Not investment advice. This is a personal learning log. The authors are not SEBI-registered investment advisers or research analysts. Nothing here is a recommendation to buy or sell any security. Intraday levels are published only after the session has closed, as a record of what was tested. The portfolio is simulated, and past results don't predict future ones. Do your own research.
Equity desk: an open notebook on Indian stock picking
This site is the working log of a small, AI-assisted equity desk that studies NSE-listed stocks. It shows how we pick stocks, what we tested, what went wrong, and which rules we changed because of it. It's updated every trading day.
There are two separate experiments:
- Quarterly fundamental screen. Every quarter, a Screener.in growth-and-quality screen produces a short list of strong businesses. They're bought only when the price trend agrees, and are tracked in a paper portfolio (simulated money, no real orders). Read the method.
- Next-day intraday candidates. Each evening, liquid NSE stocks are filtered on end-of-day momentum and noise, with fixed entry, stop and target levels for the next session. A small, strictly risk-capped real account (at most ₹150 at risk per trade) tests it. Read the method.
Where things stand
- Paper portfolio (quarterly screen), closed: HINDZINC and OFSS were opened on 21 Sep 2026 with an 8% trailing stop on closing prices, and both were closed at the 30 Sep 2026 close for −₹43,434 (−4.34%). HINDZINC lost ₹22,239 and OFSS lost ₹21,195, while the Nifty fell 3.39% over the same period. No costs were modelled. The picks were sound businesses, but we bought them at the wrong time and put too much into each. The lessons are now rules.
- IPO paper sleeve, closed: ₹0 realised. Adroit Industries wasn't allotted (about 1 in 290 odds in our category) and was refunded. A tracked-only line at the category minimum would have made +₹1.57 lakh at the NSE listing price, but it doesn't count toward P&L. German Green Steel & Power was withdrawn and refunded.
- US paper sleeve, closed: +1.49%, against −0.73% for SPY over the same period.
- Now: all three paper books were reset on 30 Sep 2026 and hold only cash. The next quarterly screen runs on about 17 Nov 2026, after the September-quarter results deadline.
- Intraday: 3 real trades so far (one target hit, two stops), about −₹56 net in total. That's far too few to judge. We'll review after about 20 trades. A 58-session backtest of common intraday setups found no positive edge after costs.
Latest changes
- 3 Oct 2026: site launched. Added the 58-session intraday backtest, the stop-width study and the 1 Jul paper-portfolio replay.
- 3 Oct 2026: corrected the status: all paper books were closed and reset to cash on 30 Sep 2026 (equity −4.34% against Nifty −3.39%, IPO sleeve ₹0, US sleeve +1.49% against SPY −0.73%).
- 1 Oct 2026: intraday: two opening stop-outs in a weak market. Added a confirmation entry while the Nifty is below a falling 50-day average (from 5 Oct).
How to read this site
- Strategy: the rules we use today, with the date each rule was added.
- Journal: every intraday brief, posted only after its session has closed, plus the outcome.
- Experiments: backtests and replays, including the ones that failed.
- Code: short snippets of the scripts, where the exact formula matters.