How were the methodologies tested?

Each of 142 codified methodologies was run across 44 instruments and roughly ten years of real end-of-day NSE data lagged at least 30 days. Every trade used next-bar-open entry, fixed 1 percent risk, a 2R target, a time stop, and the full Indian retail cost stack (brokerage, STT, GST, slippage). Results were validated out-of-sample, stress-tested with Monte Carlo, and corrected for multiple testing so that apparent edges are not artefacts of trying many rules.

Is this investment advice or a signal service?

No. These are educational findings about how documented methodologies behaved in the past on lagged historical data. Past behaviour is not indicative of future results. Nothing here is a buy, sell, or hold recommendation, a signal, or a prediction. Bharath Shiksha is an educational publisher, not a SEBI-registered investment adviser.

The Evidence Engine

Most trading patterns lose money. We tested them, so you do not have to.

Q: Do candlestick patterns actually work on Indian markets?

When tested mechanically on a decade of real, 30-day-lagged NSE data with realistic costs, most single candlestick patterns show no statistical edge, and several widely-taught ones, including Bullish Engulfing and Hanging Man, lose money after costs. A documented pattern is not automatically a profitable edge. A minority of methods, such as RSI divergence and structure breaks, do show a measurable positive expectancy that survives out-of-sample testing.

Every trading guru sells you the one chart where the pattern worked. We did the opposite: we ran 142 codified methodologies across a decade of real, 30-day-lagged NSE data, with realistic costs and proper statistics, and we publish what actually holds up, including the popular patterns that quietly lose money.

142

methods tested

10yr

real NSE data

held up

lose money

What the data debunks

These are among the most-taught patterns in retail trading. Traded mechanically with costs across a decade of real NSE data, they produced a negative expectancy: on average, they cost money. A pattern being in every textbook does not make it an edge.

Bullish Engulfing (mechanical)loses money
Hanging Man at uptrendloses money
Tweezer tops and bottomsloses money
Naive mean-reversion in trendloses money

What actually holds up

A minority of methods showed a real, positive expectancy that survived out-of-sample testing and a correction for having tried many rules. These are not magic; the edges are often thin. But they are measurable, and they are the ones worth studying.

RSI divergencegenuine edge
Break of market structuregenuine edge
Momentum and moving-average trend entriesthin but real edge

How we tested (and why you can trust it)

Real data: roughly 10 years of end-of-day NSE prices across 44 instruments, every series lagged at least 30 days.
Realistic execution: next-bar-open entry, fixed 1% risk per trade, a 2R target, a time stop, and the full Indian cost stack (brokerage, STT, GST, slippage).
Out-of-sample: the last 30% of history is held back; an edge has to survive there too.
Monte Carlo: 1,000 reshuffles of the trade order to see the typical worst drawdown and the risk of ruin.
Multiple-testing correction: we tested many rules, so we raise the bar before calling anything an edge, to avoid being fooled by luck.

See the full evidence, and test your own ideas.

The complete Evidence Engine, all 142 backtestable methods with equity curves and per-market breakdowns, plus the 1,308-method library and the practice terminal, are included with enrolment. Judge every method by its evidence, not its story.

See enrolment →

Questions

Do candlestick patterns actually work on Indian markets?

Most single candlestick patterns show no statistical edge when tested mechanically with costs, and several widely-taught ones lose money. A minority of broader methods do show a measurable, out-of-sample-positive edge.

Is this a signal service or investment advice?

No. These are educational findings about historical, lagged data. Past behaviour is not indicative of future results, and nothing here is a recommendation, signal, or prediction. Bharath Shiksha is an educational publisher, not a SEBI-registered investment adviser.

Can I run these tests myself?

Yes. The interactive Evidence Engine and Backtester are inside the Edge Terminal, included with enrolment, where you can run any method on real lagged data and read the equity curve, drawdown, and out-of-sample result yourself.