Half-Gap Reach Rate Dashboard
Jan 3, 2019 – Feb 27, 2026 | 1,774 sessions | E-mini S&P 500
Choose what counts
Zero tolerance: RTH price must reach or cross the half-gap level. Getting close does not count.
Same ES sample: 75.4% zero tolerance versus 78.1% within 1.25 points. Within tolerance includes the zero-tolerance reaches; these are not separate probabilities to add together.
The question · ES half-gap
How often did price reach the midpoint?
75.4% of eligible sessions reached or crossed the half-gap. That is 1,337 events in 1,774 sessions.
Still unreached at 7:30 AM Pacific? Among the 729 eligible sessions remaining, 292 reached it later: 40.1%. This is a different, conditional sample—not the full-session rate.
2019–February 2026 · 95% interval for the full-session rate: 73.3–77.3%. Historical frequency, not a trade win rate or an executable fill. The half-gap uses the prior RTH close, not exchange settlement.
Reach Rate by Gap Width (pts)
Full gap width (pts) — zero tolerance: reached or crossed
Evidence behind the rate
1,337 qualifying events / 1,774 sessions. 95% Wilson interval: 73.3–77.3%. Gap-up: 70.0–75.6%; gap-down: 75.5–81.2%.
Intervals describe sampling uncertainty assuming independent sessions; they do not capture changing market conditions. This historical frequency is not an entry rule, an executable fill, or a trade win rate.
Gap size & directionFull sample counts, directional rates, and uncertainty by gap bucket.
Detailed Reach Rates — ES
| Gap Width | n (Up) | Reach% (Up) | n (Dn) | Reach% (Dn) | n (Total) | Reach% (Total) | 95% interval |
|---|---|---|---|---|---|---|---|
| 0-5 | 152 | 94.7% | 153 | 98.7% | 305 | 96.7% | 94.1–98.2% |
| 5-10 | 199 | 83.9% | 150 | 88% | 349 | 85.7% | 81.6–89.0% |
| 10-15 | 155 | 80% | 113 | 85.8% | 268 | 82.5% | 77.5–86.5% |
| 15-20 | 148 | 70.9% | 88 | 78.4% | 236 | 73.7% | 67.8–78.9% |
| 20-25 | 82 | 62.2% | 54 | 70.4% | 136 | 65.4% | 57.1–72.9% |
| 25-30 | 70 | 57.1% | 45 | 68.9% | 115 | 61.7% | 52.6–70.1% |
| 30-40 | 83 | 57.8% | 61 | 57.4% | 144 | 57.6% | 49.5–65.4% |
| 40-50 | 34 | 50% | 35 | 57.1% | 69 | 53.6% | 42.0–64.9%Small sample |
| 50+ | 65 | 36.9% | 87 | 50.6% | 152 | 44.7% | 37.1–52.7% |
| TOTAL | 988 | 72.9% | 786 | 78.5% | 1774 | 75.4% | 73.3–77.3% |
Still unreached? Explore the time checkpointsRemaining-session frequencies at five Pacific-time checkpoints.
Still unreached? What happened after each checkpoint?
Only sessions with no qualifying reach before the checkpoint enter this sample. Times are Pacific, as recorded in the source chart. These are pooled across gap widths and directions, not a personalized forecast.
| Checkpoint (Pacific) | Still unreached (n) | Reached later | Rate | 95% interval |
|---|---|---|---|---|
| 07:00 | 895 | 458 | 51.2% | 47.9–54.4% |
| 07:30 | 729 | 292 | 40.1% | 36.6–43.7% |
| 08:00 | 650 | 213 | 32.8% | 29.3–36.5% |
| 09:00 | 568 | 131 | 23.1% | 19.8–26.7% |
| 10:00 | 507 | 93 | 18.3% | 15.2–21.9% |
The checkpoint minute is included in the remaining window. A session must have a bar in that minute to qualify; early closes and missing bars can change the sample. Outcomes run to that session's available RTH end. Exact and tolerance samples differ because a near approach can remove a session sooner.
Year-by-year resultsCompare historical periods without treating them as a new holdout.
Has the historical rate changed?
2019–2023: 74.4% (923/1241). 2024–February 2026: 77.7% (414/533). This is a retrospective comparison, not an untouched out-of-sample validation.
| Year | Events / sessions | Rate | 95% interval |
|---|---|---|---|
| 2019 | 167 / 237 | 70.5% | 64.4–75.9% |
| 2020 | 181 / 257 | 70.4% | 64.6–75.7% |
| 2021 | 180 / 246 | 73.2% | 67.3–78.3% |
| 2022 | 194 / 252 | 77% | 71.4–81.8% |
| 2023 | 201 / 249 | 80.7% | 75.4–85.1% |
| 2024 | 188 / 245 | 76.7% | 71.1–81.6% |
| 2025 | 194 / 249 | 77.9% | 72.4–82.6% |
| 2026 (partial) | 32 / 39 | 82.1% | 67.3–91.0% |
Definitions & how to read this studySession windows, tolerances, exclusions, and interpretation.
What is the Half-Gap?
The half-gapis the midpoint between the prior regular-session close and today's regular-session open. The research measures whether price returns to this midpoint during RTH, using the selected definition above. It uses the prior RTH close, not an independently sourced exchange settlement.
How to use this
- A possible destination, not an entry — The half-gap is below a gap-up open and above a gap-down open. Your entry and risk rules remain independent.
- Match the sample — Use the market, full opening gap width, direction, and selected measurement. Rates generally decline as gaps widen, but not in every bucket.
- Check your study settings — Existing AMT Half-Gap Alert versions use tolerance-based rates. The dashboard toggle does not change your installed study.
- Track whether it already happened — Full-session rates do not estimate a second visit or the remaining chance after an unfilled morning. Use the separate time-checkpoint table below for sessions still unreached.
Key Findings — ES
- ES has an overall reach rate of 75.4% across 1,774 sessions (1,337 qualifying events). This is not a trade win rate.
- The smallest published bucket (0-5 pts) has a reach rate of 96.7%. Historical frequency is not a guarantee.
- Gap-down: 78.5%. Gap-up: 72.9%. Directional differences vary by market and gap bucket; use the matching group.
- The largest published bucket (50+ pts) has a rate of 44.7% under the selected definition. That alone does not determine whether a trade is appropriate.
Research definitions
One-minute OHLC data, January 3, 2019 through February 27, 2026. Half-gap = (prior RTH close + current RTH open) / 2. Zero tolerance counts a reach or crossing of the unrounded midpoint; a midpoint can fall between tradable ticks. Neither measurement guarantees execution at that price.
RTH uses the source chart's Pacific time: ES/NQ/RTY 06:30–13:00; CL 06:00–11:30 (end excluded). Minimum full gap: 1 pts for ES. Bucket lower bounds are included; upper bounds are excluded. To preserve the published sample, the final “+” bucket is capped below 999 pts. The NQ sample excludes one gap beyond that cap.
Comparison tolerances: ES 1.25 points; NQ 5 points; RTY 1 point; CL $0.05. Very small gaps can already satisfy the tolerance at the open without reaching the midpoint. Changing the definition updates every statistic, chart bar, and table row above.
From the result to your chart
How to read the Half-Gap research
A midpoint revisit answers a location question. It does not tell you whether an entry at the open would survive the move before that revisit.
What is being measured?
The reference is halfway between the previous RTH close and the current RTH open. Compare the zero-tolerance and within-tolerance results on the same market and sample, then inspect gap width and direction. A near approach and an actual reach are separate outcomes.
Walk through an example
Illustrative, not a historical session
If the prior RTH close is 6,000 and the next RTH open is 6,020, the half-gap is 6,010. A move to 6,012 is not an exact reach. Whether it counts in a tolerance view depends on that view's specified tolerance.
What can change the interpretation?
An overall reach rate mixes different gap sizes and directions. Smaller filtered groups have less evidence. The result also depends on the session definition and data coverage; the prior RTH close is not necessarily the exchange settlement price.
Try a structured review
Record the open, prior RTH close, gap width, and exact midpoint before the session unfolds. Review whether price reached it, how long it took, and the excursion first. Keep those observations separate from any entry and exit rules.
All research · New to the name? What AMT Balance means.