How to Choose a GEX Levels Service (and What to Check Before You Pay)

Dealer-gamma levels are sold by a lot of people now. The numbers they publish for the same index on the same morning are not the same numbers, and the differences are not small. Before you subscribe to any of them — including mine — it is worth knowing what actually varies, because almost none of it is disclosed on a pricing page.

I build a terminal that draws these levels, so I am not neutral. What I can do is tell you exactly which decisions produce different numbers, so you can ask anyone selling you levels how they made them.

What a GEX levels service actually sells you

Every provider does roughly the same four things:

• Pull the listed options chain for an underlying.
• Take open interest at each strike.
• Run it through an options pricing model to get gamma per contract.
• Weight by open interest, sum across strikes and expirations, and hand you the strikes where hedging pressure concentrates — usually a call wall, a put wall, and a gamma flip.

That sounds standardised. It isn't. Steps 1 through 4 each contain a choice, and each choice moves the levels.

The six choices that make the numbers differ

Which chain they read. Index options and the index ETF's options are different markets with different open interest. A service quoting levels for the futures may be deriving them from the ETF chain and scaling. That scaling ratio is an assumption, and it is rarely published.

Which expirations they include. Some use the front expiry only. Some include everything inside 60 days. Adding or dropping a monthly expiry with heavy open interest can move a wall by a meaningful distance. Ask for the window.

Open interest or volume. Open interest updates once a day, after the close. Volume updates continuously but describes what traded rather than what is held. A service using yesterday's open interest is telling you where positioning was; one blending intraday volume is estimating where it is now. Both are defensible. They are not the same number.

The dealer-positioning assumption. This is the big one and it is almost never stated. Nobody can observe dealer inventory. The industry convention assumes dealers are net short customer flow on both calls and puts, which is an assumption, not data. Every level you have ever seen rests on it. A provider who tells you this openly is more trustworthy than one who presents the output as measurement.

How often it re-computes. A level published at 8:30 and never touched again is a different product from a surface re-measured through the session. Neither is wrong. But if your plan depends on knowing whether size drained off a strike by midday, a static morning number cannot tell you.

Where it lands. A dashboard you check in another tab, versus levels drawn on the chart you are already watching. This matters more in practice than it sounds — a level you have to go and look up is a level you will forget during the part of the session when it mattered.

The question almost nobody will answer

Ask any provider this: where is the published record of your levels being graded against what the market actually did?

Not testimonials. Not screenshots of good days. A dated list of the levels published before the session, marked afterwards against the session's real high and low, with the misses still on it.

This is the single most useful question you can ask, for one reason: it is the only one that cannot be answered with marketing. Anyone can describe a methodology. Publishing a record means committing to the days it went wrong, in public, permanently.

If a service cannot show you one, you are being asked to take the levels on faith. That may still be fine — plenty of useful tools do not publish records — but you should know that is what you are doing.

What a graded record honestly looks like

Mine is public, and I will use it as the example precisely because it is not flattering.

As of 31 July 2026, across 22 settled sessions I have graded 88 wall tests. 28 of them broke. That is roughly a third, and it is the number I would want to know before paying anyone. The live figure is on the record page — if it has drifted since this was written, that page is right and this sentence is stale.

That figure is the argument for using levels correctly rather than against using them at all. A wall is a place where hedging concentrates, so it is a place where a move is more likely to stall or accelerate. It is not a fence. If someone sells you levels implying price cannot pass them, the historical record — anyone's historical record — disagrees.

A checklist you can use on anyone

Before you pay, get answers to these. A provider confident in their work will answer all six quickly.

• Which chain do you read, and if you scale, what is the ratio?
• What expiration window is included?
• Open interest, volume, or a blend?
• Do you state your dealer-positioning assumption?
• How often does the surface re-compute during the session?
• Where can I see your levels graded against sessions that already happened?

The last one is the filter. The other five tell you whether the numbers suit how you trade.

What I do, stated plainly

Sharpnel draws dealer-gamma levels directly on the price chart, next to the order flow — the depth ladder, footprint and tape — rather than in a separate dashboard. The surface re-measures through the session. Every trading morning at 08:30 ET the ES and NQ levels are published, and at 16:15 ET each one is marked against that session's real high and low. The full history, misses included, is public and free to read with no account.

There is a free tier on delayed data if you want to look at the mechanics before deciding anything.

The honest summary

Most of the disagreement between GEX providers comes from unstated methodology choices, not from one being right. Ask the six questions. Weight the answer to the last one heavily. And treat any level — mine included — as a place to have a plan rather than a line the market has agreed to respect.