What we tested, and what we found

Passed · long-term

Momentum basket

Each week, the 20 strongest NSE 500 stocks by momentum (6- and 12-month returns against their volatility), equal weight, kept while they stay in the top 40, traded at the next open.

About 8 points a year over an equal-weight basket of the same stocks

NSE 500, 2019 to 2026, after costs. Run on today’s index members, so it is optimistic. Try it on paper from today or replay it from a past date.

Passed · in the forward test

Results-day drift

A stock that beats the NIFTY 500 by 5 points or more over the two days around its results, bought at the open two sessions later and held 20 sessions.

+1.19% a trade over random stocks

1,757 trades from 2019, range +0.57% to +1.82%; +2.40% after reactions of 10 points or more. A reaction 5 points or more below the index was followed by a lag of 2.35%: a warning, not a trade. Today’s index members; the effect fades after about a month.

✕Intraday order-block zones

The zone plans the app draws, tested on a year of 5-minute candles across the NSE 500 in 482 versions: none made money after costs. As drawn, they lost 0.34R a trade over 64,704 trades, and 0 of 480 variants were positive. The app still draws the zones as a map of levels, labelled “no tested edge”.

✕Intraday candle setups, six years

Every 5-minute bar from Oct 2020 to Oct 2026, with the rules written down before the data was opened. Nothing passed after 0.18% costs: an opening-range breakout lost 0.09% a trade, a gap fill 0.14%. MoneyHQ’s candle engine is well calibrated, but the odds it measures are coin tosses once costs are paid. The desk’s candle check shows this record for any setup you ask about.

✕Twelve classic chart patterns

Double bottoms, flags, wedges, head and shoulders and more, on 2,921 NSE stocks (the delisted included), Nov 2018 to Oct 2026: none has a solid edge; the bull flag and head and shoulders are a weak hint. The chart shows each pattern with this record.

✕Insider buying

Promoters and directors buying their own company’s shares in the market. In the 500 most traded stocks it passed the first half of the data but not the second: +0.43% over 60 sessions, range −1.47% to +2.14%. In smaller liquid stocks it held (+1.86%), but they are outside what MoneyHQ covers.

✕Bulk and block deals

Stocks a big client bought ₹25 crore or more of in a day went on to lag random stocks by 1.38% over the next 20 sessions (range −2.74% to +0.09%). Block deals showed no direction either way.

✕Option signals

Straddles around results priced the move about right (6.9% priced, 6.4% moved); the put/call ratio stopped working; futures build-up labels and implied-volatility rank gave no usable filter; selling the weekly NIFTY straddle made about nothing.

Every company, not just the survivors

Tests run on NSE’s own daily files for every listed company since Nov 2018: 2,850 companies, the delisted included, adjusted for splits, bonuses and demergers. Testing only today’s index members makes almost anything look good, because it leaves out the stocks that failed.

Against random stocks

Each trade is compared with random stocks from the same universe bought and sold on the same days. That separates the setup from the market simply rising.

After costs

0.35% a round trip for delivery trades and 0.18% for intraday, taken off every trade.

Rules fixed first

The newer tests write their rules down and commit them before the data is opened, choose any setting on the first half of the data only, and run the second half once for the verdict.

Ranges, not just averages

Results come with a 90% range from resampling by month. A result whose range includes zero doesn’t pass.

Checked after launch

What the app shows each day is logged and checked again each evening, so the rules keep a live record beside their tested one.