Friday, 25 September 2026. The SENSEX moved. Premiums moved. Plenty of people made money and plenty lost it. Our machine did neither, because it did not place a single order.
Zero trades. Zero rupees. A flat line on the day.
If you came here for a hero story, this is the wrong day. But if you have ever blown up a Friday afternoon trying to force one more trade out of a market that was not offering anything, this is the day worth reading about.
Our method is boring on purpose. Before the session, it maps out the levels on SENSEX where price has previously turned — the ceilings and the floors. Then it waits. It does not buy because the chart "looks bullish". It only acts when price breaks one of those mapped levels and the break is confirmed rather than a one-candle poke.
When it does take a trade, three things are decided before entry, not after: a fixed stop-loss, a scale-out point where part of the position is booked, and a trailing floor that follows the trade once it is in profit.
Today, the first condition never happened. Price wandered between the levels all session. No confirmed break, no entry. The other rules never even got a chance to run.
Think of an auto driver at a stand near the station. Trains are late, the platform is empty, nobody is walking out. He has two options.
He can sit at the stand and burn nothing. Or he can start circling the city empty, burning petrol, hoping to spot a passenger. The second option feels like work. It feels productive. At the end of the day, it costs him ₹400 of fuel and earns nothing.
Most retail F&O accounts die from circling, not from crashing. Brokerage, STT, the spread on every entry and exit, and the slow bleed of small forced trades. SEBI's own studies have repeatedly found that the overwhelming majority of individual F&O traders end up in loss. A flat day costs you nothing. Ten forced trades cost you real money before the market has even made up its mind.
Fourteen trading sessions have produced results this month. Eleven were positive, three were negative. Here is the whole month, including the parts we would rather not print.
| Date | Day P&L | Date | Day P&L |
|---|---|---|---|
| 01 Sep | −₹19,350 | 16 Sep | +₹8,507 |
| 03 Sep | +₹38,723 | 18 Sep | +₹12,150 |
| 04 Sep | +₹3,600 | 21 Sep | +₹2,070 |
| 08 Sep | +₹11,192 | 22 Sep | +₹65,408 |
| 09 Sep | +₹25,287 | 23 Sep | +₹20,705 |
| 10 Sep | −₹26,200 | 24 Sep | +₹2,295 |
| 11 Sep | +₹5,723 | 25 Sep | ₹0 — no trade |
| 15 Sep | −₹19,800 | Month to date | +₹1,30,310 |
Look at the shape of that table rather than the total. One session, 22 September, contributed ₹65,408 — half the month. Strip that single day out and the month is ordinary. That is what a level-break method looks like from the inside: a long stretch of small numbers and flat days, punctuated by the occasional session where a break runs.
The number that matters in this method is not the win rate. It is what happens on the days the machine is wrong. A fixed stop-loss decided before entry means a bad trade has a known, bounded cost. A trailing floor means a good trade that reverses does not become a bad one. And a day with no confirmed break means neither rule is needed.
The temptation on a day like today is to loosen the definition of a break so that something — anything — qualifies. That is how a rulebook quietly becomes a feeling. We logged the day as zero and closed the terminal.
Same rulebook, same levels re-mapped on the fresh data, same entry conditions. Nothing changes because today was blank, and nothing would have changed if today had been ₹50,000 either. That consistency is the only thing we are actually trying to build.