BTST option momentum on NIFTY: an option-buying strategy that passed all eight checks

The idea

Most option strategies that do well in backtests sell options: they collect premium and win small amounts often. Buying options is the opposite bet, and most rules for it lose money: premiums decay every day, and the buyer pays costs on every trade.

This one, published on a public strategy-research site, is a momentum rule. Each morning it marks the NIFTY weekly call and put trading near ₹100. If either one jumps 50% during the day, something is moving, so it buys that option, holds it overnight, and sells it the next morning. “BTST” is the Indian traders’ shorthand for buy today, sell tomorrow.

The rules

• Mark: at 9:20 am, find the NIFTY weekly call and the weekly put whose prices are nearest ₹100, on strikes that are multiples of 100. Use the nearest weekly expiry with at least one trading day left, so on expiry day it uses the following week’s.
• Entry: buy an option the first time it trades 50% above its mark (₹150 if the mark was ₹100), any time from 9:21 am to 3:29 pm. Calls and puts are separate: on a day of big swings both can trigger. At most one trade per side per day.
• Stop: sell if the option falls to 50% of the price paid, checked every minute, overnight included. A gap through the stop at the next open fills at the opening price.
• Exit: otherwise, sell at 9:25 am the next trading day.
• Cost: built up from the published charge schedule at ten lots per order: ₹20 brokerage per order, STT on the sell side, exchange charges and GST on the actual premiums, plus 0.5% slippage on each side. That comes to about 2.1 points per trade, round trip, most of it slippage.

The backtest

Data: 1-minute NIFTY weekly option prices, 881 trades from April 2021 to July 2026 (5.3 years). Fills are taken at the trigger price, or the minute’s opening price if it jumped past the trigger. 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.)

Six of the 887 signals couldn’t be priced (a missing minute of option data at the mark or the exit) and were left out.

Why it works

• The winners are much bigger than the losers. The average win was +84 points and the average loss −57. The stop caps most losses near half the price paid, about 74 points. A winner has no cap: the best trade made +645 points, buying a call at ₹139 that was worth ₹790 the next morning.
• The profit is in the overnight hold. The same rule, but sold at 3:20 pm on the day of entry, lost money (−1,172 points, t −0.65). Buying after a big move and holding it to the next morning is what earns.
• Calls and puts both work. Calls made +3,470 points, puts +4,165, with similar strength (t 1.95 and 1.99). It isn’t a disguised bet on the market rising.
• Costs are a small share. At ten lots per order, costs took about a fifth of the gross profit. Option buying at ₹150 is far cheaper to trade than a multi-leg spread.

Where it hurts

Most trades lose. Only 47% of trades made money; 255 of the 881 were stopped out. Long losing runs are normal for this kind of strategy.

Gaps through the stop. 34 trades opened the next morning below their stop. They lost 93 points on average, and the worst lost 148: a put bought at ₹169 on 21 Oct 2022 opened at ₹22 after the long weekend.

Losses include costs.

The edge has thinned. Year by year, per lot:

*From April 2021; 2026 up to July.

Every year made money, but 2024 and 2025 were weak: in 2025 the average trade was just over 2 points, about what the costs are. 2026 recovered, mostly on one strong quarter (January to March, +1,912 points, a quarter of the five-year total).

What it takes in capital. Buying options needs no margin beyond the premium: about ₹19,500 for a lot each of a call and a put at ₹150. Counting only the premium as capital would make the return look enormous and hide the risk, so for option buying we take capital as 10 times the premium, about ₹1.95 lakh per lot: a conservative rule that keeps premium to a tenth of the account. On that, the average profit of about ₹93,000 a year is a return of about 48% a year. The worst drawdown, −995 points (≈ ₹64,700 per lot), ran from April to September 2024 and took about 33% of that capital. Risk is Medium on our scale: the worst drawdown was about eight months of average profit.

Robustness tests

Changing one setting at a time:

• A broad plateau. Every nearby setting made money with a t near or above 2. Selling later in the morning was the weakest: by 9:45 am about 30% of the profit was gone.
• Both halves made money, with about the same total, but the second half is less certain (t 1.77) because the 2024–2025 trades were weak.
• 18 of 22 quarters were profitable. The worst quarter (July to September 2025, −423 points) was about the size of a typical good one.
• Not carried by a few trades. Without its five best trades it still made +5,159 points, and no quarter made more than a third of the profit.

Risks

• Overnight gaps. Every trade is held overnight, many over weekends. No stop can protect against a gap: 34 trades opened below their stop, and the worst lost almost the whole premium.
• Long losing runs. With fewer than half the trades winning, a run of stops is normal; the worst drawdown lasted five months and took about a third of the capital.
• Fading edge. The edge was strongest in 2021–2022 and weak in 2024–2025. A market with fewer big moves leaves little to catch.
• Fills. The entry is a fast-moving option at the moment it jumps 50%. Filling at the minute’s close instead of the trigger cut the profit by almost a quarter; real fills on a sharp move can be worse.
• Expiry-rule changes. NIFTY’s weekly expiry moved from Thursday to Tuesday in September 2025, and contract details change from time to time; the rule depends on a weekly expiry being available.

Pros & cons

Pros

• Profitable after costs over five years, beyond what luck explains (t 2.78).
• Losses on most trades are capped by the stop; the winners have no cap.
• Robust: every nearby setting worked, it survived double costs, and both calls and puts made money.
• Needs only the premium, no margin, and costs are a small share of the profit.

Cons

• More than half the trades lose; long losing streaks are part of the deal.
• Gaps make some losses close to the full premium.
• Weak in 2024 and 2025; the most recent profit rests on one strong quarter.
• The worst drawdown took about a third of the capital, a five-month losing stretch.

Our verdict

It passed all eight of our checks: the first option-buying strategy on The Algo Bench to do so. After realistic costs it made about 8.7 points a trade over 881 trades, every nearby setting worked, and it held up at double costs, in both halves of the test and on pessimistic fills. The profit comes from holding overnight after a big move, with the stop limiting most losses to about half the premium.

Passing every check is not the same as being safe. More than half the trades lose, gaps can take almost the whole premium, the worst drawdown took about a third of the capital, and the edge was thin in 2024 and 2025. This is a test of past data, not a forecast.