144 settings, 968 sessions: testing an opening-range breakout on Nifty options
Buy a call or put when the Nifty breaks out of its opening range with trend, volume and open-interest support, then let a trailing stop run the winner. It sounds disciplined. Across 968 sessions of 1-minute data, every one of 144 settings lost money after costs.
The rules we tested
The strategy took long calls or long puts only, with one position at a time. Every condition was evaluated on the close of a 1-minute bar, and the entry filled at the open of the next bar.
- Required: the Nifty closes beyond both the 09:15 to 09:29 range and the prior 20-bar high or low, with a candle body at least half of its range, and the 5-minute and 15-minute trends agree.
- Confirmers, at least three of four: price on the right side of session VWAP, 5-minute RSI in a band, put or call open interest building faster across strikes near the money, and the option's premium rising.
- Filters: entries from 09:30 to 14:30, expiry day skipped, premium at least ₹20, a 10-minute cooldown, three trades a day at most, and a stop for the day after two losses or a ₹1,500 loss.
- Exits: an initial stop from the option's own volatility, breakeven at +1R, then a trailing stop with no fixed target, plus a 20-minute fail-fast rule, a VWAP exit and a 60-minute limit.
How we kept the test honest
The first 60% of sessions, to March 2024, were used for choosing settings, and the last 40% were kept for testing. Settings were chosen by a plateau score, the average of a setting and its neighbours, so a lucky peak could not win. Data after September 2025 was not opened at all.
Fills were deliberately cautious. If a stop and a new high could both have happened in one bar, we assumed the stop came first. Stops updated only after a bar closed, and a gap through a stop filled at the open. We charged 0.5 point of slippage per side, tested 1.0 as well, and applied dated charges for brokerage, STT, exchange fees, stamp duty, SEBI fees and GST. Rupee results use one lot of 65.
Results
| Version | Trades | Win rate | Net per trade | Profit factor |
|---|---|---|---|---|
| Starting rules, first 60% | 1,018 | 37.3% | -₹177 | 0.51 |
| Starting rules, last 40% | 676 | 36.8% | -₹105 | 0.74 |
| Best stable setting, first 60% | 951 | 42.5% | -₹126 | 0.65 |
| Best stable setting, last 40% | 638 | 39.5% | -₹141 | 0.65 |
| Best stable setting, last 40%, slippage 1.0 | 632 | 36.2% | -₹209 | 0.53 |
None of the 144 settings was positive in either period, so there was no profitable plateau to choose from. The 95% bootstrap interval for the best setting's test-period expectancy ran from -₹203 to -₹80 per trade, entirely below zero. In the walk-forward check, with 12 months of training and three of testing, all 12 test windows were negative.
Is it the costs?
Not on their own. The best setting lost about ₹78 per trade before charges, and charges added about ₹63. Removing every fee would still have left a loss.
Where the money went
| Exit reason (test period) | Trades | Average net |
|---|---|---|
| Trailing stop | 216 | +₹674 |
| Initial stop | 315 | -₹739 |
| Gap through stop | 63 | +₹181 |
| Fail-fast | 19 | -₹355 |
| VWAP exit | 13 | -₹522 |
The trailing stop did what it was designed to do: when a trade worked, it made money. There were just too few of those trades. Calls and puts lost almost the same, ₹141 and ₹142 per trade. Every hour of entry lost money, and 17 of 19 test months were negative. The top five trades supplied only about 10% of gross profit, so this was not a case of a good strategy missing one lucky day.
Did the entry rules add anything?
This is the most useful comparison. We replaced the entries with random ones, keeping the same exits and the same number of trades, and repeated that 100 times. The random entries lost a median of ₹121 per trade, with a 90th percentile of -₹86. The rules lost ₹141. They beat only 20% of the random runs.
In other words, all those conditions, breakout, trend, VWAP, RSI and open interest, added no measurable value over chance in this test. They did reduce how often the strategy traded, which can feel like quality, but it was not visible in the results.
What to take from this
- Compare with random entries. A strategy with many conditions can look selective. The only way to know whether the conditions help is to test the same exits without them.
- Look at gross as well as net. A strategy that loses before costs has a different problem from one that loses only after costs. Cheaper execution would not fix the first kind.
- Separate entry and exit questions. A trailing exit can work while the strategy fails. Knowing which part works tells you what is worth testing next.
Limits of this study
There is no bid-ask history, so fills are bar prices plus an assumed slippage, and real spreads may be wider. The index has no volume, so VWAP here is an equal-weighted proxy. Open interest changes in only about 40% of minutes, which makes it a coarse signal. RBI and CPI dates were not excluded, though skipping Budget days and one election result day changed nothing material. These findings apply to these rules on this sample. They do not prove that no intraday option strategy can work. Full tables are on the research page.
This is research on historical data, not a recommendation. See our disclaimer and editorial policy.