Your idea: AI Algo Backtesting Specification — NSE Stocks, 5-Minute Two-Candle St…

AI Algo Backtesting Specification — NSE Stocks, 5-Minute Two-Candle Strategy 1. Objective Build and backtest an intraday equity trading strategy using 5-minute candlestick charts for individual NSE-listed stocks. * Market: NSE India cash equities. * Timeframe: 5-minute candles. * Maximum planned risk: ₹1,000 per trade, including estimated costs. * Risk–reward ratio: 1:2. * Maximum trades: One trade per day across the entire stock universe. * Trading style: Intraday cash equity; no options, futures, leverage or overnight positions. 2. Stock universe Use a predefined, configurable universe of liquid NSE-listed stocks. Start with NIFTY 50 constituent stocks as of each historical date, using historical constituents to avoid survivorship bias. Apply consistent liquidity and price filters and exclude suspended or unsuitable stocks. Do not select stocks retrospectively based on which performed best. 3. Entry signal Evaluate the first two completed 5-minute candles after the NSE cash market opens at 9:15 AM IST. Bullish signal — Buy shares 1. First candle closes below its open (red). 2. Second candle also closes below its open (red). 3. After the second candle closes, buy the selected stock. 4. Enter at the next realistically executable price after confirmation. Bearish signal — Short-sell shares 1. First candle closes above its open (green). 2. Second candle also closes above its open (green). 3. After the second candle closes, short-sell the selected stock, subject to applicable intraday short-selling rules. If either candle is a doji or the candles have different colours, that stock does not qualify. Important: The strategy is contrarian. Two red candles trigger a buy, and two green candles trigger a short sale. 4. Stock selection Scan the eligible stock universe after the second candle closes. If multiple stocks qualify: * Rank them using a predefined, configurable criterion, such as relative volume or percentage change. * Select only the highest-ranked eligible stock. * Never use future candle data to rank candidates. * If no stock qualifies, do not trade that day. Test the ranking criterion separately from the base strategy. 5. Stop-loss and target Use a 1:2 risk–reward ratio based on the actual stock entry price. For a long trade: * Stop-loss = entry price − (1% × entry price). * Target = entry price + (2% × entry price). For a short trade: * Stop-loss = entry price + (1% × entry price). * Target = entry price − (2% × entry price). These are initial test parameters and should be configurable. Model stop-loss gaps and slippage realistically. 6. Position sizing Maximum planned risk per trade, including estimated transaction costs and slippage, is ₹1,000. For a long trade: Risk per share = Entry price − Stop-loss price. For a short trade: Risk per share = Stop-loss price − Entry price. Calculate: Maximum quantity = floor(₹1,000 ÷ (risk per share + estimated costs per share)). Also ensure the position is affordable within the available capital and complies with applicable broker and exchange rules. If the calculated quantity is zero, skip the trade. 7. Exit rules * Exit when the stop-loss or target is reached. * If neither is reached, close the position before market close. * Do not carry positions overnight. * Do not average down or increase risk after a loss. * Permit only one trade per day across all scanned stocks. 8. Backtesting Backtest at least 3–5 years of reliable historical data, if available. Report: * Total trades and skipped signals. * Long versus short trades. * Win rate and average win/loss. * Gross and net profit after brokerage, STT, exchange charges, GST, stamp duty and other applicable costs. * Profit factor and maximum drawdown. * Largest loss and consecutive losing trades. * Monthly and yearly performance. * Results by stock and market regime. * The effect of realistic slippage, gaps and short-selling restrictions. Use historical stock constituents, realistic execution