What is leverage in trading?

Strip away the product names and leverage is one fraction: exposure divided by capital. Capital is the money that is genuinely yours and that absorbs the outcome. Exposure is the full size of the position that money is allowed to control once a broker, an exchange or the structure of a contract lets you post only a fraction of the value. When ₹1,00,000 of your own money carries a ₹5,00,000 position, the ratio is five to one, written 5x. The gap between the two numbers is borrowed, whether the borrowing is an explicit cash loan, a margin deposit against a futures contract, or the small premium that a bought option costs relative to the value it references. In every case you are standing behind a position larger than your money, and that is the entire idea.

The consequence follows immediately, because returns are generated by the exposure and then settled against the capital. If the asset moves by some percentage, the position gains or loses that percentage of the exposure, and that rupee amount is credited or debited to your capital alone. Divide it back out and the percentage change in your account is the market's move multiplied by the leverage. A 1 percent move at 5x is a 5 percent move on your equity; a 2 percent move is 10 percent; a 4 percent move is 20 percent. The market did nothing unusual in any of these cases. The only thing that changed the size of the outcome was the denominator you chose, the sliver of real money standing under a much larger position.

Make it concrete with a single position. Suppose you commit ₹1,00,000 of your own money and the structure lets you control ₹5,00,000 of an asset, a leverage of five. The asset rises 3 percent, so the ₹5,00,000 position gains ₹15,000, and that ₹15,000 is credited to your ₹1,00,000: a 15 percent gain on your capital from a 3 percent move in the market. Reverse it, and a 3 percent fall costs the position ₹15,000, dropping your capital to ₹85,000, down 15 percent. A 3 percent day is nothing remarkable in either direction; the five times multiple is the entire difference between a 3 percent outcome and a 15 percent one. The same position also shows where a margin call is born, because the losses are debited from your posted capital first, so a run of ordinary red days can quietly exhaust the money you put up long before the asset itself has done anything dramatic. All figures illustrative.

One clarification prevents a common miscount. Leverage is measured on total exposure, not on the net of offsetting positions, because it is the gross size that generates the swings and pays the costs. A trader holding several geared positions at once can be carrying far more total exposure than any single ticket suggests, and the number that actually governs survival is the account level leverage, the sum of every exposure divided by the capital behind all of them. It is easy to feel modestly sized on each individual trade while being dangerously geared in aggregate, which is why the honest measure is always taken across the whole account at once, every open position included, rather than one trade at a time.

This is why leverage is best read not as a way to trade but as a dial on the size of the bet. Turning the dial does not make you more likely to be right, does not sharpen your entry, and does not improve the average outcome of your method by a single basis point. It takes whatever result the position was going to produce and scales it, in both directions, by the same factor. The figure below fixes a single, ordinary 2 percent move in the asset and shows what three settings of that dial do to your equity. Notice that within each setting the up move and the down move are exactly the same size: in percent, leverage is perfectly symmetric, and that symmetry is precisely the trap the rest of this guide takes apart.

Because the multiplier is linear, a table captures the whole of it. Fix a set of ordinary moves in the asset, the kind that pass without comment on any normal day, and read across what each does to your equity at 1x, 3x and 5x. The pattern is mechanical and, crucially, the same in both directions: a good day is multiplied exactly as much as a bad day. There is nothing hidden in the up column that is missing from the down column. If leverage did only this, it would be a neutral amplifier, as fair as a magnifying glass, and much of its marketing quietly relies on you stopping your analysis right here.

Two things in that last row deserve to be sat with. First, the move that takes your account to zero is simply 100 divided by the leverage: 100 percent at 1x, about 33 percent at 3x, 20 percent at 5x. Leverage does not merely enlarge the daily swing; it pulls the point of total loss in from the far horizon to somewhere an ordinary market can actually reach. A 20 percent fall in a single stock is a bad session, not a black swan, yet at 5x it is the whole account. Second, every cost obeys the same multiplication as the returns. Brokerage, statutory charges and slippage are levied on turnover, which means on the exposure, not on your capital. A round trip that costs a tenth of a percent of turnover is a rounding error at 1x and a full half percent of your capital at 5x, and it is charged whether the trade wins or loses. The levered account therefore starts every position further behind, and the deficit compounds with how often you trade.

The frequency point is worth dwelling on, because it is where the friendly mirror first cracks. A trader who takes fifty round trips a year at a tenth of a percent of turnover each pays about 5 percent of capital in costs at 1x, an annoyance. The identical activity at 5x pays about a quarter of the capital in costs over the same year, before a single trade has been judged right or wrong, because each round trip now bites at half a percent of capital rather than a tenth. Leverage does not merely enlarge the good and bad outcomes symmetrically; it silently enlarges the guaranteed, direction-independent drag of simply participating. The up column and the down column are mirror images, but the cost column, which sits underneath both, is not neutral: it is always negative, and leverage multiplies it in lockstep with everything else.

In percent, leverage is a fair mirror. In survival, the mirror is a cliff, and the multiplier is what walks you to the edge of it.

Here the symmetry breaks, and it breaks for a reason that has nothing to do with leverage and everything to do with arithmetic. A loss and the gain that undoes it are not mirror images, because each is measured against a different base. Lose 20 percent and you have ₹80,000 left, and to get back to ₹1,00,000 that smaller pile must grow by 25 percent, not 20. Lose half and the ₹50,000 that remains must double, a 100 percent gain, merely to return to where you began. Lose three quarters and you need a 300 percent gain. The rule is exact: the gain required equals the loss divided by one minus the loss, and it accelerates without limit. As the loss approaches the whole account, the climb needed to recover runs away towards infinity, because there is almost nothing left to compound from. Lose 90 percent and the survivor must make 900 percent simply to return to par, a demand no ordinary run of trades can satisfy, which is why a loss of that depth is rarely a drawdown and almost always an ending.

This curve is the true face of leverage, and the hero figure below draws it. The dashed line is the intuition almost everyone carries, that a loss needs a gain of the same size to repair it. The solid curve is the arithmetic, and it pulls away from the intuition and then bends nearly vertical. What leverage does is decide where on this curve an ordinary market move lands you. Follow a single, unremarkable 20 percent fall in the asset across three leverage settings. At 1x it is a 20 percent drawdown, a shallow spot near the flat left end, needing a 25 percent gain to mend. At 3x the same fall is a 60 percent drawdown, far up the steepening slope, needing 150 percent. At 5x that identical 20 percent fall is the entire account, the vertical wall on the right where no recovery trade exists at all. One market event, three sizes, and only the size decided whether it was a scratch, a wound or a death.

Notice what is quietly cruel about this shape. The multiplier is felt every single day, because the account jumps around by several times the market and the trader sees it and grows used to it. The asymmetry, by contrast, stays invisible right up until the moment it is not. For every small and medium move the up and the down really do look symmetric on the screen, exactly as the first figure showed, so a beginner reasonably extrapolates that the whole relationship is symmetric. The cliff at the right of the recovery curve never shows up in ordinary experience, because ordinary experience never reaches a loss deep enough to reveal it. It is felt for the first time on the day it is fallen off, which is the worst possible moment to discover that percent and survival were never the same measurement.

The recovery curve, read together with the multiplier, delivers the single most important idea about leverage: it drags the point of no return in from a distance you would never reach towards a distance an ordinary week reaches routinely. At 1x, being wiped out requires the asset to fall the whole 100 percent, which for a diversified holding essentially never happens in one stroke. At 5x, the same fate needs only a 20 percent fall, and at 10x only 10 percent. You have not changed the market, its volatility, or your skill. You have moved the edge of the cliff closer to your feet, and left the size of your ordinary steps unchanged.

The table lays the two effects side by side. Read down the loss column as the drawdowns a levered account routinely visits, and across to see both the gain each demands and the adverse move in the asset that produces it at 1x and at 5x. The right two columns are the punchline: at 1x a 20 percent drawdown needs a 20 percent fall in the asset to create it, but at 5x that same 20 percent drawdown needs a fall of only 4 percent. Leverage compresses the distance between a normal day and a serious wound, and it keeps compressing it right up to the wall.

This is why professional risk frameworks treat position limits as non-negotiable even when conviction is total, and especially then. Conviction has no bearing whatever on where the point of no return sits; only size does. A desk that caps the loss any single position can inflict is not expressing doubt about the trade, it is refusing to let any one trade reach the steep end of the recovery curve, because it has already priced what that end costs. The retail instinct runs the opposite way: the stronger the conviction, the larger the size, which is exactly the pairing that turns a good idea and a bad week into a closed account. A size limit is the humility the arithmetic demands, installed in advance precisely so that a strong feeling cannot remove it in the moment it most wants to.

All of this would still be tolerable if leverage bought something in exchange for the risk it adds. It does not. Leverage is a multiplier applied to whatever edge a system already has, and a multiplier changes the size of a number without changing its sign. Multiply a genuine positive edge and you get a bigger, more violent version of a winning method, one that also visits the steep end of the recovery curve far more often. Multiply a break-even or negative edge and you get to ruin faster, because the bet is larger relative to the capital and the costs are larger too. The expectancy per unit of exposure is exactly what it was; only the variance, the drawdowns and the probability of never recovering have grown.

It helps to write the edge down. Expectancy per trade is the win rate multiplied by the average win, minus the loss rate multiplied by the average loss, all measured per unit of exposure. Leverage scales the exposure, so it scales both the average win and the average loss by the same factor, and the expectancy per unit is left precisely where it started. What moves is the spread of outcomes around that unchanged average, which widens in direct proportion to the leverage, and the chance that the spread carries the account into a hole from which the recovery arithmetic offers no way back. A larger bet does not tilt the coin. It only raises the stakes on each flip, and with them the odds that an ordinary streak of tails ends the game before the favourable average has had time to assert itself.

The figure below makes the point with a single computed example, and it is deliberately honest about it. One asset, with a genuine edge, grinds upward across the window. Three traders ride the identical price path at 1x, 3x and 5x, each re-sized to that multiple. The 1x rider ends up about 14 percent, the edge doing its slow work. Then one ordinary crash day arrives, a 20 percent fall of the kind that a circuit or a gap produces a few times in any trading life. The 1x rider takes the hit and recovers, because a 20 percent drawdown is survivable. The 3x rider is thrown into a 60 percent hole and claws back only to roughly flat, punished by the very drag and asymmetry the earlier figures described. The 5x rider does not get a second act: a 20 percent fall at 5x is a 100 percent loss, so the account is liquidated on that single day and takes no further part in the recovery it helped pay for. Same edge, same market, three fates, and the only variable was the size of the bet.

The deeper reason sits in how an edge actually pays out. An edge is not a promise about the next trade; it is an average that only reveals itself across a large number of trades. A method with a genuine advantage still loses on plenty of individual trades and strings several losses together on a regular basis, and that is not the edge breaking, it is the edge behaving exactly as a probabilistic thing behaves. The account curve only drifts reliably upward once enough trades have accumulated for the average to dominate the noise, and the more variable each trade is, the larger the sample you need before the mean shows through. This is ordinary statistics, and it is quietly ruthless.

Leverage attacks that process from both ends at once. It inflates the variance of every trade, which lengthens the sample you need for the edge to appear at all. And it sharpens an absorbing barrier, ruin, that the plain law of large numbers never accounts for. A coin with a small positive expectancy drifts upward over thousands of flips, but only if you are allowed to keep flipping. Bet too large a fraction on each flip and a normal losing streak takes the bankroll to zero first, at which point the favourable long-run average is meaningless, because you are no longer in the game to collect it. Leverage is precisely the knob that raises the fraction staked on each trade, so it is precisely the knob that can end the sequence before the edge has had its say.

There is a second cost that scales with leverage, quieter than the drawdown and easier to ignore because it is charged in fair weather. Borrowed exposure has a running price. In the cash market that price is explicit interest on the funded rupees. In derivatives it wears a different coat, the time decay of an option or the roll cost of a future, but it is the same kind of drain: a toll you pay for holding a position larger than your money, whether or not the position works. And because you are carrying the toll on the exposure while it eats your capital, the drag scales with the leverage just as the returns do.

The figure fixes an illustrative carry of 12 percent a year on the borrowed portion and shows what that becomes as a fraction of your own capital at each setting. At 2x you are borrowing an amount equal to your capital, so the annual carry is about 12 percent of your money. At 5x you are borrowing four times your capital, so the same rate is roughly 48 percent of your money every year, a hurdle the asset must clear before you have earned a single rupee. Expressed on the position, the asset must rise by close to 10 percent a year at 5x merely to stand still after carry. This is why a levered position that is merely flat is quietly losing, and why holding a geared trade patiently is not the free waiting that holding an unlevered share can be.

There is a third drain that compounds the interest and the decay, and it is the strangest of the three because it appears even when the market ends exactly where it started. If you re-size to a constant multiple of your equity, which is what a trader does who sets each morning's position to the maximum the margin allows, a fall of some size followed by a rise of the same size does not return you to par, because the rise acts on a smaller base. The shortfall grows with the square of the leverage. A 5 percent fall then a 5 percent rise costs an unlevered account about a quarter of a percent; the same pair re-levered to 5x becomes a 25 percent swing each way and costs roughly 6.25 percent per round trip, about twenty-five times the damage. A choppy, directionless market that an unlevered holder barely registers therefore grinds a re-levered account steadily lower. Leverage does not only tax the trend. It taxes the noise, and it taxes it by the square.

Stack the three together and the running cost of carrying leverage is larger than any one of them suggests on its own. Interest or roll takes its cut on the calendar, time decay takes its cut as expiry approaches, and volatility drag takes its cut from the chop in between. All three scale with the size of the borrowed exposure, and all three are charged regardless of whether the underlying idea turns out to be correct. A levered position is therefore not merely a bigger version of an unlevered one; it is a bigger version that is also paying rent on the part it borrowed, every single day it is held. The edge has to clear that rent before it clears anything at all, which is one more reason the same edge does less for a levered account than the naive multiplier promises.

Leverage appears across the Indian market in what look like unrelated products, and each has its own page on this site with its own mechanics. The cash-market loan against your shares is the margin trading facility, examined in our guide to margin trading in India. A derivatives contract bundles a fixed quantity of the underlying into a single unit, so the F&O lot size already carries gearing before you post anything extra. A bought option folds convex leverage into a small premium. Intraday margin products, including the cover order that pairs the leverage with a compulsory stop, let a smaller deposit control a larger position for the session. The temptation is to learn each as a separate subject. The more useful move is to see that they are one subject in four costumes.

Underneath all of them sits the same fraction, exposure divided by capital, and therefore the same three consequences. Every form multiplies the outcome by its gearing. Every form levies a toll, whether that toll is named interest, time decay or roll. And no form manufactures edge; each simply reaches whatever destination the underlying method was headed for, sooner and with larger swings. The table sets each form beside where its borrowing lives, the toll it charges and the shape of its worst case, and the schematic that follows draws the convergence: four inlets, one law, two outputs. When you next meet a leverage product you have not seen before, you do not need a new mental model for it. You need to find its gearing, find its toll, and remember that it cannot add an edge you did not bring.

One distinction inside that table matters more than the others, because it decides how the worst case behaves. A bought option is the gentle end of the range: its gearing is convex and often extreme, yet the loss is capped at the premium paid, so it can never cost more than the money put in, however violently the underlying moves. The written, or short, option is the dangerous mirror of the same instrument. It is margined like a future, it collects a small premium in exchange for a large and open-ended obligation, and its loss is not capped at all. The two sit at opposite ends of one product, and a trader who has learned that options are the safe, capped form of leverage from the buyer's side can walk into the seller's side and meet exactly the uncapped exposure this whole guide warns about. Read the direction of the trade, not just the name of the instrument, before you decide which law applies.

Because the law is shared, the study of leverage transfers in a way the study of any single product does not. Learn the mechanics of the margin trading facility in isolation and you have learned one product. Learn to read exposure over capital, to find the toll, and to expect no edge bonus, and you have learned to price every geared instrument you will ever meet, including ones that do not exist yet. The specific rules and the specific numbers belong on the specific product pages, and they will change with each revision the exchange or the regulator makes. The law on this page does not change, which is precisely why it is worth holding on to more tightly than any single margin percentage.

The general law is universal, but the amount of leverage a retail trader can actually reach in India has been falling for years, by deliberate regulatory design, and any account of the subject that ignores this is describing a market that no longer exists. The Securities and Exchange Board of India has run two long campaigns to reduce available leverage, and it has published the loss data that motivated them. Treat the specifics below as principle plus date, not as numbers to memorise, because the exact thresholds are revised periodically and the mechanics of each product belong on that product's own page.

The first campaign ended the era of very high broker-granted intraday leverage. Under the peak margin framework, introduced by a SEBI circular dated 20 July 2020 and phased to full upfront margin collection by September 2021, brokers must collect the required margin before the trade rather than extending a large position on a thin deposit and checking later. The practical effect is straightforward: intraday leverage in the cash segment is a small multiple today, not the double-digit figure some older material still quotes. The second campaign targeted index derivatives, where embedded leverage concentrates. SEBI's framework of 1 October 2024 raised minimum contract sizes and required option premium to be collected upfront from buyers, among other changes on a published timetable. Both campaigns rest on the same evidence, and it is stark: a SEBI study from September 2024 found that about 93 percent of individual traders in equity derivatives made net losses over FY22 to FY24, with aggregate net losses exceeding ₹1.8 lakh crore. Leverage did not choose those losing trades, but it set their size, and the derivatives segment where gearing is highest is exactly where the losses concentrate.

There is a lesson in the direction of travel itself, quite apart from any single threshold. When a regulator with full visibility into retail outcomes spends years methodically reducing the leverage available to individual traders, and publishes loss figures of this magnitude as its reason, it is making from data the same argument this guide makes from arithmetic: that the gearing was not helping the people using it. The rules do not assume traders lack an edge. They assume that whatever edge exists is fragile, and that multiplying a fragile thing by a large number mostly multiplies the damage. A trader does not have to wait for a rule to reach that conclusion privately, position by position, and choose to use less than the maximum on offer.

Put the pieces together and leverage stops looking like an opportunity and starts looking like what it is, a decision about the size of the bet that has been dressed up as a feature. It offers a bigger number in the same direction your method was already going, and in exchange it multiplies the toll, steepens every drawdown, pulls the point of no return towards you and does nothing whatever for your edge. None of this makes leverage forbidden. It makes leverage an output of a plan rather than an input to it, something you discover at the end of a sizing calculation rather than something you reach for at the start.

Run in that honest order, the arithmetic serves you. You begin from a fixed risk budget, the small fraction of capital a single loss is allowed to cost, derive the position from the distance to your stop, and only then read off the leverage that size happens to imply, which is almost always far below whatever the product permits. Sizing from a risk budget in this way is the discipline at the centre of risk management, and it quietly dissolves most of the dangers on this page, because an account that never visits the steep end of the recovery curve never needs the miracle gain to climb back. Deciding exposure last, after the risk budget and the stop, is exactly the ordering the method we teach is built around; the multiplication itself is the trivial part.

It is worth naming why leverage is so reliably sold as an opportunity rather than a size decision, because the framing is not an accident. A larger position produces larger absolute gains on the winning days, and those are the days that generate screenshots, testimonials and attention. The toll, the drag and the ruin arrive quietly, on the losing days and the flat ones, and they are borne in private. So the visible surface of leverage is almost all upside while the cost is almost all hidden, which is precisely the shape of a thing that will be marketed hard and understood late. The discipline argued for here is really just the refusal to accept that framing: to read the leverage multiple as a statement about how much of the account is exposed to a single bad sequence, rather than as a measure of how serious a trader you are.

A quick worked example shows how small the honest number usually turns out to be. Say the account is ₹5,00,000 and the plan risks 1 percent, or ₹5,000, on a trade whose stop sits 4 percent from the entry. The position that puts exactly ₹5,000 at risk over a 4 percent move is ₹1,25,000 of exposure, one quarter of the account, a leverage of 0.25 rather than five. Even doubling the risk budget and halving the stop distance lifts the implied leverage only to one. The number the product advertised, five or more, never entered the calculation at all, because size was derived from the loss the plan could tolerate rather than from the exposure the margin permitted. That inversion, risk budget first and exposure last, is the entire difference between using leverage and being used by it. All figures illustrative.

The professional question is never how much leverage is available. It is how little leverage completes the plan, so that a normal run of losses can never end the account.

• SEBI peak margin framework. Securities and Exchange Board of India circular dated 20 July 2020, introducing upfront margin collection verified through intraday snapshots and phased to full collection by September 2021, the measure that ended very high broker-granted intraday leverage. Cited at principle level; confirm current requirements at source. sebi.gov.in
• SEBI equity index derivatives framework. Securities and Exchange Board of India circular dated 1 October 2024, raising minimum contract sizes and requiring option premium to be collected upfront from buyers, among other changes on a published timetable. Cited at principle level. sebi.gov.in
• SEBI study of individual trader outcomes. Securities and Exchange Board of India study, September 2024: about 93 percent of individual traders in equity derivatives made net losses over FY22 to FY24, with aggregate net losses exceeding ₹1.8 lakh crore. sebi.gov.in
• SEBI margin trading facility framework. Securities and Exchange Board of India circular dated 13 June 2017, the regulated route to leveraged delivery buying against pledged Group I securities, with the product mechanics covered in our margin trading guide. sebi.gov.in
• The arithmetic of recovery, carry and ruin. The recovery-gap function, the carry drag and the equity paths in this guide are standard, self-contained calculations, computed directly from the definitions of leverage, compounding and the illustrative inputs stated on each figure.