Rollover in Futures: How It Works for Indian Traders

The first time a trader watches a futures contract drift toward expiry, the question is rarely technical. It is emotional: do I close it, roll it, or just wait and hope the broker handles it?

Every NSE futures contract has a fixed last day. As of 2025 that day is the last Tuesday of the contract month — the rule moved from the old last-Thursday expiry. If the view behind the trade has not died with the contract, the position has to be moved.

Hold one Reliance future bought a fortnight ago. Expiry Tuesday arrives, and the thesis still looks fine on the chart.

Here is the part that catches beginners out. A single-stock future like Reliance does not quietly cash out if you do nothing. Held to expiry, it is physically settled — a long position has to take delivery of the actual shares and pay their full value, and a short position has to hand the shares over.

Roll the position instead, and you stay in the same underlying, the same direction, the same approximate size, on a contract that has another month of life in it — and you sidestep the delivery obligation entirely.

This article walks through what a rollover actually is, how the two-leg trade gets executed on the NSE, the rupee math on a real Reliance roll, when to roll, and how the rollover percentage that the financial press reports every expiry week is actually useful.

Rollover is two trades dressed as one

Strip away the jargon and a rollover is the simplest thing in the Futures and Options (F&O) book — the part of the market where you trade contracts on a stock or index rather than the shares themselves.

Sell the contract you are about to lose. Buy the same direction, same underlying, in the next expiry. That is the whole act.

A useful way to picture it: rolling a futures position is like renewing a one-month rental agreement. You are not buying a new house — you are extending your stay. The new month just comes with a fresh agreement and new rent, and that rent is the rollover spread plus charges.

So if you are long one lot of Reliance May future, rolling means selling that May future and buying one lot of Reliance June future. The May leg books your near-month profit or loss; the June leg opens a fresh position with a new reference price and a fresh expiry to live with.

If you are short the May future, the legs flip. You buy back the May to close the short, and you sell the June to re-establish the same bearish bet for another month.

What makes the operation feel like one thing rather than two is the timing. The two legs are executed within seconds of each other, often through a single broker order, so the trader never spends a real minute uncovered. The exchange treats them as separate transactions for tax and brokerage, but the trader treats them as one decision.

The underlying never moves. The lot size never changes, and the direction is identical.

The only thing that changes is which row in the NSE F&O symbol list the position now sits on.

How the two-leg roll trades on the NSE

The NSE F&O segment lists three running expiries at any time: the near month, the next month, and the far month — the contract expiring soon, the one after that, and the one after that. On most retail terminals a Reliance futures search returns RELIANCE25MAYFUT, RELIANCE25JUNFUT and RELIANCE25JULFUT side by side.

Here is what actually happens on the day a position gets rolled.

1. Decide the leg sequence. A long roll sells the near month and buys the next month; a short roll buys back the near month and sells the next month. The directions are mirror images.
2. Pick the order type. Two market orders fired in sequence will get the job done, but you take on the risk that one fills and the other does not, or that the price moves between the two legs. The cleaner route is a calendar spread or pair order, which many Indian brokers expose under names like "Spread", "Pair" or "Rollover" on the F&O segment.
3. Submit the roll as one instruction. A spread order carries both legs together, priced as a single spread, so both go to market at once instead of one at a time. This lowers the risk of one leg filling while the other slips — but the exact mechanics, and even whether a native spread order exists, vary by broker, so check what your platform actually supports.
4. Confirm the contract notes. Two separate trades show up on the contract note, one in the near month and one in the next, each with its own brokerage, Securities Transaction Tax (STT) and exchange transaction charges. The cash impact of the roll is the spread between the two prices, multiplied by the lot size, plus those statutory charges.
5. Let the new position carry its own margin and MTM. The near-month margin is released after that leg closes, and the next-month margin is blocked when the new leg opens. From the next session, the daily mark-to-market (MTM) profit-or-loss settlement runs against the next-month contract's price, on its own expiry calendar.

The one part beginners stumble on is that the legs flip depending on which way the position points. This is the same roll, side by side, for a long and a short.

None of this requires the position to be flat in between. Done right, the trader is long Reliance from the moment the order is filled, every minute, all the way through the roll.

Rolling one Reliance future, the rupee math

Take a worked example. A trader is long one lot of Reliance May future, bought at ₹2,800, with a lot size of 500.

Expiry Tuesday is two days away. The cash share is trading at ₹2,840, and the trader still thinks the rally has more in it — so the plan is to roll, not to close.

One session before expiry, the trader rolls. The screen quotes look like this: the May future trades at ₹2,845; the June future trades at ₹2,860 — a spread of ₹15.

The May leg books the gain on the original trade. Forty-five points multiplied by a lot size of 500 is ₹22,500 of realised profit, before charges.

The June leg opens the new position at ₹2,860. Buying the next month at ₹15 above the May price is the cost of carry priced into the curve. That ₹15 spread on 500 shares is ₹7,500 paid out of the May gain to stay in the trade.

Two contract notes mean two sets of charges. Each leg carries brokerage, Securities Transaction Tax (STT), exchange transaction charges, GST, stamp duty and the SEBI turnover fee — so a roll pays these on both the closing trade and the opening trade.

There is no single fixed number to quote here, because the statutory rates change with almost every Union Budget — the STT on a futures sale, for instance, was raised again with effect from 1 April 2026. Treat the roll cost as a formula — the spread times the lot size, plus the per-leg charges your broker lists — and read the exact figure off your broker's brokerage calculator before you trade.

Net effect on the ledger is roughly ₹14,350 of realised profit, plus an open June long entered at ₹2,860, carrying its own June margin and its own June MTM cycle from that evening onwards. The thesis lives, but about ₹8,000 of the May gain just paid the rent for the next month.

The arithmetic flips in backwardation. When the next-month future trades at a discount to the near month, which happens around heavy dividend dates and when the market is in a sharply bearish mood, the roll is a small credit instead of a debit. The trader is paid a few rupees per share to stay in.

When to roll, and when to let the contract expire

Most retail rollover decisions get answered by one question. Is the original reason for the trade still on the chart?

If the answer is yes, the roll is the default. If the answer is no, the contract expires and the position is gone. Rolling a tired thesis just to "give it more time" turns a clean exit into a paid extension of a losing view.

The timing of a yes roll matters almost as much as the answer. The classical Indian-market practice is to roll in the last two or three sessions before expiry, with the heaviest volume clustering in the final sessions before the contract's last day. Two reasons.

First, near-month time decay on the small premium that still hangs on the May future bleeds quickly in the final week. Rolling a week early gives that decay away to whoever is on the other side.

Second, the spread book is densest when the institutional flow is heaviest. Spreads are tightest in the last couple of sessions before expiry, because every fund manager who needs to roll is in the same window, and the calendar-spread market makers are quoting their best prices.

Rolling on the expiry afternoon itself is a different game. Liquidity in the dying contract thins out after lunch, the spreads widen, and an impatient retail roll executed at 3:15 pm can pay several rupees of slippage that an earlier roll, a session or two before, would have avoided.

The same logic feeds back into the sizing decision covered in the guide on initial and exposure margin. A position that needs to be rolled three or four times to ride a longer-term trend is paying carry every month, and that carry has to be smaller than the move the thesis expects to capture.

Rollover percentage as a signal, and three mistakes

The number the financial press talks about during expiry week is the rollover percentage. It is the share of outstanding open interest — the contracts still open and not yet closed — in the dying contract that has migrated to the next month. A Nifty future running at a 78 per cent rollover by the afternoon before expiry means roughly three out of every four open positions in the May contract have been carried forward to June.

One thing worth being precise about: the NSE itself does not publish a "rollover percentage". What it publishes is contract-wise price, volume and open-interest data. Brokers and analytics platforms take the near-month, next-month and far-month open interest from that data and compute the percentage, then put it out in their rollover trackers around expiry.

The signal sits in the comparison, not in the absolute number. A reading well above the three-month average tells you institutions are committed to carrying the view; a reading well below it tells you they are letting positions die instead of paying carry.

Either way, the rollover number is a vote on the next month, cast by people with real money on the line.

Three mistakes recur in retail rollover behaviour often enough to be worth naming.

Mistake one is rolling a losing trade for the wrong reason. The position is down, the thesis has weakened, and instead of taking the loss the trader rolls to "give it another month." The new month starts with the same broken thesis plus a fresh spread cost on top. A losing trade does not improve by being moved to a different contract.

Mistake two is firing two market orders instead of a spread. One leg fills cleanly. The price moves while the trader is clicking. The second leg fills four or five rupees worse than expected.

On a single lot of Reliance, that slippage is ₹2,500. On a Nifty lot it is ₹250. On ten lots it adds up to the cost of a small course.

Mistake three is ignoring the tax clock. Each roll books a realised gain or loss on the closing leg of the near-month contract. For tax purposes that is a finished trade, separate from the new position. A trader rolling four times across a financial year has eight reportable transactions and a quarterly advance-tax obligation that does not exist for someone who simply holds a delivery share for the same period.

None of these are about getting the market direction wrong. They are about the small operational decisions that turn a clean trend trade into a slow leak of slippage, spread and tax friction.