Do premarket runners hold their gains through the session and then spike again around noon? And are the after-hours moves really more frequent lately? Two questions, one twelve-month pass over every archived tape. One answer is no; the other is an emphatic yes.
Roughly one day in eight. Just over half hold the open at all (54.6%), and only a fifth of those put in a genuine noon leg. Over the last three months the pattern ran at 8.4% against 14.6% before — a real decline (p = 0.008), not a rise.
The share of PM runners printing an after-hours push of ≥ 10% above the 4pm close has doubled in the last three months. It is the single strongest shift in the whole dataset, it holds in every cohort, and it is visible in the wider market too.
Before testing the pattern it is worth fixing the shape of the underlying day, because it is not a gentle one. A stock that doubles in the premarket typically opens 39% below its premarket high, trades another 18% below that open before noon, and closes the regular session 48% under the premarket high. The move you see on the premarket scanner is mostly gone by the time the bell rings, and the question of "holding the gains" is really a question about holding the open, which is already a much lower price.
Percent moves. The control column is a deterministic 5% sample of every archived day in the window — itself already a selected population of 15%+ range days.
| Median measure | PM ≥ 100% | PM 50–100% | All days (control) |
|---|---|---|---|
| Open vs. premarket high | −39.2% | −28.2% | −3.7% |
| Deepest dip vs. open, before 11:30 | −18.4% | −11.1% | −6.0% |
| Deepest dip vs. open, whole session | −20.4% | −12.0% | −7.0% |
| Session high vs. open | +16.1% | +11.6% | +6.6% |
| 4pm close vs. open | −9.3% | −1.9% | −0.9% |
| 4pm close vs. premarket high | −47.6% | −32.1% | −7.4% |
"Doesn't drop more than 20% below the open" is a coin flip for a PM ≥ 100% runner — 54.6% clear it. The ladder below shows why the threshold matters: move the line to 15% and only 40% qualify; move it to 30% and 77% do. The gradient across cohorts is the real finding. The larger the premarket move, the worse the holding: names up more than 200% premarket hold the open only 39% of the time, versus 86% for the modest 30–50% group and 95% for an ordinary corpus day.
Bars are drawn at the 20% threshold — the one in the original question. Read across for the whole ladder.
| Cohort | n | −10% | −15% | −20% | −25% | −30% | at −20% |
|---|
If runners systematically spiked again around midday, the day's high would cluster there. It does the opposite. Over half of all PM ≥ 100% days set their session high in the first thirty minutes — 53% against 46% for an ordinary day — and the noon window (11:30–14:00) is the quietest stretch of the session for new highs, holding just 12.8% of them. The last half hour is a bigger cluster than any noon bucket. Runners are front-loaded; the tape's centre of gravity is the open, not the middle of the day.
Shaded band marks the noon window used throughout this report. Hover any bucket for exact counts.
The pattern is rare, but it is not random — and the "hold" half of it is the part that carries information. Among PM ≥ 100% runners that held the open into 11:30, 22.4% went on to make a new session high in the noon window with a leg of at least 10% off the 11:30 price. Among those that had already broken 20% below the open, only 7.4% did. That is a 3.0× lift from a condition you can observe live at 11:30, and the lift survives in every cohort. On a control day the same split gives 6.2% against 3.2% — the filter barely does anything when the tape is ordinary.
Same measure, two conditions. The gap is the edge.
Hold rate × noon-spike rate. The 100–200% band is the sweet spot: big enough to run, not so big that it collapses at the open.
| Cohort | n | Held | Spike | Both |
|---|
The recent impression does not survive contact with the monthly series. The joint pattern peaked at 23% in October 2025 and has been drifting down since; the last three months ran 8.4% against 14.6% for the nine before them (z = −2.64, p = 0.008). The noon-spike leg on its own is down too (12.2% vs 17.6%, p = 0.038), while the hold rate itself is statistically flat.
The reconciliation is in the second chart: June 2026 produced 162 qualifying runners, the most of any month in the window. Fifteen of them ran the full pattern — the third largest absolute count of the year — even though the hit rate was a below-average 9.3%. More candidates at a lower rate still means more instances crossing the screen. That is almost certainly what "we're seeing a few of those lately" is picking up.
Percent of scored PM ≥ 100% days. Monthly n runs 27–162, so single months are noisy; the trend is the signal.
Qualifying PM ≥ 100% tapes per month, with the count that ran the full pattern overlaid.
This is the finding of the study. Across the whole window, 29.3% of PM runners printed an after-hours push of at least 10% over their 4pm close. But the series is not stationary: it sat between 9% and 24% for the first six months of the window and has spent the last four between 38% and 48%. The last three months against the prior nine: 42.4% vs 21.5%, z = 6.40, p < 0.0001. After-hours moves of any direction ≥ 10% went from 60.5% to 75.1% of days, and the share of runners setting a fresh day high after the bell nearly doubled, 7.7% → 14.6%.
It is not confined to the runners. The 50–100% cohort shows the same doubling (15.8% → 27.9%, p < 0.0001) and even the broad control sample moved (9.9% → 14.0%, p = 0.011). Whatever changed, it changed the after-hours session market-wide, and it is amplified in the small-cap momentum names.
Runners against the control sample of all archived days. The two series track each other until March 2026, then separate.
Distribution of the after-hours high above the 4pm close, PM ≥ 100% runners, by month.
When the after-hours high prints, 30-minute buckets. The early window still dominates, but the tail has thickened.
Two qualifiers worth keeping. First, after-hours on these names is still net down: 49.0% drop at least 10% below the close against 29.3% that rise that much, and the median after-hours close sits 4.5% under the 4pm print. What has grown is the upside tail, not the drift. Second, after-hours liquidity is thin — the median runner does 1.6% of its daily volume after the bell across 224 traded minutes, so a 10% print is a real quote but not necessarily a fillable size.
The one place the two questions meet: on days that did run the hold-and-noon-spike pattern, 79% had an after-hours move of 10% or more — the most after-hours-active subset in the study.