Quantitative Trading Strategy Backtest

A long-only quantitative equity backtester for Indian NSE stocks. The strategy combines relative strength filtering, residual momentum scoring, and mean-variance portfolio optimisation, benchmarked against the Nifty 500 index.

---

How It Works

At each rebalance, the strategy runs through four steps:

1. Relative Strength Filter — Ranks all stocks against the Nifty 500 and keeps the top 200.
2. Residual Momentum Score — From those 200, selects the top 20 using a log-linear regression slope × R² score.
3. Mean-Variance Optimisation — Allocates weights across the 20 stocks to maximise the Sharpe ratio (weights constrained to 2%–10% per stock).
4. Rebalancing & Risk Rules — Rebalances every 21 trading days (or earlier on a volume spike), and cuts any position that drops 30% from entry.

Benchmark data (Nifty 500 / ^CRSLDX) is downloaded automatically from Yahoo Finance at runtime.

---

Requirements

Python 3.8+ and an active internet connection (for benchmark download).

Install dependencies:

pip install pandas numpy scipy matplotlib tqdm yfinance

---

Input Data

A .csv or .parquet file with one row per stock per trading day.

Column	Accepted names
Date	date, tradedate, trading_date
Symbol	symbol, ticker, fid
OHLCV	open, high, low, close, volume
Sector	sector

Note: The script requires ~252 trading days of warmup data before the active trading start date. If your active window begins on 2010-01-01, your data file should include data from at least 2009-01-01.

---

Usage

Basic:

python strategy.py --data NSE_Data_2010_2020.csv

With all options:

python strategy.py \
  --data data.csv \
  --output my_results \
  --start 2020-01-01 \
  --end 2025-01-01 \
  --capital 1000000

Arguments

Argument	Description	Default
--data	Path to data file (.csv or .parquet)	Required
--output	Folder to save results	results
--start	Active trading start date (YYYY-MM-DD)	2020-01-01
--end	Active trading end date (YYYY-MM-DD)	2025-01-01
--capital	Starting capital in rupees	5,000,000

---

Outputs

All files are saved to the --output directory.

Charts

• backtest_performance.png — NAV curve, drawdown, Sharpe, and turnover
• sector_allocation_rebalance.png — Sector mix at every rebalance
• strategy_vs_nifty500.png — Side-by-side benchmark comparison

CSV Files

• portfolio_values.csv — Daily portfolio value and returns
• rebalance_log.csv — Every rebalance event and its trigger
• trade_log.csv — Every individual buy/sell trade
• portfolio_snapshots.csv — Holdings at each rebalance
• sector_allocations.csv — Daily sector weights
• nifty500_benchmark.csv — Strategy vs. benchmark side by side
• rolling_performance_metrics.csv — 1Y / 3Y / 5Y rolling outperformance

Report

• performance_report.txt — Full stats summary (CAGR, Sharpe, max drawdown, etc.)

---

Configuration

Key parameters can be adjusted in the CONFIG dict at the top of strategy.py:

Parameter	Default	Description
top_k_rs	200	Stocks kept after Relative Strength filter
top_n_final	20	Final portfolio size
min_weight	0.02	Minimum weight per stock (2%)
max_weight	0.10	Maximum weight per stock (10%)
static_rebalance_days	21	Rebalance every N days
volume_spike_threshold	1.35	Multiplier to trigger early rebalance
transaction_cost	0.00268	Per-side cost per trade
stop_loss_pct	0.30	Cut position if it falls 30% from entry
use_trailing_stop	False	Set True to use a trailing stop instead
scoring_lookback_days	252	Lookback window for momentum scoring
optimization_lookback_days	252	Lookback window for covariance estimation