The SENSEX overnight ₹30 strangle: nine wins in ten, and a tail the backtest never met

The idea

Far out-of-the-money weekly options lose most of their value in their last day or two. A strategy circulating among traders on social media tries to collect that last bit of decay on SENSEX: the day before the weekly expiry, sell a call and a put that are each worth only about ₹30, hold them overnight, and buy them back on expiry afternoon, when they are usually worth almost nothing.

It is the opposite of a lottery ticket. Most weeks the options simply expire towards zero and the seller keeps most of the premium. The catch is that the seller has written the lottery ticket: on the rare week the index runs a long way, the loss has no ceiling.

The rules

• Day: the trading session before each SENSEX weekly expiry.
• Entry: at 9:20 am, sell one SENSEX call and one SENSEX put of the expiring weekly series, each the strike whose price at 9:19 am is nearest ₹30 (strikes in steps of 100, at least about 1% out of the money).
• No stop. The position is held through the rest of that day and overnight, whatever happens.
• Exit: buy both back at 3:00 pm on expiry day.
• One position a week.
• Cost: built up from the published charge schedule at ten lots per order: ₹20 brokerage per order, STT of 0.15% on the premium sold, exchange charges of 0.0325% on both sides, SEBI fee, stamp duty on the buys and GST, plus 0.5 point of slippage per option on each side. That comes to about 2.6 points per strangle, round trip: 0.6 of charges and 2.0 of slippage.

The backtest

Data: 1-minute SENSEX option prices, 156 weekly trades from August 2023 to August 2026 (just under three years). Fills are at the close of the entry and exit minutes. All figures are per lot in SENSEX points (one lot = 20 units, today’s lot size, applied throughout).

(t measures whether the average trade is distinguishable from zero; around 2 or more is needed before we call a result more than luck.)

The two options together brought in about 60 points on average, and the strategy kept a little over half of it before costs. On average the position was flat at the close of the entry day: the profit came overnight and on expiry day, as the options decayed.

What it does well

• It wins nine times in ten. 141 of 156 trades made money, and every calendar year was profitable.
• The statistics are strong. An average of +31 points a trade with a t of 4.82, well past our bar, and both halves of the test made money on their own.
• Costs are a small share. About 2.6 points of a 34-point gross average. Doubling them still left t at 4.42.
• The strikes adapt to the market. Because the rule picks a price, not a distance, the strikes sit further away in nervous weeks, when options are dearer, and closer in calm ones. The median call was about 1,200 points (1.5%) above the index and the median put about 1,300 points (1.7%) below, but the range ran from under 1% to about 5%.
• It even profited from a crash. On 7 April 2025 SENSEX opened 4.7% lower. That gap came before the 9:20 am entry, so the strategy sold into the panic at almost double the usual premium, with strikes 4.5% away, and made its best trade, +110 points.

Where it hurts

Each loss costs about three and a half wins. The average win was +54 points, the average loss −186. Fifteen trades lost. Almost all were steady moves of 1% to 3% between entry and exit rather than gaps, and four of the five worst came in calm weeks, when one of the strikes was only about 1% from the index:

Losses include costs, per lot in SENSEX points.

The worst moment didn’t show in the result. On 8 July 2026, SENSEX fell about 2% after the 9:20 am entry, through the put strike. At that day’s close the position was down about 527 points, and at its worst about 650 points (₹13,000 per lot). The next morning the index recovered just enough to finish above the strike, and the trade ended +30. With no stop, how a week ends is luck, not design.

The worst drawdown was −594 points (about ₹11,900 per lot), from late July to mid-September 2024. Year by year, per lot:

*From August 2023; 2026 up to 6 August.

What it takes in capital. We estimate the margin for one lot of a SENSEX short strangle held into expiry day at about ₹2.5 lakh. For option selling we add the worst mark-to-market loss seen at any minute, 650 points × 20 ≈ ₹13,000, making the capital about ₹2.6 lakh. On that, the average profit of about ₹32,400 a year is a return of about 12% a year. On our scale the drawdown would rate Low, but uncapped losses make the risk Extreme, whatever the backtest shows.

The tail the sample never met

Three years of data contain 156 nights. The biggest gap at the open of an expiry day was −2.5% (19 March 2026), and it arrived in a nervous week when the put strike was 2.9% away: the trade lost 191 points. The 7 April 2025 crash came just before an entry, not during one. On 4 June 2024, election-results day, SENSEX closed almost 6% lower, but the strategy was not holding a position that night.

So the sample tells us what happens in normal weeks. It cannot tell us what a large gap in a calm week would do, when the strikes sit about 1.5% away. The table below is an illustration, not a backtest: it assumes SENSEX at about 81,500, the median strike distances (1,255 points to the call, 1,369 to the put), the median premium of 60 points, and a move that is still in place at 3 pm on expiry day, when the options are worth about their intrinsic value.

Loss = the option’s intrinsic value at the moved level, minus the premium received, plus costs. Applied instead to each of the 156 trades’ own strikes and premiums, a 5% move would have lost a median of about 2,500 points (₹50,000 per lot) and up to about 3,700 (₹74,000); only in the few nervous weeks with strikes nearly 5% away would it have been close to break-even.

One 5% night in a calm week would take back about 1.7 years of average profit and about a fifth of the capital. Moves that size are rare, but SENSEX has had several in the past decade, and this position is open for one night in every week of the year.

Robustness tests

Changing the premium target:

• A plateau, with the usual trade-off. Cheaper options win more often and with a steadier curve, but make less; dearer ones make more, with bigger losses. All three are well past our bar.
• 10 of 13 quarters were profitable (77%), below the 85% we ask of option-selling strategies. The losing quarters were July to September 2023 (a part-quarter, −221 points), July to September 2024 (−76) and October to December 2025 (−53).
• Not carried by a few trades. Without its five best trades it still made +4,475 points, and no quarter made more than 15% of the profit.
• Just under three years. The test covers 156 trades but 2.99 years, a whisker short of our three-year minimum: SENSEX weekly options of this kind only have usable history from mid-2023, and our 90-day rule cuts off the latest weeks.

Risks

• Uncapped losses. Both options are sold without protection. A large move costs many times the premium received.
• Overnight gaps, with no stop. Every trade is held overnight. No stop could act before the next open, and this rule has none anyway.
• An unsampled tail. The sample has no large adverse gap in a calm week. The stress table above shows what one could cost.
• Margin. Short options need large margins, more on expiry day; a sharp move can bring a margin call on top of the loss.
• Short history and rule changes. About three years of data, and SENSEX’s weekly expiry day has moved more than once in that time. The exchange can change contract rules, lot sizes and margins again.

Pros & cons

Pros

• Wins about nine trades in ten; every calendar year was profitable.
• Profitable after costs well beyond luck (t 4.82), in both halves and at double costs.
• The ₹30 rule places strikes further away in nervous weeks.
• One trade a week, two orders in and two out.

Cons

• Losses are uncapped, and there is no stop.
• An average loss is about 3½ times an average win.
• The tail risk, a big overnight gap in a calm week, never appeared in the sample.
• Under three years of data; only 77% of quarters were profitable.

Our verdict

It passed six of our eight checks. Over 156 weeks it made about 31 points a trade after costs, with nine wins in ten, a t-statistic near 5, profits in both halves, every nearby setting working, and little sensitivity to costs or fills. It missed two checks narrowly: 10 of 13 quarters were profitable against our 85% bar for option selling, and the data covers just under three years.

The numbers are the easy part. This is a naked short strangle held overnight with no stop, and the sample never met the event that would hurt it most: a large gap in a quiet week. On our illustration, one 5% night would take back about 1.7 years of average profit. Its risk is Extreme because of that, not because of anything the backtest shows. This is a test of past data, not a forecast.