How we test strategies

1. The data

Backtests run on NSE price candles: open, high, low, close and volume for each stock, from one-minute candles up to monthly ones, going back to 2017 for the stocks that traded then. New candles arrive through the trading day and are checked for gaps and repeated candles; the admin team sees a daily health report and fixes holes before they reach a result. A stock that stops trading (suspended or delisted) is not filled in or guessed.

Daily scan pages use only stocks whose latest candle is from the most recent trading day, so a stock that has not traded never appears in "today's" list.

2. The charges

Every trade pays what a real trade would. By default we use a discount broker's published rate card, read on 15 September 2026 and rechecked line by line on 28 September 2026:

• Brokerage: intraday and futures 0.03% of the order value or ₹20 per executed order, whichever is lower; delivery ₹0; options ₹20 per executed order.
• STT at the statutory rate (for intraday equity, 0.025% on the sell side).
• Exchange transaction charges, SEBI turnover fee and stamp duty at their published rates.
• GST at 18% on brokerage, exchange and SEBI charges.

A round trip is two orders, so it pays brokerage twice. You can change any rate for your own broker before a run. The free brokerage calculator and STT calculator use the same numbers.

3. How a trade is filled

• Entry price: by default a trade enters at the close of the candle that gave the signal, as Zerodha Streak does. You can switch to the next candle's open, which is TradingView's default.
• Slippage: 0.01% of the order value on the entry and again on the exit, unless you set your own.
• Stops and targets are checked against each candle's high and low. When one candle touches both, a strategy that sets its own stop takes the stop first (as Freqtrade does); every other run follows the price path inside the candle, as TradingView does.
• Intraday positions are squared off at 15:25 IST by default. The NSE session runs 09:15 to 15:30 IST.

4. One account, one pool of money

A backtest starts with one account (₹1,00,000 unless you change it). Each trade is sized from the money in the account at that moment, after earlier profits and losses. When a strategy runs on a basket of stocks, every stock draws from the same account, so a losing streak on one stock leaves less money for the others, as it would in real life. A trade the account cannot pay for (or, for futures, cannot cover the margin of) is not taken.

5. What we publish

• Every result, good or bad. Our 7-strategy study shows each strategy's real result after charges, whatever it was.
• Every trade. Each result lists its trades with entry, exit, charges and the reason for the exit, and you can run it again yourself.
• The period and the stocks. A result always names the stocks, the dates and the candle size it was tested on.
• No cherry-picking. We do not tune a strategy on the period we then show you. Where a strategy is tuned, the results shown are from months it never saw.

6. Limits you should know

A backtest is the past. It cannot know about news, gaps that happen while the market is shut, a stock you could not actually buy because it was locked in a circuit, or your own nerves. Treat a good result as a reason to paper trade an idea, not as proof that it will make money. If you find a number that looks wrong, write to [email protected] with the page or the backtest; we check it and correct it in public.