Love Vtrender Charts? Check out our new offer!
In October no index rebalance landed. For the first time since the Closing Auction Session began on 3 August, what the mechanism does on an ordinary day is visible without a large event sitting on top of it.
That makes the October edition of our market structure report the most useful of the three so far. Auction turnover fell by nearly half, and what is left underneath is the baseline — the thing the first two months could not show you.
If you have not come across it yet, here is the whole thing in one paragraph. Since 3 August 2026, every stock that carries a futures or options contract stops trading continuously at 3:15 pm. For the next twenty minutes orders collect but nothing matches. At about 3:35 the exchange works out the single price at which the most shares can change hands, matches everything there, and that price becomes the official close. It is the price your holdings are marked at, the price the index is built from, and the price your stock futures and options settle against at expiry.
We have covered ground on CAS earlier at - https://vtrender.com/posts/what-the-closing-auction-cas-has-done-to-your-close
Two details decide everything else in this report. The auction does not start from the last traded price. It starts from a reference price, which is the average price weighted by volume between 3:00 and 3:15. And orders in the auction cannot go further than 3% either side of that reference price.
So there are two prices every day for every stock: the one the auction inherited at 3:15, and the one it produced at 3:35. The distance between them is the mechanism's own footprint. NSE publishes both, every day, for every stock. That distance is what we measure.
Index providers rebalance on a schedule. When MSCI, FTSE or the Nifty review takes effect, the funds that track those indices have to buy and sell at the closing price, because the closing price is what they are measured against. The auction is where that money prints. In August one such day, the MSCI review on 31 August, carried Rs 39,718 crore through the auction, which was 22% of everything that traded in the NSE cash market that session. In September three event days carried two-thirds of the month.
October had none. No MSCI, no FTSE, no Nifty review. Total auction turnover fell to Rs 37,813 crore from Rs 68,007 crore in September. The largest session of the month was monthly expiry on 27 October at Rs 5,980 crore, an eighth of the September peak, and even that was spread across a wide set of names rather than concentrated in a handful.
Strip the event days out of all three months and the ordinary session is doing something the headline totals were hiding. The median ordinary session cleared Rs 1,137 crore in August, Rs 1,354 crore in September and Rs 1,551 crore in October. That is a rise of 36% in two months, in each month, while cash market turnover barely moved.
We said in the September edition that two months showed no established trend in that baseline. Three months do. The headline number for any given month is decided by the index calendar, which is a scheduling fact rather than a market one. The baseline is the number that describes the mechanism, and it is going up.
The three-month baseline table, the full event-session list and the share of cash turnover the auction clears are on page 3 of the report.
The headline measure of this series is how far the auction moves the close away from the price the stock was trading at between 3:00 and 3:15. In August that distance averaged 0.323% on an ordinary session. In September it fell to 0.269%, which we reported as the auction settling down. In October it rose to 0.313%.
Taken alone that reads as the mechanism giving back a month of improvement. It is significant in the statistical sense and it is broad: of the 210 stocks measured in both months, 144 widened. On a Rs 1,000 stock the typical gap went from about Rs 2.69 to about Rs 3.13.
But an auction that has to find a price in a market that is moving will move the price further, and October was a more volatile month than September. Measured across the whole universe, the spread of daily stock returns ran at 1.49% in August, 1.58% in September and 1.79% in October. The three months are not comparable until you account for that.
Divide the auction's move by how much the market itself was moving and the series reads 0.217 in August, 0.170 in September and 0.174 in October. On that basis September's improvement is essentially intact, and October sits 20% below August rather than 3% below it. The raw widening is mostly a busier tape, not a mechanism going backwards.
We report both, because they answer different questions. If you have to execute into the close in October, you face the raw number, and it is wider. If you want to know whether the auction is working better than it was in August, the adjusted number is the one that answers it, and it says yes.
The futures-to-cash basis does exactly the same thing. Raw, it widened 24% against September. Adjusted for volatility it reads 0.321, 0.234 and 0.253 across the three months — the same shape, from a completely separate calculation on a different set of files.
The side-by-side charts, the significance tests and the paired symbol-by-symbol comparison are on page 2 of the report; the basis work is on page 4.
full report at - https://acrobat.adobe.com/id/urn:aaid:sc:ap:93b6065c-6b8b-452d-be73-1f0bb6813b1e
There is a second thing the auction does besides moving the close, and it is more interesting than the first. It moves it in a direction.
If the auction were purely a matching mechanism with no tilt, the final price would land above its reference price about half the time. It does not. In August it closed above 60.0% of the time, with an average signed move of +0.138%. In September that fell to 54.8% and +0.039%, and we reported the upward tilt as more than halved.
In October it went to 64.6% and +0.143%. That is the highest of the three months, and volatility does not explain it. A more volatile market makes moves bigger in both directions; it does not push them one way. This is the one measure in the October data where the September improvement did not hold, and the report says so plainly rather than burying it under the adjusted figures that did hold.
What it means is not settled. One month is one month, and an upward tilt in the close is consistent with several things, including simply that buyers were more willing than sellers to pay for certainty of execution in a rising month. It is on the list to measure again in November.
Single-stock futures do not trade evenly across a month. They bunch into the few days before expiry, and the reason is mechanical rather than emotional. A futures contract has a fixed end date. A trader who wants to stay in the position past that date has to close the front-month contract and open the next one. That is one decision and two trades, so the volume roughly doubles by construction. It is the system at work, not a surge of conviction.
In August and September the four sessions to and including expiry carried around 44 lakh contracts a day against a quiet body of the month nearer 9 lakh. October was no different in that respect — the rollover window ran about four and a half times the quiet sessions, the same as the two months before it.
What changed is where the extra activity went. Across the whole single-stock futures universe, October averaged 17.58 lakh contracts a day against 15.50 lakh in September and 15.62 lakh in August. Line the months up by distance from expiry rather than by calendar date, and the early cycle rose 18.5%, the middle rose 17.3%, the rollover window rose 11.4%, and expiry day itself fell 7.2%.
That is the opposite of September, where the early sessions drained and expiry day gained. September moved activity around inside the month. October added it across the whole month, and added most of it to the quiet part. Concentration into the sessions around expiry fell for the first time since this series began.
This matters if you read volume for context. A month that gains only on rollover tells you nothing about ordinary participation, because the rollover is arithmetic. A month that gains on the quiet sessions is telling you something different.
20 October was a market holiday. The NIFTY weekly option series that would have settled that Tuesday settled instead on Monday 19 October — the first index expiry to fall on a day other than Tuesday since the closing auction began. The cycle leading into it ran four sessions instead of five.
The shortened cycle carried less premium in total than the full cycles on either side of it, which is what you would expect from losing a day. But expiry day itself carried Rs 75,918 crore, more than either neighbouring Tuesday expiry. Compressing the week concentrated the settlement rather than reducing it.
One observation is not a finding, and we have said so in the report. The next displaced expiry is the test. We have noted it because it is the kind of thing that is easy to miss and impossible to reconstruct later if nobody wrote the number down at the time.
That session also recorded the second-lowest auction turnover of the month. Index options settle against a computed index level, not against the stock auction prices, so an index expiry places no particular demand on the closing auction. The two mechanisms happen to share the same clock and otherwise leave each other alone — which is itself worth knowing if you have assumed they interact.
The full October edition runs to eight pages. Beyond what is covered above it carries the session-by-session auction chart across all three months, the complete list of index event sessions with what drove each one, the cash and derivatives turnover comparison — cash flat, derivatives up 18.1%, the ratio between them moving from 1.48 to 1.75 in a single month — and a sector breakdown showing which parts of the market carried October's increase and how much of it sat in ordinary sessions rather than rollover.
It also carries a short revisions note. The September edition described its expiry-concentration figure as covering four sessions when the number quoted was a five-session figure. The numbers themselves were right and consistent between August and September; the label was not. Both bases are published in the October edition so the series ties either way. We would rather print that than quietly change a number.
The companion study published alongside it ranks the hundred most active single-stock futures contracts by participation, with the median session and the quiet-day floor next to the average, and adds a new layer this month on what the cash market actually delivered behind each contract.
The October report is attached as a PDF. The three-month tables, the significance tests, the method notes and the full session data are all inside it.
link - https://acrobat.adobe.com/id/urn:aaid:sc:ap:93b6065c-6b8b-452d-be73-1f0bb6813b1e
Vtrender Market Structure Desk · Vtrender Technologies Private Limited
Data: NSE Data- , 1 August to 30 October 2026.
Market-structure research and charting context only. Nothing here is a buy or sell recommendation, and nothing here says what any price will do next.
Markets don't move because indicators change.
Indicators change because markets move.
A practical reference for NSE & BSE derivatives traders. 366 pages, built to sit open beside live charts.