On Tuesday 16 Jun 2026, the day the weekly Nifty contract expired, the strike with the most open interest was 24,000. Calls and puts there held 5.84 crore units between them. The index closed at 23,989.15, only 10.85 points away. A screenshot of that chain says the strike pulled the index in. One expiry proves nothing, so this lesson asks three separate questions of the files. Do closes cluster at strikes? Does the heaviest strike attract the close? Do the big call and put strikes hold as a ceiling and a floor? Every figure comes from NSE end of day files and is dated. Nothing here is live and nothing is a view on where the index will settle next.
3By the end you can
- State the pinning claim as three separate tests, each with its own baseline.
- Work out the chance share for closes near a strike and the band that counts as evidence.
- Say which settlement regime a pinning test covers and why a new regime restarts it.
The idea, plainly
What pinning would mean
Pinning is the claim that an underlying finishes near an option strike more often than chance allows. The usual story has two parts. Option dealers hedge with the underlying, and near a strike with heavy open interest their hedges push against a move: they buy as the price dips and sell as it rises. Some traders also say large writers defend a strike. The first part is a mechanism that has been measured in US stocks. The second is a story, because the files do not say who holds which side.
Read one more paragraph, 74 words
India adds a complication. NSE publishes open interest by strike, and a separate participant file that splits index calls and puts by category, summed over every index, strike and expiry, never by strike. So nobody outside the exchange can tell whether the open interest at 24,000 sits with dealers who hedge, with spread traders, or with buyers who will let it lapse. The module on the other side of the trade goes into that.
Heaviest strike, max pain and a wall are three different things
The heaviest strike adds calls and puts at one strike. Max pain is the strike where the payout to all buyers is smallest, which uses every strike at once. A wall is the biggest call strike above the index or the biggest put strike below it. On the expiry day chain of 16 Jun 2026 the heaviest strike was 24,000 and max pain was 23,950. They differ, and the stories told about them differ too. Say which one a claim means before testing it. The max pain module tests the second.
A magnet or fuel: the sign problem
A dealer who is long options gains from movement and hedges against it, buying dips and selling rallies. That damps moves and can hold a close near a strike. A dealer who is short options hedges the other way, selling dips and buying rallies, which feeds a move. The same open interest can therefore be read as a magnet or as fuel, and the files do not say which dealer holds which side. That is the same sign problem the gamma exposure module shows with three conventions, and it is why a chart of open interest alone cannot settle pinning.
Three claims, three tests, three kinds of file
Which close is being tested
The expiry settlement for Nifty options is the index close of expiry day, so a test of pinning is a test of that close. The method behind the close is changing. Before 3 Aug 2026 the close came from a last half hour volume weighted average of the constituents. A closing auction session set it from 3 Aug 2026, and SEBI's circular of 15 Oct 2026 moves index expiries back to a last half hour average of continuous trading from 7 Dec 2026. Every expiry tested below falls before 3 Aug 2026. A result for one regime says nothing yet about the next, and the test starts again for it.
The maths of the first test
Strikes near the money are fifty points apart. If closes had no link to strikes, the distance from a close to the nearest strike would be spread evenly between zero and twenty five points. The chance of landing within five points is then ten out of fifty, one in five.
Chance within a band
= 2 x band / gap
Spread of a share
= sqrt(p x (1 - p) / n)
Here the band is five points, the gap fifty, so p is one in five. With 55 expiry closes the spread is about 5.4 percentage points. A share has to fall outside about two spreads either side of the chance level before it says anything, which is a band of roughly 9 to 31 percent.
Worked example: a year of expiries, then nine chains
Test one: do closes cluster at strikes?
The files hold every Nifty close from 1 Jul 2025. Weekly expiries fell on Thursdays until Aug 2025 and on Tuesdays after, moving to the day before when a Tuesday was a holiday. That gives 55 expiry closes up to 14 Jul 2026 and 201 other closes in the same window. For each close I took the distance to the nearest multiple of fifty and asked how often it was within five points.
Expiry closes were no closer to a strike than the other days
Read one more paragraph, 78 words
Seven out of those expiries is lower than the eleven that chance gives, and lower than the other days. It also sits inside the band from the fold, so I do not read it as an anti pin either. The honest result is no clustering found. US studies of single stocks have found clustering at strikes on expiry dates. This sample of one index finds none, and a sample this size could only have shown a strong effect.
Test two: does the heaviest strike attract the close?
Now the chains. For each expiry in the same window that has a chain and a close in the files, I took the last chain before expiry day and found the strike with the most open interest, calls and puts together. The baseline is the plain forecast: the close of the chain date. The score is which of the two sat nearer to the settlement. Nine expiries qualify, from 31 Jul 2025 to 7 Jul 2026, with chains from one to fifty six days out. The 28 Jul 2026 expiry falls just after the window; adding it makes six of ten here and eight of ten in test three, and changes no conclusion.
The heaviest strike was nearer on five of nine expiries, a coin flip
Read one more paragraph, 56 words
Five wins and four losses is what a fair coin gives half the time. The biggest miss, 30 Mar 2026, came from a chain eight weeks out, when the heaviest strike sat far above an index that went on to fall. Longer chains are not a fair test of a strike that traders adjust every week.
Test three: ceiling and floor
The wall claim says the largest call strike above the index acts as a ceiling and the largest put strike below it as a floor. On 29 Jun 2026 the largest call open interest was at 25,000 with 1.85 crore units, and the index closed at 23,946.25 that day. Across the nine expiries the settlement stayed between the largest put strike below the index and the largest call strike above it in seven. That sounds like support until you see how wide the walls are.
Open interest is spread over many strikes, so a wall is a wide zone
How many expiries would settle it?
Say the heaviest strike really were nearer than the plain forecast on seven expiries in ten. To tell that from a coin at a one sided five percent level, with an eighty percent chance of detecting it, takes roughly 37 expiries. Nine cannot do it, and the expiry closes can only test the clustering claim. A fair verdict on the second needs a year or two of chains, saved before each expiry.
Read one more paragraph, 74 words
The two walls sat between five hundred and twenty eight hundred points apart, wide enough to hold an ordinary move over the life of most of these chains whether or not anyone defends them. Both breaches, on 31 Jul 2025 and 30 Mar 2026, fell through the put wall, and both came from chains about two months out. Containment inside a wide band is a low bar. It does not show the strike acts.
Read the chart yourself
Take the open interest chart above. Count how many strikes carry at least half the open interest of the heaviest one. Then ask what you would conclude if the index had closed at the second heaviest strike instead. If the same story fits either outcome, the chart is not evidence. Do the same with the largest call strike: would a close above it have been called a breakout, and a close below it a ceiling that held?
The common mistake
The claim: The strike with the highest open interest is where the index gets pinned, so it acts as support or resistance on expiry.
One vivid case, 16 Jun 2026, fits the story. The tests do not: closes did not cluster at strikes on a year of expiries, the heaviest strike was nearer on five of nine, and both breaches of the walls came from chains about two months out. The verdict is not shown, which is not the same as disproved. What would change it is about forty expiries under the method that starts on 7 Dec 2026, a rule written before the data is pulled, and a clear win over the last close. The question page gives the short answer.
Key takeaways
- Pinning is three claims, and each needs its own baseline.
- Closes near strikes can be tested with index closes alone.
- A year of expiry closes showed no clustering at strikes.
- Nine chains cannot separate the heaviest strike from a coin.
- A new settlement method restarts the test.
3 terms this section turns on
Before you trade. Derivatives trading carries a high risk of loss. SEBI's study "Profitability of Individual Traders in the Equity Derivatives Segment (FY25 to FY26)" (August 2026) found that 87.7% of individual traders in equity F&O made net losses in FY26, with aggregate net losses of Rs 91,685 crore. Read the SEBI study