Gamma exposure ("GEX") tries to answer one question: when the index moves, does dealer hedging push it further, or pull it back? The idea is sound. Most published numbers are built on assumptions that are rarely stated, and on at least one arithmetic error that is easy to make and hard to notice.
This page explains how we compute it for the Nikkei 225 and the S&P 500, what the number is worth, and the two mistakes we found in our own implementation.
A dealer who is short options must hedge. If they are short gamma, hedging means selling into declines and buying into rallies — the hedge amplifies the move. If they are long gamma, they do the opposite and dampen it.
Gamma is largest near the strike and near expiry, so the effect concentrates around heavily traded strikes in the front month. Aggregate gamma across every strike, sign it by assumed dealer positioning, and you get a single number: yen (or dollars) of hedging flow per 1% move in the index.
Most gamma estimates need an options pricing model, which needs implied volatility, which usually means paying for data. JPX removes that step: its daily settlement price file publishes an implied volatility for every single strike, along with days to expiry and the reference index level. Black-Scholes gamma follows directly from official data.
We compute gamma per strike, weight it by open interest and the contract multiplier (¥1,000 for large, ¥100 for mini), and sum across the nearest expiries within ±10% of spot.
Here is the honest limitation, stated plainly: dealer positioning is not public. Every GEX calculation, ours included, substitutes a convention — dealers are assumed long calls and short puts. Nobody publishes whether that is true on any given day.
It is also incomplete by construction. Structured products hedged over the counter never appear in listed open interest. In the US, covered-call funds illustrate the scale of the gap: JEPI runs roughly $45bn largely through OTC equity-linked notes, invisible to any exchange-data calculation, while listed-option funds like QYLD (~$8.4bn) are visible. The same asymmetry exists in Japan through structured notes.
So treat the output as a sign and a shape, not a quantity. Is hedging flow amplifying or dampening? Where does it flip relative to spot? Those survive the assumptions. The absolute yen figure does not.
The aggregate hides the useful part. Our Nikkei readings for 21 August 2026:
| Region | Gamma per 1% move | Effect |
|---|---|---|
| Above spot | +¥39.5bn | dampening |
| Below spot | −¥58.4bn | amplifying |
| Net | −¥18.8bn | amplifying |
The signs are opposite. Hedging flow would cushion a rally and accelerate a decline — a market that grinds up and drops fast. A single net figure of −¥18.8bn tells you none of that, which is why we chart the profile across strikes rather than publishing one number.
The shape is also more stable than the level. On 20 August the same readings were +¥42.5bn above and −¥57.9bn below. The Nikkei moved +890 yen that day and −200 the next, yet the downside figure changed by less than ¥1bn. Net gamma has been on the amplifying side since 19 August (−¥44.0bn, then −¥15.4bn, then −¥18.8bn) even as the Nikkei Volatility Index drifted down from 29.7 to 28.4.
Options chains from most sources include contracts that have already expired. If you filter only by "days to expiry ≤ N" without also requiring expiry ≥ today, negative day counts pass through and contribute gamma that no longer exists.
We had this bug in our SPX calculation. Adding a single condition — expiry must be today or later — moved published SPX gamma exposure from $178.1bn to $87.1bn. The original figure was overstated by 51%. If a published GEX number looks large, this is the first thing to check.
Nikkei mini options expire weekly, and it is natural to assume that omitting them understates gamma badly. We assumed exactly that, then measured it.
About 85% of mini open interest already sits on the monthly SQ expiry and was being captured. Adding every weekly expiry contributed 571 large-equivalent contracts — roughly 0.3% of the total. The intuition was wrong, and we would not have known without checking.
One detail worth recording if you build this yourself: JPX labels mini open interest by last trading day, while the settlement file labels the same series by SQ day. They differ by exactly one calendar day. Join on the raw code and the weekly series silently disappears.
Every business day, for both markets:
Both are labelled as estimates, with the dealer-positioning assumption stated on the page.
→ Live Nikkei gamma profile ・ → SPX gamma and COT ・ → Nikkei options field guide