Posted in My Daily Journal · 30 October 2026
A full day of backtests on NIFTY option buying — the 2.33× partial, weekly ITM-150 versus monthly ITM-500, how long winners run versus losers, and what SEBI's FY25-26 study says about traders like us.
We buy deep in-the-money NIFTY options on a pivot-breakout signal and trail the stop on index pivots. Should we book one of two lots early, the way Sahil describes, and let the other run?
Yes, and the number is 2.33 times the entry box. We defined the "box" as the distance between the last pivot high and pivot low at entry. Selling one lot when the index has moved 2.33 boxes in our favour, and trailing the second lot on the pivot stop with no target, kept 90–95% of the money of two trailing lots while cutting the maximum drawdown by about 20%. Every multiple between 2.3 and 3.0 behaved alike; 2.0 was worse. On an intraday breakout system with a premium trail the same partial failed on both data sources — it only works when the runner is a multi-day position.
Weekly or monthly contracts for a system that holds overnight?
On seven years of real option prices they are almost identical when the weekly is bought with at least five days to expiry and rolled a day before it expires: ₹10.1L versus ₹10.2L per lot, with the weekly's profit-to-drawdown slightly better (5.0 vs 4.4) and a lot costing roughly ₹16k instead of ₹38k of premium. The weekly loses if you buy it too close to expiry — the version that rolled into the same expiring contract bled every year, and that was a bug in the first test, not a property of weeklies. The deciding factor is the bid-ask spread you actually pay, which a backtest cannot know; we measure it in the live market before trusting either.
How long do winners run compared with losers?
Winning signals held a median of 516 minutes, losers 177 — about three to one. All of the profit came from positions held two or more nights (+₹25.6L); the intraday-only part of the same signals lost ₹17.25L. The twenty largest winners produced 95% of the profit and were held a median of four nights. That is the whole case for holding through the night and for never cutting winners early.
Can we book one lot early and hold the second lot to the weekly expiry, as Sahil literally says?
Tested. Holding the runner to expiry from a break-even stop gave a better profit-to-drawdown ratio (7.5) but a lower total and two flat years; on monthly contracts it was a disaster (drawdown doubled). We kept the trailing runner.
Does the same idea work on BankNifty?
The partial keeps only 83–89% of the two-lot money on BankNifty — its winners run further relative to the box — so there we would trail both lots. BankNifty has had no weekly expiry since November 2024; the tradeable twin is the monthly contract.
SEBI published a study of individual F&O traders. What does it say about option buyers like us?
Of individual traders in FY25-26, 93% traded only options as buyers and about 90% of them lost money; the median return was −114%. The worst cell was traders turning over ₹1–10 crore of premium a year on ₹1–5 lakh of capital — 92% lose, average loss ₹3.24 lakh. A two-lot system of ours turns over roughly that much, which is why the capital rule we use is premium plus the measured maximum drawdown per lot, never premium alone. The report does not mention algorithmic or API traders at all.
Could the overnight edge on NIFTY exist on commodities too?
We ran Sahil's daily 5-EMA stop-and-reverse on MCX futures from 2021. Silver worked on every measure and beat buy-and-hold in both halves of the sample; gold made money but less than simply holding it; crude, natural gas, copper, zinc and aluminium failed. Five trades out of sixty-five made all of silver's profit, so it is a lead for collecting intraday and option data, not a system to trade.