YM, RTY). The directional asymmetry finding and the Episode warning signal are the highest-priority candidates for further validation before drawing operational conclusions.
Methodology note: All analysis performed in Python using Databento MBO tick data.
Value Area computed using standard 70% single-tick expansion from POC.
All times Eastern. 5-minute bars use left-label / left-close convention.
Breakout candle = first bar closing outside filter range from 10:05 ET onward.
Deviation multiple = max favorable move from breakout price / filter range size.
Disclaimer: This paper is for research and informational purposes only.
Nothing herein constitutes financial advice or a recommendation to trade any instrument.
Past statistical patterns do not guarantee future results.