The 5-delta weekly NIFTY strangle: wins eight trades in ten, and one weekend took 40% of the profit

The idea

Options far from the market rarely end up worth anything. A setup popular with traders sells the NIFTY weekly call and put whose delta is about 0.05, which the options market itself prices as roughly a 1-in-20 chance of each one ending in the money. It sells them four trading sessions before expiry, holds them to expiry day, and adds a stop: if the two options together become worth 3.5 times what they were sold for, close both.

The appeal is the win rate. It is often shown winning nine trades in ten. The question this test asks is the one a high win rate tends to hide: how big are the losses when they come?

There are no protective options bought further out, and the position is held over nights and weekends.

The rules

• Entry: at 11:15 am, four trading sessions before each NIFTY weekly expiry (for a Thursday expiry, the Friday before), sell one call and one put of that weekly expiry.
• Strikes: the call and the put whose delta is nearest 0.05, on 50-point strikes. Delta is worked out with the Black-Scholes formula from each option’s own price at 11:15 am, NIFTY at 11:15 am, a 6.5% interest rate and the calendar time left to 3:30 pm on expiry day.
• Stop: close both legs the first minute the two options together close at 3.5 times the credit received. Checked on every 1-minute close, including the first minute after an overnight or weekend gap.
• Exit: otherwise, buy both back at 3:15 pm on expiry day (by then the pair is usually worth 10 to 15 paise).
• Cost: built up from the published charge schedule at ten lots per order: ₹20 brokerage on each of the four orders, STT of 0.15% on the premium sold, exchange charges, SEBI fees, stamp duty and GST, plus a flat 0.25 point of slippage per option on each side. That comes to about 1.2 points per trade, round trip, of which 1.0 point is slippage. See Wrong costs, in either direction.

The backtest

Data: 1-minute NIFTY weekly option prices, 277 weekly trades from April 2021 to August 2026 (5.3 years). Fills are taken at the 1-minute closes. All figures are per lot in NIFTY points (one lot = 65 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.)

Two weeks were skipped because the entry day was a short special session (a Saturday session in March 2024 and the Diwali session in October 2025) with no 11:15 am price.

What the trades look like

• Small premiums. The average credit was about 18 points for the pair. The call was on average about 650 points (3.2%) above NIFTY and the put about 770 points (3.8%) below it; puts this far out cost more, so a 0.05-delta put sits further away than a 0.05-delta call.
• Wins are small and frequent. 229 of 277 trades (83%) made money. The average win was +16.7 points.
• Losses are rare and large. 48 trades were stopped out. The average loss was −60.5 points, about 3.6 times the average win. Over the five years the winners made +3,823 points and the losers gave back 2,906 of them.
• Costs take a quarter. About 1.2 points on an 18-point credit took about a quarter of the gross profit.

The high-win-rate trap

The stop is set at 3.5 times the credit, which on a typical trade means a loss of about 45 to 50 points. But the stop can only act when the market is open. The worst trades were the ones where NIFTY moved over a weekend or overnight:

Losses include costs. “Share” is the loss as a percentage of the total net profit over five years (+917 points).

The worst trade sold the pair for 20 points and bought it back for 386, about 19 times the credit, because the first price after the weekend was already far beyond the stop. That single trade cost about 40% of five years’ net profit, the same as 22 average wins. It also put 2025 in the red for the year even though 85% of that year’s trades won.

Year by year, per lot:

*From April 2021; 2026 up to 11 August.

The Tuesday era so far. NIFTY’s weekly expiry moved from Thursday to Tuesday in September 2025. The 228 Thursday-era trades made +504 points (t 0.87); the 49 Tuesday-era trades made +413 (t 2.0), on slightly larger premiums (21 points against 17.5). One good year is not enough to tell whether the change helped.

What it takes in capital. A short strangle needs about ₹2.25 lakh of margin per lot. Losses on sold options can arrive within minutes of the open, so we size the buffer on the worst intraday drawdown: 592 points × 65 ≈ ₹0.39 lakh, from March to May 2025. That makes the capital about ₹2.6 lakh, on which the average profit of about ₹11,200 a year is a return of about 4.2% a year, below a deposit. That drawdown took about 15% of the capital, and equals about 3½ years of average profit.

Robustness tests

Changing one setting at a time:

• Every setting made money, but the plateau is uneven. Further out (delta 0.03) and a day later (three sessions before) cleared our bar; closer in (delta 0.08) barely did better than nothing, with a worst drawdown of about 1,000 points. Picking the best row after the fact would be fitting the rule to the past.
• The stop cost more than it saved, in this sample. With no stop at all, the same trades made three times as much and won 94% of the time: most stop-outs were spikes that reversed before expiry. That is exactly the result that tempts traders to drop the stop. It also means every position is then exposed, with nothing to close it, to a move bigger than anything in these five years.
• Only 17 of 22 quarters were profitable, short of the 85% we expect from option selling. The worst quarter (April to June 2025, −343 points) was about four times a typical good one.
• Not carried by a few winners. Without its five best trades it still made +713 points, and no quarter made more than about a quarter of the profit. The problem is the other end: a few losers.

Risks

• Uncapped losses. Nothing is bought further out to limit a move. The stop is a market exit and fills wherever the market is: in April 2025 that was about 19 times the credit.
• Overnight and weekend gaps. Positions are held for about a week, including a weekend. Most of the worst losses happened at the open after a gap.
• Thin premiums. With about 18 points of credit, a point of extra slippage per trade takes a large share of the profit; on pessimistic fills almost nothing was left.
• Margin. Margins rise on volatile days, exactly when the position is losing.
• Rule changes. Weekly expiries, lot sizes and margins have all changed in recent years and may change again.

Pros & cons

Pros

• Wins about 83% of trades; most weeks are quiet and profitable.
• Simple rules: one entry a week, one stop, no adjustments.
• Profitable before costs, and in both halves of the test.
• Losses on most stopped trades stayed near the planned size.

Cons

• The average loss is about 3.6 times the average win.
• One weekend gap cost about 40% of five years’ profit.
• After costs, the profit is not distinguishable from luck (t 1.49), and about 4% a year on capital is below a deposit.
• Very sensitive to fills and to how far out the strikes are.

Our verdict

It passed three of our eight checks: it was profitable in both halves, its profit did not rest on a few winners, and it traded often enough to judge. But after realistic costs the edge was too small to rule out luck, the return on capital fell short of a deposit, only 17 of 22 quarters were profitable, and doubling the costs or assuming worse fills left little or nothing.

The 83% win rate is real. So is the trade-off behind it: many small wins, paid for by a few large losses, the biggest of which arrived over a weekend, when no stop could act. On this record, the win rate describes the quiet weeks; the result depends on the others.