SPX GEX Report: what happenedafter 382 gamma flip crossings.
We took 62 SPX sessions of DealerEdge production records — 21,947 regular-hours minute observations — and measured what price actually did in each GEX regime. Some findings back the standard playbook. One cuts directly against it. We are publishing all of them.
Median minute-to-minute volatility in Rating 1–2 regimes vs 4–5
1.98 vs 1.13 bps per minute. The regime taxonomy shows up clearly in realized moves.
Flip crossings still on the new side 30 minutes later
Of 382 confirmed crossings. A single cross of the Flip Point is a coin flip, not a signal.
Sessions where price came within 0.10% of the Anchor
Every session in the sample interacted with the largest-GEX strike at some point during regular hours.
Does the GEX regime map describe what SPX actually does?
The gamma exposure playbook makes three claims you will find repeated across the GEX literature: positive-gamma regimes are quiet and negative-gamma regimes are violent; crossing the Flip Point changes market character; and price gravitates toward the strike with the heaviest dealer positioning. These claims are testable, and platforms that sell GEX data rarely test them in public — the incentives point the other way.
So we tested them on our own production data. The sample: 62 SPX sessions from May 27 – August 21, 2026, yielding 21,947 regular-hours minute observations sampled from DealerEdge session records — the same Flip Point, Anchor, and GEX Rating values members saw live, not a backfit reconstruction. DealerEdge itself updates in real time; for this study we sampled those records at one-minute resolution, which is the natural grain for asking regime-level questions.
For each minute we recorded the GEX Rating in effect, the price position relative to the Flip Point and Anchor, and the absolute price move over that minute, measured in basis points. Then we asked the three questions above and let the medians answer. Two of the three claims held up well. The third — the one that gets the most airtime — mostly did not.
The rating bands are real volatility bands.
Median absolute 1-minute SPX move by GEX Rating band, 21,947 minutes.
This is the cleanest result in the study. In Rating 4–5 minutes (positive-gamma, stabilizing), the median absolute 1-minute move was 1.13 bps (n = 11,776; mean 1.69, 90th percentile 3.67). In Rating 3 minutes it was 1.29 bps (n = 2,527). In Rating 1–2 minutes (negative-gamma, amplifying) it was 1.98 bps (n = 7,521; mean 2.97, 90th percentile 6.29).
Rating 1–2 minutes were roughly 75% more volatile at the median and about 71% more volatile at the 90th percentile than Rating 4–5 minutes. The ordering is monotonic — 4–5 quietest, 3 in the middle, 1–2 hottest — and it holds at the tails, not just the center. That is what you would expect if the rating is genuinely tracking a hedging-pressure regime rather than noise, and it is consistent with the rating being a volatility-regime score, not a directional signal: nothing here says which way price went, only how hard it moved.
For scale: the sample spent 8,226 minutes at Rating 5, 3,550 at Rating 4, 2,527 at Rating 3, 2,202 at Rating 2, and 5,319 at Rating 1 — a summer that offered plenty of both calm and stress to measure.
The flip is a regime line, not a trade signal.
Only 47% of flip crossings held for 30 minutes.
Across the 62 sessions we found 382 confirmed Flip Point crossings in 37 sessions, using a 0.05% hysteresis band to filter out quote noise (price had to be at least 0.05% beyond the Flip after having been at least 0.05% on the other side). Thirty minutes after a confirmed crossing, price was still on the new side just 47% of the time.
That is a coin flip. If you treat a flip cross as an entry trigger, the data says you are trading noise. We also checked whether crossings ignite volatility: the median 30-minute realized range was 22.3 bps before a crossing and 20.7 bps after. No ignition effect — if anything, a hair quieter.
We want to be precise about what this does and does not say. It does not say the Flip Point is meaningless — section 04 shows the two sides of the line are measurably different volatility environments, and that difference is the entire point of the level. It says the moment of crossing carries far less information than the marketing around gamma flips implies. Roughly 382 crossings in 37 sessions also means that on days when SPX trades near its flip, it tends to cross repeatedly — chop around a boundary, not a clean one-way regime break.
The practical read: use the flip to know which regime you are in — how much movement to expect, how to size, how far to trail — not as a buy or sell trigger. Our Flip Point explainer covers the regime framing in depth, and trading the crossing itself sits high on our list of common GEX mistakes for exactly the reason this data shows.
Below the flip runs about 55% hotter.
Here the classic claim holds. Splitting every minute by which side of the Flip Point price was on: minutes spent above the flip had a median absolute move of 1.19 bps (n = 14,272); minutes spent below the flip ran 1.84 bps (n = 7,552) — roughly 55% hotter.
Put sections 03 and 04 together and the picture is coherent: the Flip Point separates two genuinely different volatility environments, but the act of moving between them is not itself an event. A thermostat boundary, not a tripwire. The regime is the information; the crossing is mostly noise.
Price found the Anchor in 62 of 62 sessions.
| Measure | Result | Reading |
|---|---|---|
| Touched within 0.10% | 62 of 62 sessions | Price came within 0.10% of the Anchor at some point in every session sampled. |
| Close within 0.25%, Rating ≥ 4 days | 34 of 35 (97%) | High-rating (pin-prone) days closed near the final Anchor almost every time. |
| Close within 0.25%, other days | 25 of 27 (93%) | Lower-rating days closed near it only slightly less often — a modest differential, reported as-is. |
The gravity claim looks strong at first glance — a 100% touch rate — but this is where an honest caveat matters most. The Anchor is recomputed intraday as positioning changes. It is the strike carrying the largest absolute net GEX at that moment, and that strike migrates through the day, often toward wherever open interest is concentrating — which is frequently near where price already is. So these touch and pin statistics describe how price behaves around a migrating center of gravity, not around a fixed line drawn at the open. That makes the 62-of-62 figure partly a statement about the Anchor chasing price, not only price chasing the Anchor. A fixed-morning-Anchor version of this test is on our list for a future update.
The close-proximity split is more informative because it is anchored to the end of day: 34 of 35 (97%) of Rating-4-or-5 days closed within 0.25% of the final Anchor, versus 25 of 27 (93%) of other days. That is directionally what pinning theory predicts — stronger positive gamma, tighter close to the heavy strike — but 97% versus 93% on a 62-session sample is a modest differential, and we would rather report it plainly than round it up to a slogan.
Method & limitations.
How the study was run
- Source. Minute-resolution samples of DealerEdge production session records for SPX — the levels and ratings as they were computed live, not recomputed after the fact. How those levels are built is documented in our GEX methodology.
- Window. Regular trading hours only, 62 sessions, May 27 – August 21, 2026. 21,947 minute observations after filtering to minutes with a valid price and level state.
- Crossing definition. A confirmed crossing requires price to move at least 0.05% beyond the Flip Point after having been at least 0.05% on the other side. The hysteresis band exists so that quote flutter around the level does not count as ten crossings.
- Statistics. Medians are quoted ahead of means throughout, because minute-level move distributions are heavily right-skewed and means are dragged by a handful of violent minutes. Realized moves are expressed in basis points of the index level.
What this study cannot tell you
- One sample, one summer. 62 sessions from a single May–August window. A different volatility regime — a real drawdown, a rate shock — could shift every number here. The page updates in place as the sample grows.
- Levels recompute intraday. As discussed above, the Flip Point, Anchor, and Rating all update as positioning changes. These results describe behavior around live, moving levels — the thing members actually see — not around static morning lines.
- Association, not causation. Negative-gamma minutes being hotter does not prove dealer hedging caused the heat; regimes and volatility share drivers.
- Descriptive statistics, not trade advice. Nothing on this page is a recommendation to buy or sell anything. See our risk disclosure. How we decide what to publish — including the obligation to publish results that cut against our own product marketing — is in our editorial policy.
What one of these sessions looked like live.
Aggregates compress away the texture of an actual trading day. For a single-session view of the same data — a flip loss, the chop the crossing statistics predict, and a reclaim into a positive-gamma afternoon — read the SPX June 11 reclaim case study, which walks one session from this sample minute by minute.
The same levels, in real time.
Every number above was measured from the levels DealerEdge publishes to members in real time — the Flip Point, the Anchor, and the 1–5 GEX Rating, live on every covered ticker.