# Gamma weather is not a real options term. We made it up — here is exactly what it measures. > Two facts about a chain — which way dealer hedging pushes, and how close the boundary is — collapsed into one word. The full definition, the thresholds behind it, the window they are measured over, and the point where it stops being useful. *2026-08-31 · Written by the person building GammaGrid · * --- If you searched for "gamma weather" expecting an established options concept, the honest answer comes first: there isn't one. It is not in the Cboe glossary, no desk uses it, and nobody was calling anything this before we shipped it on 31 August 2026. We named it. Anyone telling you it is standard terminology is repeating us. What is not made up is the arithmetic underneath. Every number in a gamma weather card is a number the tool already computed and already showed you somewhere: net dealer gamma exposure, the gamma flip level, the call and put walls. The name is new. The measurement is the same one that has been on the screen the whole time. This article is the definition, so that the term can be used precisely or argued with precisely, rather than absorbed as vibes. ## What is gamma weather? **Gamma weather is a one-word summary of an option chain, built from two facts: whether dealer hedging is currently damping price moves or amplifying them, and how far the price sits from the level where that flips.** It comes out as one of five states — Clear, Fair, Unsettled, Showers or Storm — measured across the expiries falling inside the next thirty days. Everything after this paragraph is that sentence taken apart. **Fact one: the sign of net gamma exposure.** Sum estimated dealer gamma across the chain. Positive means dealer hedging works against the move — they sell into strength and buy into weakness, which damps price. Negative means hedging works with the move, buying into strength and selling into weakness, which amplifies it. This is the ordinary GEX calculation, described in full in [Gamma Exposure Explained](https://gammagrid.io/learn/gamma-exposure-explained/), with every limitation that article spends several paragraphs on. **Fact two: how far the price is from the gamma flip, in percent.** The gamma flip is the underlying price at which the cumulative profile crosses zero — the boundary between the two regimes. Distance is expressed as a percentage of the current price, because a two-dollar gap means something different on a $30 ETF than on a $700 index. The second fact is the entire reason this is not just a coloured plus-or-minus sign. A ticker sitting two tenths of a percent from its flip and a ticker sitting five percent from it are both "positive gamma", and they are not remotely the same day. The first can change regime before lunch on nothing in particular. The second takes an event. A summary that reports both as the same state is not summarising, it is rounding away the part that mattered. ## Why there is a made-up word here at all The app has eight views of a ticker. Every one of them is a lens: the GEX heatmap, max pain, the IV surface, open interest deltas, unusual activity. Each answers a specific question, and each assumes you arrived already knowing which question you had. What was missing was the line before all of that — *what kind of day is this ticker having.* Not a recommendation, not a score out of ten. The one-sentence description you would give a colleague who leaned over and asked what SPY looked like this morning, before either of you opened anything. That sentence always came out the same shape: the direction, then the distance, and nothing else. Small enough to name — and once it has a name it can sit above the view switcher rather than inside a view, because a summary you have to click for stops being a summary. ## The five states The two facts cross into five states. Green always means damping and purple always means amplification — the same two colours those things carry on every chart in the app, so a card that ends up in somebody's timeline as a screenshot is readable without the legend. | State | Rule | What the card says | |---|---|---| | **Clear** | positive gamma, flip ≥ 3% away | Dealer hedging is damping moves. | | **Fair** | positive gamma, flip 1–3% away | The damping may not hold. | | **Unsettled** | flip within 1%, either sign | The regime can turn either way. | | **Showers** | negative gamma, flip 1–3% away | Moves are amplified for now. | | **Storm** | negative gamma, flip ≥ 3% away | No boundary nearby; hedging amplifies moves. | Note what Unsettled does: inside one percent, the sign stops mattering enough to name. The card does not say "positive but fragile", it says the regime is the thing least worth relying on today, whichever side it currently falls on. That is a deliberate refusal to report a number more precisely than it deserves. Alongside the state, the card shows the four levels it was computed from — price, call wall, put wall, gamma flip — so the word is never the only thing on offer. If you disagree with the summary, the inputs are right there to disagree with. ## Where 1% and 3% came from, and what would replace them Straight answer: they are not fitted. Nobody ran a backtest, found that 1.0% maximised something and wrote it down. They are anchored to the size of an ordinary trading day — a level within about one percent is one a normal session can reach without help, and a level three percent out generally needs a reason. That is a judgement, and a judgement is worse than a measurement. It is worth saying so plainly rather than presenting round numbers as if they fell out of an optimisation. Both are settings rather than numbers baked into the calculation, for exactly that reason: once there is enough stored history to measure how often a flip at a given distance was actually reached the same day, they get replaced by the answer and this paragraph gets rewritten. The practical consequence for a reader: the boundaries are soft. A ticker at 0.99% and a ticker at 1.01% get different words and are the same situation. Read the distance, not just the label — it is printed on the card. ## The window: thirty days, not "the nearest few expiries" Every gamma number depends on which expiries you include, and this is where the first version was wrong in a way worth describing, because it is the kind of wrong that looks fine. The card originally read the entire chain. The GEX heatmap directly below it reads the nearest ten expiries. On AAPL that produced walls of 300/250 on the card against 360/300 on the tab immediately underneath, and a flip of 313.46 against 309.79, with the price at 319.70. Same labels, different numbers, one screen. Neither was miscalculated; 67% of AAPL's open interest sits beyond the tenth expiry, so "the whole chain" is mostly a description of contracts expiring next year. On GLD it was worse than cosmetic: the card put the flip at 414.90, above a price of 408.89, while the tab put it at 399.42, below. Two panels disagreeing about which regime you were in. The obvious fix is "use the nearest N expiries, same as the heatmap". A measurement killed it. On 31 August, the nearest three expiries meant **two days on SPY and 137 days on CPER** — a 68× spread. A count of expiries measures how densely the exchange chose to list them, not how far ahead anyone is looking, and one label cannot honestly cover both. So the window is stated in days. Thirty of them, and the reason is how much of the chain's gamma actually falls inside: | Ticker | Gamma inside 14 days | Inside 30 days | |---|---|---| | SPY | 37% | **64%** | | AAPL | 25% | **49%** | | SLV | 14% | **47%** | | PBR | 3% | **25%** | At fourteen days, PBR's card would be computed from three percent of its gamma. The flip is the most sensitive number on the card, and on that little of a chain it is noise — on GLD, going from one expiry to two moved the flip from 402.27 to 360.65. Thirty days is also the familiar front-month horizon, which means the label reads without explanation. The cost of that choice is real and worth knowing when you read the walls: a wider window pulls them toward distant strikes carrying years-old protective open interest. SPY's put wall under this window sat at 535 while the price was 769. That is not a level anyone is defending this month; it is a monument. One more rule, for thin calendars. CPER lists a single expiry inside thirty days, and a card built on one expiry is not a summary of anything — so the floor is two expiries even when the second falls outside the window, at 46 days in that case. The card then prints the range it *actually* used rather than the one it asked for, read back from the stored row. The two differ only where the floor fired, which is the only case the floor exists for. ## No flip at all is not the same as a distant flip Some chains have no gamma flip. The cumulative profile never crosses zero inside the listed strike range — SPX behaves this way, because below its lowest listed strike the profile is flat zero rather than negative. There is then no boundary to be near or far from, and the honest move is to say that rather than pick a state that implies a distance nobody measured. The card drops to the sign alone and states it in words: this chain has no gamma flip inside its strike range, so only the sign of gamma is known. Half a definition, labelled as half a definition. ## How to actually use it The card is a triage instrument. It answers "does this ticker need my attention today, and which view should I open first" — and it is designed to be wrong-proof in the sense that it always shows the levels it used, so the summary can be checked in about two seconds. - **Scan the watchlist for the state that changed.** A ticker that was Clear yesterday and is Unsettled today has had something happen to its positioning, and that is a more useful filter than any absolute state. The states themselves are mostly stable day to day; the transitions are the information. - **Let the state pick the view.** Unsettled or Showers means the flip is doing the work — open the GEX heatmap and look at how thick the profile is around it. Clear means the walls are the more interesting object, because that is the regime where price ranging between them is the ordinary outcome. - **Read the distance, not only the word.** Five states out of a continuous number is a lossy compression, on purpose. The number is on the card. - **Check the scope line.** It says how many expiries and how many days went into the card. When it reads something other than 30 days, the floor fired and you are looking at a thin calendar — treat the flip with more suspicion than usual. - **Check the timestamp.** The card is computed by the collector, right after each successful pass, not when you loaded the page. It is not a live tape. - **Share the card rather than a screenshot of it.** There is a Share button on the card [in the app](https://app.gammagrid.io/) that exports a square image with the state, the four levels and the collection time drawn into it. The point is that it survives being posted: a crop of somebody's dashboard needs the dashboard explained, and this does not. What it is not for: it does not say buy or sell, and it does not say up or down. Negative gamma means moves get amplified — in whichever direction they happen to go. That confusion is the single most common way people misread gamma exposure, and putting a weather icon on it does not fix that, so it is worth repeating in the article that introduces the icon. One deliberate non-decision: the card is not frozen at the open, even though "weather" suggests a morning forecast. It is recomputed on every collection pass, because a card that stopped updating at 9:35 would spend the afternoon contradicting the charts directly beneath it — and two numbers arguing on one screen is worse than a stale number honestly labelled. It only recomputes for tickers that collected successfully, so a failed pass leaves the previous card standing rather than dressing a failed collection up as a fresh calculation. ## Where it stops being useful Everything true of gamma exposure is true of this, because it is gamma exposure with a hat on: - **It is an estimate, not dealer positioning.** Nobody publishes real market-maker books. This is computed from public open interest and a Black-Scholes approximation of the greeks, like every other GEX number you have seen, including the expensive ones. - **The sign convention is a heuristic.** The model assumes dealers are net short calls and net long puts. Reasonable prior, unverifiable from public data, occasionally wrong. - **Five words are a lossy summary of a continuous surface.** That is the trade being made. It buys a glanceable first line and costs precision, which is why the numbers stay printed next to the word. - **Open interest updates once a day.** The regime described is the one implied by positions already on the books, which is a lagging picture of a forward-looking market. - **The thresholds are judgement, not measurement** — see above. Nothing about the state boundaries has been validated against realised outcomes yet. And the risk specific to naming things: a familiar metaphor makes an estimate feel more certain than it is. A weather icon carries an implicit promise of forecasting that this number cannot keep. It is a description of current structural conditions, in the same sense that "it is raining" is a description and not a prediction of when it will stop. ## Why weather, and not a score The alternative was a number — a 0–100 gauge, a letter grade, something rankable. We did not build that, for a reason that survives being stated plainly: a score invites you to compare tickers as if the scale meant the same thing on each of them, and it does not. It also strongly implies that higher is better, and neither regime here is better. Positive gamma is not good. It is one description of how hedging flow behaves, and there are people whose entire position wants the other one. Weather has the right shape precisely because it is descriptive and nobody mistakes it for advice. A forecast of rain is not an instruction to stay home; it is the reason you check whether you brought a coat. That is the whole of what this card claims to be. Gamma weather is our name for a summary of two estimated quantities, built from delayed public data, with thresholds we chose by judgement and intend to replace with a measurement. It is not investment advice, it is not a signal, and it is not a standard term you should expect a broker or a desk to recognise. It is a first line on a screen, which is exactly as much as it was built to be. [Gamma Exposure Explained: What Dealer GEX Actually Tells You](https://gammagrid.io/learn/gamma-exposure-explained/) — the number underneath all of this: where it comes from, why 0DTE made it matter more, and the four places it stops being reliable. --- Source: https://gammagrid.io/learn/what-is-gamma-weather/ GammaGrid — options positioning dashboard. Educational content, not investment advice.