cat ~/projects/nse-algo-trading-system/README.md

NSE Algo Trading System

Automated stock trading system for Indian NSE markets. Screens 2400+ stocks daily across 10 strategies with multi-checkpoint validation.

Python1pushed 1d ago
Quantitative FinanceTechnical AnalysisSignal ProcessingBacktestingPortfolio Optimization2400+ NSE-listed stocks, daily OHLCVSolo Developer
signal.overview

The Indian stock market moves fast — 2400+ listed equities, each generating a new candle every day. Manually scanning for setups across that universe is impossible. This system automates the entire pipeline: ingest, screen, validate, and surface only the highest-conviction trades.

Under the hood, ten distinct strategy modules run in parallel — momentum breakouts, mean reversion, RSI divergence, volume spikes, moving average crossovers, and more. Each candidate signal passes through a multi-checkpoint validation gate before it surfaces. The system doesn't just find patterns; it stress-tests them.

The output is a daily briefing: ranked trade candidates with entry/exit levels, position sizing via Kelly Criterion, and a backtest performance snapshot for each strategy. It turns a 2400-stock haystack into a 5-10 trade shortlist every morning.

run.simulation()
nse-algo-trading-system — interactive demo
OHLC — 60 day simulated NSE data
equity curve — backtest P&L
Total Trades
2
Win Rate
50%
Max Drawdown
3.57%
Return
+6.56%
screener: 72 of 2400 stocks pass filter
cat ARCHITECTURE.md
PythonpandasNumPyyfinancePlotlyStreamlit
  • Pipeline architecture: data ingestion → strategy modules → validation gate → ranking → output. Each strategy is a pluggable module conforming to a common interface.
  • Backtester runs walk-forward optimization to avoid overfitting. Position sizing uses fractional Kelly with a 0.5x safety multiplier.
  • Screener uses vectorized pandas operations across all 2400+ stocks simultaneously — no row-level loops.
© 2026 Om Gorakhiasys.uptime: ∞