Straddle or Strangle: You Are Not Buying Direction, You Are Buying Movement

The premium is the hurdle, not the price of admission

Start with what the two structures are, because the mechanical description is short and the interesting part comes afterwards. A long straddle is a bought call and a bought put at the same strike, conventionally the strike nearest the spot. A long strangle is a bought call and a bought put at different strikes, both out of the money, the call above the spot and the put below it. In each case both legs are bought, so the whole cost is paid up front and the whole cost is the maximum that can be lost. Neither structure has a view on direction, and neither can be made to have one.

What they do have a view on is size. Because the call profits when the underlying rises and the put profits when it falls, the combined position pays whenever the move is large, regardless of sign. But the two legs were not free, and the premium paid has to be recovered out of whatever the move produces before a single point of profit exists. That recovery threshold is the whole trade. Everything else, the Greeks, the strike selection, the choice of expiry, is machinery around the question of whether the underlying will cover the distance.

The distance is not arbitrary either, and this is where the position stops being a simple bet on a busy market. The premium is set by people who are also estimating how far the underlying will move, and their estimate is encoded in implied volatility. So the buyer is not betting that the index will move. The buyer is betting that it will move further than the price already assumes. Those are different propositions, and only the second one is on the table. Our guide to reading an option chain works through how that estimate is read off the screen; the concern here is what it implies for a position, once you turn it into a percentage.

Framing it as a percentage is the single change that makes these structures legible. A premium of 743 points sounds like a number you would have to look up. The statement that the index must travel 2.97 percent in 30 days is a claim about the world that you can immediately test against how the index has behaved. The first form invites you to think about whether the premium seems expensive. The second form invites you to think about whether the move is likely. Only the second question has an answer.

This page sits underneath the site's framework for deciding which structure fits a view, which classifies these two together as a single family with a defined loss and a pure volatility view. That classification is correct and it is not repeated here. What follows instead is the arithmetic underneath the classification: what the two structures actually cost, what they actually require, where they differ from one another, and how a realistic distribution of moves compares to the requirement.

The setup, stated in full so it can be checked

Every number on this page comes from one deterministic model, seeded so it reproduces. The underlying is an illustrative index rather than a live instrument, deliberately: a stated model keeps the arithmetic reproducible and avoids implying that any particular Indian index would have produced any particular outcome. Options are priced with a standard lognormal model at a zero interest rate, which keeps the forward equal to the spot and makes the at the money structure exactly symmetric. A non zero rate shifts the strikes marginally and changes nothing about the argument.

Two choices in that list deserve a note. The strangle strikes sit 400 points out, which is 1.6 percent of the index and corresponds to roughly a 34 delta call and a 33 delta put on these inputs. That is a conventional width rather than an extreme one, and the comparison is repeated at seven other widths later on the page so that nothing rests on the particular choice. The synthetic series carries an annualised volatility of 12.0 percent against an implied volatility of 13 percent, which is the ordinary relationship between the two in an option market. The gap between them is not an error in the model. It is the thing the buyer is paying, and putting it in the setup rather than discovering it later is the honest way to run the exercise.

Both payoffs, computed

With those inputs the straddle costs 371.7 points for the call and the same again for the put, a total of 743.4 points, which is 2.97 percent of the index. The strangle costs 208.1 for the 25,400 call and 202.7 for the 24,600 put, a total of 410.8 points, or 1.64 percent. The strangle is 44.7 percent cheaper. Held to expiry, the straddle breaks even at 24,257 and 25,743; the strangle breaks even at 24,189 and 25,811.

Read the breakevens as distances rather than levels and the comparison sharpens. The straddle requires 2.97 percent in either direction. The strangle requires 3.24 percent, because although it costs 332.5 points less, its profitable zone starts 400 points further from the spot on each side, and the net of those two effects is 67.5 points of extra distance. The losing band is 1,487 points wide for the straddle and 1,622 points wide for the strangle: the cheaper structure has the wider region in which it makes nothing at all.

The shapes differ in a way the numbers alone do not convey. The straddle is a V with its point at the strike, so its worst outcome occurs at exactly one index level and every step away from that level improves the result immediately. The strangle is a U with a flat floor between its two strikes, so its worst outcome is the same across an 800 point range and the position does not begin to improve at all until the index clears one of the strikes. Between 24,600 and 25,400 the strangle is completely insensitive to the index, which is a form of dead zone the straddle does not have.

One row in that table is worth stating in words because it is the row that gets skipped. The maximum loss is the premium, in full, and it is reached whenever the index finishes at the strike for a straddle or anywhere between the strikes for a strangle. This is not a remote scenario reserved for strange markets. It is what happens on a quiet month, which is the most common kind of month there is.

Where the cheaper structure stops being the better one

The natural next question is which of the two comes out ahead, and the answer has a shape that is easy to state and easy to get wrong. Subtract the straddle's profit at expiry from the strangle's profit at expiry, at every index level, and the difference is not a curve. It is a tent with a flat brim.

Below the lower strike and above the upper strike, the difference is a constant. The straddle is worth exactly 67.5 points more than the strangle, at 26,000 and at 30,000 and at any level you care to name. The gap does not narrow with distance, because past the strikes both structures gain one point for every point the index moves, so whatever gap existed when they both went in the money is preserved forever. Between the strikes the difference rises to a peak of 332.5 points at the strike itself, which is precisely the premium the strangle saved.

So the crossovers sit 332.5 points either side of the strike, at 24,668 and 25,333. Inside that window the strangle is the better of the two; outside it the straddle is. And here is the part that changes how the comparison should be read: the window in which the strangle leads lies entirely inside its own losing zone. At the crossover the strangle is down its full 410.8 point premium. Its breakeven is another 478 points further out. The strangle's advantage is real, it is arithmetically exact, and it is available only in the region where the position has lost money.

This is not a quirk of the particular strikes chosen. It follows from a property of option prices. The premium the strangle saves is the straddle premium minus the strangle premium, which after applying put and call parity reduces to twice the call at the strike, less the call at each of the two outer strikes, plus the strike offset. Option value is a convex function of the strike, which means the first three terms sum to something that cannot be positive. The saving is therefore always smaller than the offset, at every width, on any set of inputs where option prices are convex in the strike. That is a mathematical fact rather than a feature of this model.

The model confirms it at eight different widths. Moving the strikes 100 points out saves 95.7 points against a 100 point give-up; 400 points out saves 332.5 against 400; 800 points out saves 541.6 against 800; and 1,500 points out saves 700.9 against 1,500. The saving grows as the strikes widen but it grows more slowly than the distance surrendered, so the flat gap in the tails widens too. A very wide strangle is a very cheap position that is very much worse than a straddle on any move that actually arrives.

What the strangle is genuinely buying, then, is not upside. It is a smaller number on the downside. A position that risks 411 instead of 743 is a smaller position, and if the honest reason for choosing it is that the smaller number is the one that can be afforded, that is a legitimate reason and it should be stated as such rather than dressed up as a better structure. The trap is the reasoning that runs from cheaper to more efficient. On the evidence above, the cheaper structure is strictly worse in every state of the world where the underlying moves far enough to matter.

The question nobody computes: how far does the index actually go?

A payoff diagram will tell you what happens at every index level. It will not tell you how likely any of those levels is, and that omission is what makes the diagram comforting. Both wings run away to the top of the chart, the shape looks generous, and nothing on it indicates that the profitable region begins beyond a distance the underlying rarely covers.

So the requirement has to be compared against a distribution of moves. The series used here is 24 years of synthetic daily sessions, built with a two state volatility process so that quiet stretches and stressed stretches alternate, plus an occasional jump so the tails are heavier than a normal distribution. Its annualised volatility is 12.0 percent and the kurtosis of its daily returns is 8.94, against 3.00 for a normal series. Taking every overlapping 21 session window gives 6,027 holding periods to measure.

The median absolute move across those windows is 1.83 percent. The straddle needs 2.97 percent, which is 1.63 times the median. The strangle needs 3.24 percent, which is 1.78 times the median. The index reached the straddle's requirement in 1,802 of the 6,027 windows, or 29.9 percent, and the strangle's in 1,586, or 26.3 percent. It finished inside both strangle strikes, where the entire strangle premium is gone, in 44.8 percent of them.

Now the detail that matters more than any of those figures. The index moved at all, meaning by more than half a percent, in 84.8 percent of the windows. Movement is not scarce. Movement of the required size is. The gap between those two statements is the whole psychology of the trade: the market is busy almost every month, the position is watched moving up and down every day, and none of that activity is the thing being paid for.

There is a further result here that runs against the intuition of anyone who has heard that markets have fat tails. Fat tails should make a distance threshold easier to reach, not harder. Yet a normal distribution with the same implied volatility would have put the chance of clearing the straddle breakeven at 42.5 percent, and the fat tailed series delivered 29.9 percent. The reason is that variance can be delivered in two ways. A series can move a moderate amount most of the time, or it can be quiet most of the time and then move violently. Both can carry identical volatility. The second shape spends more of its life below any fixed threshold, and it is the shape real markets have. Clustering and jumps make the extremes larger and simultaneously make the ordinary month quieter, and a position whose payoff begins at a fixed distance is hurt by the second effect.

The honest counterweight belongs here, because leaving it out would produce a tidier page and a less true one. When the series did clear the straddle's breakeven, it did not clear it by a little. The median excess was 1.52 percent of the index, about 379 points beyond the breakeven, and the ninetieth percentile of the excess was 5.62 percent, about 1,406 points. The largest window in 24 years moved 22.10 percent. The top five percent of windows account for 20.6 percent of all the distance the series travelled. So the outcome distribution is strongly skewed: the requirement is met less often than a naive reading suggests, and when it is met the result is frequently far past the threshold rather than just over it. Any honest account of these structures has to hold both halves of that at once.

What the arithmetic will not support is the middle position, which is the one most buyers actually hold: that the index is likely to move enough, and that a moderately busy month will do. It will not. On this series a moderately busy month leaves the position short, and the outcomes that pay are the ones that were unusual at the time they happened. That is the difference between buying movement and buying an unusual amount of movement, and only the second one is for sale.

What waiting costs

The distribution above measures the move over the full holding period. It says nothing about the path, and the path has a cost of its own, because a bought option is a wasting asset from the moment it is paid for. Hold the index perfectly still at 25,000 and revalue both structures each day and the shape of that cost becomes visible.

Time value falls roughly with the square root of the time remaining, which produces a curve that is gentle at first and then very steep. At the halfway point, 15 days in, the straddle is still worth 70.7 percent of what was paid for it. That sounds like a slow bleed until you read it the other way round: half the calendar has gone and 70.7 percent of the cost is still to come, so the second half of the holding period is more than twice as expensive as the first. The first full day of waiting costs 12.5 points. The last day costs 135.7, a ratio of about eleven to one.

The strangle decays faster in proportional terms, and by a wide margin. With a week to go it retains 22.5 percent of what was paid, against 48.3 percent for the straddle. With one day left it retains 0.3 percent, effectively nothing, while the straddle still holds 18.3 percent. The reason is structural: the strangle's value is entirely time value, since neither leg has any intrinsic value while the index sits between the strikes, and time value is the part that goes to zero. The straddle always has one leg with intrinsic value once the index moves at all, and that part does not decay.

This changes what the cheaper structure is. A strangle held into the final week with the index inside the strikes has become an almost pure lottery ticket: a position whose value depends on a large move arriving in the next few sessions and which will be worth essentially nothing otherwise. That may be an accurate description of what was intended, but it is rarely what the buyer believed they were holding when they compared the two premiums at entry.

There is a practical consequence in the interaction between decay and the distribution. The requirement of 2.97 percent was computed at expiry. Exit early and the residual time value works in the holder's favour, so the effective threshold before expiry is slightly nearer than the expiry breakeven. Hold to the end and there is no time value left to soften anything, and the payoff is the raw distance. The two effects pull in opposite directions and neither is large enough to change the conclusion, but a position sized on the expiry payoff and then held for its full life is being asked to do the hardest version of the job.

The event trade, and why being right is not enough

The most instructive failure of a bought volatility structure is the one that occurs around a scheduled event, because it is the case where the buyer is right about the thing they were betting on and loses anyway. It is worth working through in full.

Ahead of a known date, the outcome is unknown by definition, so implied volatility rises and premiums swell. Our page on implied volatility covers why that happens and what it means for pricing. The consequence for a structure with two bought legs is that both of them are being purchased at the same inflated level at the same time, so the position is doubly exposed to whatever happens to implied volatility afterwards. And what happens afterwards is not in doubt. The instant the outcome is public, the uncertainty that justified the premium is gone.

Model it. Buy the at the money straddle eight days before the event with implied volatility at 22 percent rather than the ordinary 13. The structure costs 649.7 points, so the expiry breakeven already requires a 2.60 percent move. One day later the event has happened, the index has risen 2.0 percent to 25,500, and implied volatility has fallen back to 12 percent.

The position is now worth 547.9 points, against 649.7 paid. It has lost 101.8 points, or 15.7 percent of the premium, on a day when the index moved 500 points in a single session. Decomposing the change shows exactly where it went. One day of decay took 42.0 points. The 2.0 percent move added 131.3 points, which is the position doing precisely what it was bought to do. The fall in implied volatility from 22 percent to 12 percent removed 191.1 points, more than the move contributed. The three add to the 101.8 point loss.

Read the middle line again, because it is the one that gets lost. The move was worth a genuine 131 points. The trade thesis was correct. The event produced a large single day move in the underlying, which is the outcome the structure exists to capture, and the position still finished down because the price paid for it embedded a bigger move than the one that arrived, and because the repricing of every option on the underlying happened simultaneously.

The arithmetic of what would have been needed is worth stating. To return the entry premium after the crush, the index had to be 2.50 percent away, about 624 points, rather than the 500 points it delivered. A correct directional call, on the right day, in the right size of move, that was still a quarter short of the requirement. This is the specific reason that entering a bought volatility structure into an elevated implied volatility is a materially different proposition from entering the same structure in a quiet market, and the difference is not visible anywhere on a payoff diagram, which is drawn at expiry and assumes implied volatility no longer exists.

The same mechanism runs in reverse, and honesty requires saying so. A structure bought when implied volatility is low and held into a rise in implied volatility gains on the repricing as well as on the move. The asymmetry is not in the mechanics, it is in the timing: implied volatility is highest exactly when a scheduled event makes a large move seem most likely, which is exactly when a buyer is most tempted, so the moments of maximum interest and maximum price coincide.

The charge stack, and the cost that is not a charge

A two leg structure pays the charge stack twice on the way in and twice on the way out, so a single round trip is four charge events rather than the two an outright option pays. That doubling is the reason several multi leg structures are unviable at retail size, and it is worth computing rather than assuming.

The result complicates the usual story, so it is worth stating plainly. The statutory stack on a bought option structure is small. It comes to 0.23 percent of the premium on both structures and it moves the required move from 2.97 percent to about 3.00 percent for the straddle and from 3.24 percent to 3.27 percent for the strangle. On this evidence, charges are not what makes buying volatility difficult. The breakeven distance is. Any account of these structures that blames the charge stack for the outcome is pointing at the wrong line.

The line that does the damage is the one that is not a charge at all. Four leg events at an illustrative half spread cost the straddle four points and the strangle six, several times the entire statutory stack in both cases. And the ordering reverses. As a share of the premium committed, the spread costs the cheaper structure more than twice what it costs the dearer one, 1.69 percent against 0.77 percent, because out of the money options are quoted wider and the premium the spread is measured against is smaller. The structure that looked cheaper on the premium line is the more expensive one to get in and out of.

One further line deserves correction rather than repetition. Holding to expiry rather than squaring off attracts the transaction tax on the intrinsic value of the exercised option at the same 0.15 percent rate that applies to the premium on a sale, payable by the purchaser. At expiry an option's premium is its intrinsic value, so the two routes cost the same. Square off with time value still in the price and the tax is charged on the larger number, which makes selling the leg marginally the dearer of the two rather than the cheaper. The folk rule that an in the money option must be closed before expiry to avoid a punitive exercise charge does not survive the current rate structure.

What each one needs, and how each one fails

The two structures fail in different ways, and the difference is more useful than the similarity. Setting the failure modes side by side is also the fastest way to see what the entry decision is actually deciding.

One row is doing more work than the rest. The subtle failure of the strangle, a move that clears the spot but not the strike, has no counterpart in the straddle, and it is the failure a buyer is least prepared for. The index moved. It moved in a direction. It moved by a full percent and a half. And the position returned nothing, because the payoff does not begin until 1.6 percent and the premium was never recovered. There is no equivalent experience with a straddle, where a move of any size at all changes the value of the position immediately, even if not by enough.

The context for all of this is the base rate published by the regulator: about 93% of individual traders in equity derivatives made net losses over FY22 to FY24, with aggregate net losses exceeding ₹1.8 lakh crore (SEBI, September 2024). A material part of that sits in short dated bought options, which are the same structures as these with less time and a proportionally larger required move. Nothing on this page is an argument that the arithmetic can be beaten by choosing the other structure. The arithmetic is the same arithmetic for both.

What the two structures are actually for

It would be reasonable to reach the end of this page and conclude that buying volatility is simply a bad idea. That is not the finding, and it is not what the numbers say. The finding is narrower and more useful: the payoff diagram is not the trade. The trade is the relationship between a required distance and a distribution of distances, and that relationship can be computed in five minutes before any money is committed.

Everything on this page reduces to three numbers that a buyer can produce for any bought volatility structure, on any underlying, at any expiry. The first is the premium as a percentage of the underlying, which is the required move. The second is how that percentage compares to the underlying's typical move over the same holding period, which is the ratio that tells you whether you are asking for something ordinary or something unusual. The third is where implied volatility currently sits relative to where it usually sits, which tells you whether the repricing risk is on your side or against you. On the illustrative structure here those three numbers are 2.97 percent, 1.63 times the median move, and an implied level of 13 percent against a realised 12.0 percent. None of them required a model more complicated than arithmetic.

The comparison between the straddle and the strangle turns out to be the least interesting question of the three, which is itself worth knowing, since it is the question most comparisons spend all their time on. The convexity result settles it: the cheaper structure is not the more efficient one, it is the smaller one, and it is strictly worse in every state of the world where the underlying moves far enough to pay. If the reason for preferring it is that the smaller premium is what can be committed, that is an honest reason and it belongs in the position sizing conversation rather than the structure selection one.

What remains after all the arithmetic is a discipline rather than a technique. Compute the required move before looking at the payoff diagram. Compare it to the distribution rather than to a feeling about whether the market seems likely to be busy. Check where implied volatility sits before assuming the structure will behave the way the diagram suggests. Those three habits are unglamorous, they take a few minutes, and they are what separates a considered position from an expensive opinion about the news. They are also most of what a serious curriculum in derivatives spends its time on, and if the arithmetic here was the interesting part rather than the tedious part, that is the method we teach.

Every figure comes from a single deterministic script, seeded so it reproduces identically on each run. Options are valued with a standard lognormal model at a zero interest rate on an illustrative index at 25,000, with 30 calendar days to expiry and an implied volatility of 13 percent held flat across strikes. The distribution work uses a synthetic series of 6,048 daily sessions built from a two state volatility process with an added jump component, scaled to an annualised volatility of 12.0 percent, from which every overlapping 21 session move was measured. Charge rates are taken from the site's verified statutory research and applied per leg; the bid and ask spread and the brokerage assumption are stated illustrative inputs, not verified rates.

All results are illustrative and simulated. They are not a track record, they are not a forecast, and they are not an indication of what any structure would produce on any real index. The purpose is to demonstrate the relationship between a premium, a required distance and a distribution of distances, which is a property of the instruments rather than of any particular market. Contract specifications, lot sizes and statutory rates change; verify all of them at source before relying on any figure here.

• The framework for choosing a structure →
• Reading an option chain, column by column →
• What implied volatility actually measures →
• Lot sizes and the contract value band →