Max Pain Measures a Payout, Not a Destination.

Every month, a chart circulates with a single number on it and a story attached: max pain is here, so price is going there. The number is usually real. The story does not follow from it.

Max pain is an arithmetic result, and a narrow one. Knowing exactly what it counts is the difference between a useful reading and a confident mistake.

What the calculation actually does

Take every open option on one expiry. For each listed strike, ask a single question: if the contract settled at that price, what would writers owe holders in total?

For a call, the holder is owed the settlement price minus the strike, when that is positive. For a put, the strike minus the settlement price. Multiply each by that instrument's open interest, add them all together, and you have the payout at that candidate price. Repeat for every strike. The one with the smallest total is max pain.

That is the whole method. It uses open interest and strikes. It does not use price history, volume, order flow, funding, or anybody's opinion. It answers one question — at which listed price would the least money change hands at settlement — and nothing else.

A live reading, with the gap on display

Taken from Deribit on 16 September 2026 at 15:43 UTC, with the BTC index at 75,697:

Nine days before the largest expiry of the board, the max-pain strike sits nearly five per cent below the market. If the number were a magnet, it is a weak one.

This is the ordinary case rather than an exception. Max pain moves as open interest is written and closed, and spot moves for its own reasons; the two are not tied together by anything mechanical. A gap of several per cent a week or two out is unremarkable. Treating that gap as a forecast means reading a distance as a direction.

The assumption the story needs, and crypto does not supply

The "price gravitates to max pain" argument depends on a chain: market makers are short the options, they hedge continuously, their hedging pushes price, and the net of that pressure lands near the strike that minimises payout.

Each link is an assumption, and in crypto none of them is observable. There is no published data on who holds which side of an open contract. A large block of calls might be a fund's covered position against spot it already owns, a market maker's inventory, or a directional bet. The open-interest number is identical in all three cases, and the hedging behaviour is completely different.

Equity index options at least have consolidated exchange reporting that supports some inference about dealer positioning. Crypto venues publish open interest by instrument. Everything beyond that — who is short gamma, who is hedging, in which direction — is modelled, not measured.

That is why Basis labels dealer gamma, vanna and charm as modelled figures and prints the positioning assumption beside them. A model is worth reading. A model presented as an observation is how a trader ends up defending a level that was never there.

What the same numbers do support

Drop the forecast and the data is still useful:

Where the market has written contracts. On that 25 September expiry the largest single positions are 70,000 calls (10,944), 85,000 calls (9,804), 90,000 calls (9,437) and 70,000 puts (9,326). Strikes run from 30,000 to 320,000 across 66 listings. That distribution tells you which prices the market bothered to trade around — a map of attention, not of destiny.

Which expiry matters. One date holds 43.3% of all open interest. The others are noise by comparison. If an expiry is going to matter at all, it is that one, and knowing the date is worth more than knowing the strike.

What a ratio does not say. Calls outnumber puts on that expiry by roughly two to one, a put/call ratio of 0.53. It is tempting to read that as bullish. It is also the board's normal state: across every listed BTC expiry the same day the ratio was 0.55, and each of the five largest expiries fell between 0.53 and 0.70. Calls routinely carry more open interest than puts here, so the informative question is not whether the ratio is below one but whether it has moved against its own recent range. A reading compared to nothing is a reading that means nothing.

Implied against realized, carefully

On the same view, Deribit's DVOL — a thirty-day implied volatility index — read 38.2, while Deribit's own historical volatility series read 33.9.

The obvious move is to subtract and call the difference a premium. Resist it long enough to check one thing: the two series are not measured over the same window or by the same method. Implied volatility is a forward-looking price; realized volatility is a backward-looking calculation over some fixed lookback. Comparing them is legitimate and routine, but the comparison is between two different objects, and the number that comes out of the subtraction is smaller than the uncertainty in the comparison more often than people admit.

Implied sitting above realized is the usual state of an options market, not a signal on its own.

A workflow in Basis

1. Open the options workspace and fix one venue. Merged views are convenient and hide which venue a number came from; volatility in particular is read from one reference venue rather than averaged.
2. Find the expiry holding most of the open interest before looking at any strike. The date is the finding.
3. Read max pain as the payout-minimising strike it is. Write down its distance from spot, then write down what would have to be true for that distance to close.
4. Look at open interest by strike for concentration, not for direction. Note the two or three strikes that dominate.
5. If you read dealer gamma, read the assumption printed with it in the same breath. If you cannot state the assumption, you cannot use the number.
6. Set the invalidation level before the expiry, not during it. An expiry is a scheduled event; deciding in advance is the only part of it you control.

The honest summary

Max pain tells you where the least money would change hands. That is a real fact about the board as it stands today, and it will be a different fact tomorrow as contracts are written and closed.

It is not a target, not a probability, and not a plan. The most useful thing an options board gives a trader is a calendar and a map of concentration. The least useful thing is a single number repeated with confidence.

Method note: figures read from Deribit on 16 September 2026 at 15:43 UTC via the Basis options workspace, and they will have changed since. Max pain computed as the listed strike minimising total intrinsic payout across open calls and puts on that expiry, in coin-denominated open interest. Deribit is one venue and the largest for BTC options; other venues list their own boards. Nothing here is a forecast or personalized investment advice.