Loading…
Market closedThe market is closed. Showing today's close, 16:00 ET. These numbers start moving again when it opens Thu 24 Sept at 09:30 ET.
Loading…
What everything here means, for someone new to the market.
no jargon, promise
When you buy or sell an option, someone has to take the other side. That someone is usually a dealer, and dealers do not want to bet on the market — they want to stay flat. So every time price moves, they buy or sell shares to cancel out the risk they just picked up.
That hedging is huge, it is automatic, and it is fairly predictable. On some days it pushes back against moves and the market feels sticky. On other days it pushes with them and the market feels slippery. GammaDesk works out which kind of day it is, and where the big piles of options sit that price tends to get stuck on.
That is the whole idea: it tells you what kind of day it is, not what to buy. Everywhere you see a ? next to a number, tap it for an explanation of what it means.
Open Positioning and look at two numbers: Gamma Regime and Gamma Flip.
CALM (positive gamma) means the market has padding today. Pushes get absorbed, and price tends to drift back to where it came from. WILD (negative gamma) means the padding is off. A push keeps going, and moves feed on themselves.
It is the difference between a car with working suspension and one without. Same bump in the road, very different ride.
The Gamma Flip is the price where one turns into the other. Above it, calm. Below it, jumpy. If price is sitting right on top of it, expect the day to change character.
Now find Magnet Above and Magnet Below. These are the two biggest piles of options either side of today’s price.
Price tends to drift toward a pile and then stall there, because that is where dealer hedging is strongest. Think of them as a ceiling above and a floor below.
Drawn in pencil, though, not ink. They are the levels most likely to matter today — not levels that are guaranteed to hold.
Knowing where a wall is does not tell you what price will do when it gets there. So do nothing until it arrives and shows you.
If price reaches the wall above and stalls, the wall is behaving as expected. If it slices straight through without pausing, the wall was not real today — and that is useful information too, because it usually means the move has more behind it than options positioning.
Waiting for confirmation costs you the first part of a move. Not waiting costs you the whole thing when you are wrong.
Where today’s walls are, price by price. This is the main page — start here.
A thousand pretend versions of the next month, so you can see what counts as a normal move.
The same read for baskets — big tech, chip makers, the indexes — plus how the wider market is breathing.
Every stock we track scored out of 100 and ranked, strongest to weakest.
Your own shortlist. Starred names are kept in this browser — no account needed.
Strikes that traded far more than usual today. It tells you something happened, not what to do.
What changed overnight — which walls grew, which shrank, and which are brand new.
The site’s own report card. What it said each morning, and what the market actually did after.
Three more you will see in the menu: Dashboard puts the headline numbers from all of the above on one screen, Ticker runs the full read on any stock you type in, and Daily is the whole day written out in a few sentences, with the postable version underneath it.
This list is the same text as the ? bubbles around the site — one file, so the two can never drift apart.
This teaches you to read conditions. It is not a buy or sell signal, and there is no page here that will ever tell you to buy or sell anything.
Knowing that today is a calm day with a wall at 775 does not tell you which way price goes. It tells you what kind of behaviour to expect if it gets there. Two people can read the same screen correctly and do opposite things with it.
Every number here is a model of the market, not the market. The gamma figures rest on an assumption about who is on the other side of each option trade, which is usually about right and sometimes wrong. The forecast holds volatility still, so it understates how bad a bad day can get. None of it knows about earnings, news, or anything a human decides tomorrow morning.
Quotes are delayed, and the heavy pages are worked out once a day rather than live. Treat this as a way to understand the shape of a market, not as a live trading tool.
Nothing on this site is investment advice. It is for information and education only. If you are about to risk money you cannot afford to lose, speak to someone licensed to advise you — that is not us.
What every number here is built from, and the assumption it all rests on.
Read the assumption first. Dealer positioning is assumed, not observed. The model takes the customer to be a buyer of puts and a seller of calls, which puts the dealer long calls and short puts. Nothing in an option chain records who was on which side of a trade, so this is a convention that fits index and large-cap books and can be exactly backwards on a heavily retail-traded single name.
Everything below inherits it. If it is wrong for the ticker you are looking at, every level on that page is wrong with it — not slightly off, but pointing the other way.
One option-chain snapshot per underlying, from Cboe’s delayed public feed. Every number on a positioning page is built from that single snapshot, so the price, the levels and the timestamp always describe one moment rather than three.
Calls and puts, both, at every strike inside the window: the nearest 30 strikes either side of spot, across the nearest 5 expirations (the forecast widens this to 20 from the same snapshot, at no extra upstream cost).
Open interest is as of the prior session’s settlement. It is published after the close and does not move during the day. That is worth sitting with: a level built from open interest describes positions carried into today, not positions opened during it. On a day with heavy new activity the map is one session behind the market it is describing.
Implied volatility is resolved per strike in this order: the quoted out-of-the-money volatility, then the quoted in-the-money one, then a value solved from the mid price, then a modelled surface. Which of the four produced each strike is counted and shown — when a quarter or more of the chain is modelled, the pages say so unprompted.
The walls on the plain-English view are picked nearest strong, in that order: look at the 8 strikes closest to spot on that side, and take the first one carrying at least 40% of that neighbourhood’s largest exposure.
Both halves of the rule are load-bearing. The nearest strike alone can be a trivial one price walks straight through; the largest alone can sit five percent away and have nothing to do with the next hour. One helper decides this for every page, so the homepage and /decision can never name different levels for the same book.
The gamma flip is not a strike. It is the price at which the modelled net exposure changes sign — a solved point on a curve, which is why no dollar figure is ever printed beside it. Printing $0 there would claim the level was measured and found empty, a different and false statement from “no figure applies”.
Gamma exposure is open interest weighted by each contract’s gamma — not a contract count. A strike with enormous open interest but negligible gamma barely registers, and that is correct: the question is how much hedging a price move forces, not how many contracts exist.
Gamma comes from Black-Scholes, using the risk-free rate and dividend yield shown in each page’s drawer. It is a model output, not a reported figure.
What it is not: a prediction, a measurement of anyone’s actual book, or a claim about direction. It describes where hedging pressure would concentrate if the assumption above holds.
The rule: Session volume divided by open interest, per contract A contract is listed when it trades at least 250 contracts, carries at least 50 open interest, and its ratio is at least 1.0×.
Why that ratio means anything: Open interest is yesterday’s settled position count So a contract trading more than its own open interest in one session means most of today’s activity is opening new exposure rather than shuffling existing positions.
Open interest as of: The prior session’s settlement It is published after the close and does not move intraday, which is precisely what makes it a usable denominator for today’s volume.
Volume as of: not yet computed Today’s traded contracts, from the same snapshot as the open interest.
This level is derived from: Volume — contracts traded during the session, not positions held. Unlike the positioning levels, nothing here is weighted by gamma or by any greek.
What it cannot tell you: Direction, or who traded A large print can be an opening buy, an opening sell, a hedge leg, or a roll. The chain records the trade, never the intent behind it.
The comparison not being made: Volume versus its own recent average That is the better screen and it is not available here: the source publishes today’s volume but no history of it, so it would take a stored daily series this app does not keep. Stated rather than replaced with a worse proxy.
In short: at least 250 contracts traded, at least 50 open interest, ratio at least 1.0×.
Every page showing positioning grades its snapshot against the market clock and shows a red banner when the answer is bad. The reference is the last moment the feed should have had something new to say — the current time while the market is open, the last session’s close otherwise — with 90 minutes of tolerance for the delayed feed, the cache, and a late scheduled job.
The simpler rule, “older than the last close”, was rejected deliberately: after 16:00 it condemns the correct end-of-day snapshot every evening, and a warning that appears nightly is not read on the morning it matters.
Which scheduled jobs have actually run is on /status.
It does not know who traded, or why. An option chain records contracts, never intent — every inference about dealers is the convention at the top of this page, applied.
It does not account for scheduled news. Positioning levels describe hedging pressure in an ordinary session; a Fed decision or a CPI print is a repricing that runs straight through them.
It is not advice, and no part of it says what to do. Nothing here is a forecast, including the pages with the word forecast in the title — those show a spread of simulated outcomes, which is a description of uncertainty rather than a prediction.