Rohit, if you have ever sat in front of a screen from 9:15 to 3:30, watched a trade go your way, waited for "a little more", and then watched it come back to zero — this session will read strangely to you.
On Friday, 18 September 2026, our SENSEX engine was inside the market for a total of six minutes. Three trades. That was the whole day.
The machine does one thing. Before the session it marks the price levels where the SENSEX has repeatedly turned around — the ceilings and the floors. Then it waits. It only enters when the index confirms a break of one of those levels. Every entry carries a fixed stop-loss, a planned scale-out, and a trailing floor that ratchets up as the trade moves in favour.
Here is the full book for the day, nothing hidden:
| In | Out | Option | Entry | Exit | Move | P&L |
|---|---|---|---|---|---|---|
| 09:34 | 09:34 | 74500 PE | 487.85 | 488.10 | +0.25 | +₹225 |
| 10:28 | 10:33 | 74500 PE | 450.35 | 454.40 | +4.05 | +₹3,645 |
| 11:59 | 12:00 | 74400 CE | 535.00 | 544.20 | +9.20 | +₹8,280 |
Our model book is 900 quantity, so every one-point move in the option premium is ₹900. Check the arithmetic yourself: 0.25 × 900 = ₹225. 4.05 × 900 = ₹3,645. 9.20 × 900 = ₹8,280. Nothing is rounded, nothing is smoothed.
All three trades exited on the trailing floor, not on a profit target. That distinction matters more than the number at the bottom of the page.
A target means we decided in advance where to get out. A trailing floor means we never decided — we simply kept raising the line under the trade, and when the price fell back to touch that line, the position closed. The market chose the exit. We only chose where we would refuse to give more back.
And the honest footnote: the planned scale-out did not fire on a single trade. None of the three moves ran far enough to reach the first scale-out step. Trade one — the 09:34 put — made ₹225. That is a rounding error, effectively a scratch. On a 900 lot, it barely covered the cost of being there. We are not going to dress it up as a win.
Ten sessions have been recorded in September so far. Seven were green, three were red. The red ones were not small: −₹19,350 on the 1st, −₹26,200 on the 10th, −₹19,800 on the 15th. That last one stung particularly, because the same level that lost money on one lane made money on another.
Month to date the book stands at ₹39,832 across those ten sessions. Averaged out that is about ₹3,983 a session — a number nobody would put on a poster, and exactly the kind of number that is worth trusting more than a screenshot of one spectacular afternoon.
The machine was idle for 99.7% of Friday. It sat through the entire post-lunch session without taking a position, because the index never confirmed a break of a marked level after 12:00.
Six minutes of exposure. Three entries, each with a stop-loss already sitting in the market before the trade was live. Three exits handed to a trailing floor. That is the whole method, and it is deliberately boring.