This repository presents a structured research study of six vanilla technical indicator–based strategies applied cross-sectionally across a broad Indian equity universe.
Each strategy is implemented and evaluated independently in a self-contained Jupyter notebook. The objective is to systematically examine the standalone effectiveness of widely used technical indicators by evaluating the signals they generate in isolation using a consistent performance and risk evaluation framework.
The focus of this research is signal validation, cross-sectional robustness, and structural behavior — not production deployment.
All strategies are evaluated using a consistent methodology:
- Cross-sectional daily signal evaluation
- Long-only framework
- Comparison against a buy-and-hold benchmark
- Equal-weighted aggregation across 276 stocks
- Uniform performance and risk metrics across strategies
Performance metrics are computed at the individual stock level and then aggregated cross-sectionally.
- Metrics calculated per stock
- Final results represent the cross-sectional mean across all 276 stocks
- Equal-weighted evaluation framework
- CAGR
- Annual Volatility
- Total Return (%)
- Average Win Size
- Average Loss Size
- Win-Loss Ratio
- Profit Factor
- Maximum Drawdown (%)
- Win Rate
- Value of 1
- Sharpe Ratio
- Calmar Ratio
- Sortino Ratio
- Average Active %
- Median Active %
- Maximum Active %
- % Stocks Profitable
- Median Stocks Profitable
This ensures comparability and robustness assessment across all strategies.
A dual moving average crossover framework designed to capture medium-term directional trends.
A momentum-based signal using MACD crossovers to identify acceleration shifts in price behavior.
A mean-reversion strategy capturing short-term oversold and overbought conditions.
A breakout-based approach capturing volatility expansion and directional continuation.
A volatility compression and mean-reversion framework using band extremes.
A trend-following overlay designed to capture sustained directional movement while limiting downside exposure.
Each notebook contains:
- Strategy logic
- Objective
- Observations
- Performance analysis
- Conclusion
- Most standalone vanilla indicators struggle to consistently outperform buy-and-hold in absolute return terms.
- Trend-following strategies perform better during sustained directional regimes but suffer during sideways markets.
- Mean-reversion strategies demonstrate improved drawdown control but lower long-term compounding.
- Cross-sectional dispersion highlights meaningful variation in stock-level outcomes.
- Risk-adjusted performance varies meaningfully across strategy types.
These findings reinforce the importance of regime conditioning and portfolio construction rather than reliance on single-indicator signals.
- No transaction cost or slippage modeling
- No regime filters applied
- No portfolio-level capital allocation
- No multi-strategy blending
This repository represents structured signal research rather than a deployable trading system.
- Statistical validation of signal edge
- Regime-based conditioning
- Multi-strategy portfolio construction
- Dynamic capital allocation framework
This project reflects a systematic approach toward quantitative signal research and forms the foundation for more advanced portfolio-level strategy development.