What is rollover in futures trading?

A futures contract is dated. It has a defined expiry after which it is settled and ceases to exist, so unlike a share you cannot simply hold it forever. If your view runs longer than the current contract, expiry forces a binary choice: close the exposure, or move it into the next contract. Rolling is the second option, and it is purely mechanical, a consequence of trading a dated instrument rather than any kind of trading signal. Nothing about the market has told you to roll; the calendar has.

It is worth being precise about why this is forced. A share is a claim on a company that persists until you choose to sell it; a futures contract is a time-limited agreement to transact at a set price on a set date, and the exchange extinguishes it at expiry by settling every open position. There is no version of holding the same contract indefinitely, because the contract itself has a death date stamped on it. Index and stock futures in India are listed as a short series of monthly contracts, typically the near, next and far month, so a roll normally moves a position from the expiring near month into the next month. A trader used to equities can find this jarring: the position does not renew itself, no broker quietly extends it, and if you take no action the exchange will close or settle it on its own terms. The roll exists precisely to put that decision back in your hands, at the price of doing it deliberately, and paying for it, every single cycle.

To roll, you close the expiring contract, the near month, and open the identical position in the next month, so the market exposure continues without a break. The lot framework does not change across the roll: you carry the same number of lots into the new series. Crucially the two legs are done together. Selling the near and buying the far as separate orders leaves a gap, seconds or minutes, in which you are either flat or double-exposed, and price can move against you in that window. Placing the pair as one calendar spread, sell near and buy far in a single instruction, transfers the position cleanly and pins the trade to the spread you actually pay. The same close-and-reopen idea applies to an options position, though the arithmetic there also drags in the strike and the time value that decays into expiry, a difference the guide on futures versus options draws out.

This is why rollover clusters in the final sessions before expiry and is treated as routine position maintenance rather than an advanced manoeuvre. The schematic below traces a long being carried past expiry, and the table breaks the same roll into its five deliberate steps.

Because a roll is two trades, it is never free, and the cost is not a one-off. Every expiry you choose to stay in the market you cross the spread again and pay charges on two more legs. For a trader whose view spans a single afternoon this is irrelevant. For anyone carrying a position across months it is a recurring toll that quietly compounds, a bill that arrives every expiry whether the position has worked or not. The friction most people picture, the brokerage line on the contract note, is the smallest part of it. The real weight sits in the spread, hidden in the price rather than printed as a fee.

It helps to put real, if illustrative, numbers on a single roll. Take one long lot of a single-stock future worth about ₹8,00,000. At a net cost of carry near 7 percent a year, a one-month roll spread is roughly ₹9 to ₹10 a share, on the order of ₹4,800 for the lot. Securities transaction tax on the sale leg, at 0.05 percent of the sold value, adds about ₹400, and brokerage with exchange and other statutory charges on the two legs adds a modest amount more. The figure below itemises that one roll and then stacks three of them, a full quarter of holding, so the compounding is visible rather than assumed.

The one piece of good news is that the roll is cheap to execute well, because the calendar spread is not merely two orders you happen to send together. On Indian exchanges it is a tradable instrument in its own right, with its own quoted price, the spread itself, so the exchange can match your sell-near-and-buy-far as a single fill at one spread price rather than leaving you to leg into it. That matters because the danger in a roll is almost entirely leg risk: the chance that price lurches in the seconds between closing the near and opening the far. Trading the spread as one instrument removes that gap, and because the roll window concentrates enormous two-sided volume into exactly these spreads, it usually offers the tightest pricing of the month. The catch is not any single roll, which is small in fees and clean in execution. It is the repetition, expiry after expiry, that turns a modest, well-executed cost into a material drag on a position held for the long run.

A futures position is not a thing you own. It is a thing you keep renewing, and the exchange charges you at every renewal.

Where does the spread come from? The near-month and next-month futures almost never trade at the same price, and the gap is not arbitrary. A futures price is tethered to the spot price of the underlying by the cost of carry, the net cost of financing the exposure until the contract expires. The standard relation is F = S e^((r − d) t): F is the fair futures price, S the spot price, r the financing rate to expiry, d the dividend yield the underlying pays before expiry, and t the time remaining.

Read plainly, the far contract has a longer t, so for a positive net carry it prices a little higher than the near contract. That is why, in a normal market, rolling a long means selling the cheaper near contract and buying the pricier far one, and the difference is a real cost you pay. The gap between futures and spot is the basis, and it decays to zero at expiry, because a contract with no time left cannot carry anything: futures and spot must meet. That convergence is why the roll spread is widest early in a contract's life and thin right at expiry, and it is also why each roll matters. Rolling does not just pay a spread once; it re-opens a fresh, full-length basis in the far contract that you then pay down again over the next month, only to roll and re-open it once more. The recurrence in the previous figure is this convergence-and-reset, made into rupees.

It helps to be concrete about the two rates. The financing rate r is the cost of the money you would otherwise tie up buying the underlying outright: think of it as the short-term interest you pay or forgo to hold the exposure on margin instead of in full. The dividend yield d is what the underlying pays out before the far expiry, which a futures holder does not receive, so it is netted off the financing cost. For an index future the dividend term is a blended yield across many constituents and is usually small, which is why index futures tend to sit in gentle contango. For a single-stock future the dividend term is lumpy: it is close to zero for most of the year and then, in the weeks around a large dividend, it can spike and briefly overwhelm the financing term. That is why the sign of a stock-future roll spread is far more variable than an index one, and why the spread is something you read live off the two contract prices rather than a constant you can memorise.

The roll spread is not always a cost to a long. Its sign depends on whether financing or dividends dominate the cost of carry, and that gives the two states the market has names for. When the financing rate r is larger than the dividend yield d, the net carry is positive and the far contract sits above the near: this is contango, and a long pays the spread to roll forward. When a large dividend is due before the far expiry, the yield term d can dominate and push the far contract below the near: this is backwardation, and a long is effectively paid the spread to roll. The same formula produces both; only the balance of financing against dividend flips which way the spread points.

There is a practical refinement worth carrying into the roll window: annualise the spread before you judge it. Because the spread reflects carry over the specific gap between the two expiries, it is most comparable once you scale it to a yearly rate. A spread of about ₹9 to ₹10 a share on a ₹1,600 stock over one month works out to roughly 7 percent annualised, which you can weigh against the prevailing short-term financing rate to decide whether this roll looks rich or cheap. When the implied carry sits well above financing, the far contract is expensive and a long is paying up to renew; when it sits below, or turns negative around a dividend, the roll is unusually favourable. None of this forecasts price. It answers only the one question the spread can honestly answer, whether this month's renewal is dear or cheap relative to the cost of money.

The practical lesson is to read the two contract prices rather than assume. Most of the time, and for index futures in particular, the market sits in gentle contango and the long pays a modest spread to renew. But the sign can flip, especially in single-stock futures around a dividend, and a trader who blindly assumes rolling always costs money will misjudge both the price and the timing of the roll. The spread is information, and it is free to read on the order book of the two contracts you are about to trade.

In India, stock and index futures run on a monthly expiry cycle, and the roll window is the last few sessions before that monthly settlement, when roll liquidity and rollover commentary both peak. That timing is the practical heart of the roll, and the mechanics of expiry day itself are set out in the guide on expiry-day trading. What matters for rollover is that a dated contract forces a decision, and that the nature of the decision depends on how the contract settles.

Settlement type is not uniform. Index futures are cash-settled: at expiry the position is closed against the index level and no delivery changes hands. Single-stock futures are physically settled: since the phased transition completed at the October 2019 expiry, an expiring single-stock futures position is settled by actual delivery of the shares rather than in cash, as of 17 July 2026. That distinction turns expiry from an accounting event into a delivery obligation, and it is a concrete reason a stock-future holder must act before the last day. Let a single-stock long simply lapse and you are on the hook to take delivery of the full underlying value, not merely settle a small cash difference. The figure lays out the three ways an expiring single-stock long can end.

The concentration of activity into the roll window is not incidental; it is what makes the roll cheap to execute. As expiry nears, hedgers, arbitrageurs and directional traders all move their positions in the same few sessions, and that crowd is what deepens the spread market and tightens its pricing. Roll too early and you may pay a wider spread into thinner liquidity; roll at the very last moment and you risk the near contract settling before you have acted, which for a single-stock future means the delivery obligation rather than a clean transfer. Knowing the exact expiry date, then, is not pedantry. It defines the window in which the roll is both cheapest and safest, and it is the one date around which the whole exercise is organised, which is exactly why an out-of-date calendar is more than a cosmetic error.

Around every expiry, analysts publish a rollover percentage for index and single-stock futures. It answers one narrow question: of the positioning that was open in the expiring contract, how much was carried into the next series rather than closed? Because it is built on open interest, the count of contracts still open, it is a measure of quantity carried, nothing more and nothing less. A common construction is

The single figure is close to meaningless in isolation; the comparison is where the information lives. A rollover reading is described as heavy when it sits above the recent average and light when it sits below, and it is that deviation, not the absolute percentage, that commentary leans on. Heavy rollover is generally read as participants choosing to carry the existing bias into the next series rather than close out at expiry, which is loosely framed as conviction. Light rollover is read as more participants stepping aside at expiry. Alongside the percentage, the roll spread adds a second dimension: whether the spread is unusually wide or narrow hints at how eagerly one side is paying up to carry.

Two levels of the statistic are worth separating. A single-name rollover percentage describes how much of one contract's open interest moved forward, read against that name's own history. A market-wide rollover, the figure the financial press tends to headline around expiry, aggregates the roll across the index or across a basket of large single stocks, and it is read against the market's recent average as a broad gauge of how much positioning the market as a whole chose to carry. When commentary says rollover was heavy this expiry, it usually means this aggregate sat above its three-month average. That aggregate is a genuine sentiment tell in the narrow sense that it records collective behaviour at a decision point, which is why it is reported at all. It remains, however, a measure of quantity carried, and the next section is about the specific thing it still cannot tell you, however wide the basket.

Here is the part that separates a working understanding of rollover data from the headline version. The rollover percentage measures how much positioning was carried forward. It says nothing about which way that positioning is pointed. Open interest, and therefore the rollover built on it, nets longs and shorts together: every contract has a buyer and a seller. A high rollover number can be produced by committed longs carrying a bullish bet, by committed shorts carrying a bearish one, by hedgers rolling protection that has nothing to do with a directional view, or by any mixture of the three. The statistic cannot distinguish between them, so high rollover is not, by itself, bullish.

This is why "high rollover means the smart money is long, so buy" is a weak heuristic dressed as a rule. It smuggles in an assumption, that the carried positions are net long, that the number itself does not contain. Rollover is also a coincident and descriptive figure: it records what participants did as expiry arrived, not what price will do next. It lags the decisions it summarises and it looks backwards. Treated as a standalone trigger it is close to noise; treated as one input among several, it is genuinely useful context. The honest way to read it is layered: the percentage against its average tells you how much was carried, the roll spread hints at how hard one side is paying to carry, and only price and structure, read separately, speak to direction. Distinguishing a descriptive statistic from a predictive one, and refusing to let the first masquerade as the second, is exactly the kind of judgement the method we teach is built around.

A concrete example makes the blind spot impossible to miss. Suppose two expiries print an identical, above-average rollover. In the first, the carried open interest is dominated by longs who rode a rally and want to stay with it into the next series. In the second, it is dominated by shorts who are pressing a decline and want to keep the bet on. The rollover percentage is the same in both, yet the positioning behind it points in opposite directions, and price could do anything next in either case. No amount of staring at the single number resolves which world you are in, because the direction was never encoded in it to begin with. This is why serious desks pair rollover with data that does carry a directional lean, and never let the roll figure stand in for a view on its own.

The reason rollover nets to a directionless number becomes concrete once you notice how differently two participants use the very same roll. A hedger holds futures against an underlying exposure, a portfolio, an inventory, a delivery commitment, and rolls the hedge mechanically because the thing being protected still exists. For them the roll toll is an insurance premium: an accepted, budgeted cost of keeping protection in place, and the direction of their futures leg is deliberately opposite to a position they hold elsewhere. A speculator holds futures as the bet itself, so for them the roll toll comes straight out of the edge, and every renewal has to be justified by a view that still holds. Same instrument, same roll, opposite purpose.

Both of these show up in the same open interest and the same rollover percentage, indistinguishable from the outside. A heavy roll might be hedgers renewing protection into an uncertain quarter, which says nothing bullish or bearish; it might be speculators pressing a one-sided bet; it is almost always a blend of the two. This is the deeper reason the number cannot carry direction: it aggregates flows whose intentions are not merely unknown but actively opposed to one another. For your own trading, the useful version of this split is internal. Be honest, on every roll, about which of the two you are, because a hedge roll and a speculative roll deserve completely different tests before you agree to pay the toll again.

A concrete version makes the point. A long-term investor holding a basket of large-cap shares might short index futures to protect against a broad fall over an uncertain few months, and roll that short at each expiry until the risk has passed. The roll toll there is simply the running cost of the insurance, and paying it is rational for as long as the underlying holding, and the danger to it, both still exist. Contrast a trader who is short the same index futures purely as a bet that the market will drop: for them the identical roll toll is pure drag on a directional wager that has to keep being right to justify the renewal. The published rollover shows one number for both, and cannot tell the protective short from the speculative one.

Pull the threads together and the thesis is simple. A futures position is not a durable thing you own; it is a lease on exposure that expires on a fixed day, and to keep it you must renew it, at a price, again and again. Because renewing is easy and routine, it is easy to do for the wrong reason. Each roll carries a real, recurring cost, the roll spread plus two legs of transaction costs and slippage, so carrying a position is never a neutral act. The classic trap is rolling a losing position forward purely to avoid booking the loss. That converts an admission that the trade is wrong into a monthly bill, keeps leveraged capital pinned to a view that has not worked, and lets the roll costs accumulate on top of the loss.

The leverage is the reason to be careful. Futures sit in the equity derivatives segment SEBI has repeatedly flagged for retail losses: in its September 2024 study of individual traders in that segment, SEBI found that about 93% of individual traders in equity derivatives made net losses over FY22 to FY24, with aggregate net losses exceeding ₹1.8 lakh crore. Against that backdrop, the discipline the renewal ledger asks for is not optional. Rollover is a mechanical tool for managing the timing of a position across a dated-contract boundary. It is not a way to rescue a bad trade, and it is not a source of edge. Read this way, rollover data and the roll itself are context and plumbing, useful, unavoidable, and no substitute for the analysis that decides whether the exposure is worth carrying, and paying to renew, at all.

• Cost-of-carry and futures pricing. The forward-price relation, F equals S times e to the power of the net carry rate times time, with the net rate the financing rate less the dividend yield, and the convergence of futures to spot at expiry, are the standard results that set the roll spread and its sign in contango or backwardation.
• Securities transaction tax on futures. As of 17 July 2026, the rate on the sale of a futures contract is 0.05 percent, following the Finance Act 2026 (Act No. 4 of 2026) with effect from 1 April 2026. Verify the current rate against the enacted Finance Act and your contract note. indiabudget.gov.in
• Physical settlement of single-stock derivatives. Single-stock futures and options in India are settled by physical delivery of shares, following a phased transition that applied to all such contracts from the October 2019 expiry; index derivatives remain cash-settled. Confirm the current settlement rules for the specific contract with the exchange. nseindia.com
• SEBI study on derivatives-segment losses. SEBI's September 2024 study of individual traders in the equity derivatives segment reported that about 93 percent of individual traders made net losses over FY22 to FY24, with aggregate net losses exceeding 1.8 lakh crore rupees, the context for treating leveraged futures positions cautiously. sebi.gov.in
• Rollover percentage definition. The rollover percentage as open interest carried to the next series divided by the total open interest in the expiring series, read against a recent average, reflects the standard construction published by Indian exchanges and analysts around each monthly expiry.