User-Friendly Quant Strategies for Small Retail Traders (₹5–10 Lakh)

 Below are three of the easiest yet best-performing strategies distilled from Ernest P. Chan and Rishi Narang’s work, tailored for solo quant traders with limited capital. Each strategy includes rationale, step-by-step implementation, and an Upstox API script for safe backtest, paper-trade, and live deployment.

1. Intraday Buy-on-Gap Mean Reversion

Source: Chan’s Buy-on-Gap model
Why It’s Easy: Simple entry on large overnight gaps; only one signal per day per stock.
Target Capital: ₹5–10 Lakh across 2–4 stocks
Expected Metrics: APR ~8–12%, Sharpe ~1.5

Strategy Details

  • Universe: 3–5 high-liquidity NSE stocks (Reliance, TCS, HDFC Bank, Infosys)

  • Entry: If today’s open < previous close × 0.99 (gap down ≥1%)

  • Exit: Market close on same day

  • Position Sizing: Equal capital per stock; max 25% per ticker

Implementation Steps

  1. Environment Setup

    bash
    python3 -m venv quant_env
source quant_env/bin/activate
pip install upstox-api pandas numpy vectorbt

  2. Upstox Authentication

    python
    from upstox_api.api import Upstox, Session
API_KEY, API_SECRET, REDIRECT_URI = 'YOUR_KEY','YOUR_SECRET','https://app/callback'
session = Session(API_KEY, API_SECRET); session.set_redirect_uri(REDIRECT_URI)
print(Upstox(API_KEY,None).get_login_url(redirect_uri=REDIRECT_URI,response_type='code'))
# After login, exchange code:
session.set_code('RECEIVED_CODE'); token = session.retrieve_access_token()
u = Upstox(API_KEY, token)

  3. Data Retrieval & Backtest

    python
    import pandas as pd, numpy as np, vectorbt as vbt
def get_ohlc(symbol,start,end):
    bars = u.get_ohlc(u.get_instrument_by_symbol('NSE_EQ',symbol),'ONE_MINUTE',start,end)
    df=pd.DataFrame([{ 'dt':b.timestamp,'open':b.open,'close':b.close } for b in bars]).set_index('dt')
    return df.resample('1D').agg({'open':'first','close':'last'})
def backtest_gap(df):
    df['prev_close']=df.close.shift(1)
    df['gap_pct']=(df.open-df.prev_close)/df.prev_close
    df['pos']=(df.gap_pct< -0.01).astype(int)
    df['ret']=df.close.pct_change()
    return (df.pos.shift(1)*df.ret).fillna(0)
symbols=['RELIANCE','TCS']
rets=[backtest_gap(get_ohlc(s,'2023-01-01','2025-10-14')) for s in symbols]
port=vbt.Portfolio.from_returns(sum(rets))
print(port.stats())

  4. Paper Trading

    python
    from datetime import date
today = date.today().isoformat()
for sym in symbols:
    df = get_ohlc(sym,today,today)
    if (df.open.iloc[0]/df.close.shift(1).iloc[0])<0.99:
        print(f"Signal BUY {sym}")

  5. Live Order Placement

    python
    from upstox_api.api import TransactionType,OrderType,ProductType
def place_buy(sym,qty):
    inst=u.get_instrument_by_symbol('NSE_EQ',sym)
    u.place_order(inst,qty,TransactionType.Buy,ProductType.Intraday,OrderType.Market)
# On signal day:
capital=500000; per_stock=capital/len(symbols)
for s in symbols:
    price=get_ohlc(s,today,today).open.iloc[0]
    qty=int(per_stock/price)
    if qty: place_buy(s,qty)

  6. Risk Controls & Shutdown

    • Kill switch if drawdown >5%

    • Flat positions at 15:20 IST via market‐sell

2. Simple Moving-Average Crossover (Trend-Following)

Source: Narang’s concept in Inside the Black Box
Why It’s Easy: Only two moving averages; daily signals; works on liquid ETFs/stocks.
Target Capital: ₹5–10 Lakh
Expected Metrics: APR ~10–15%, Sharpe ~1.2

Strategy Details

  • Universe: Nifty 50 ETF (e.g., NIFTYBEES) or liquid blue-chips

  • Fast MA: 20-day; Slow MA: 50-day

  • Entry: Fast MA crosses above Slow MA (Go Long)

  • Exit: Fast MA crosses below Slow MA (Go Flat)

Implementation Steps

  1. Data & Signal Generation

    python
    def backtest_macross(df):
    df['MA20']=df.close.rolling(20).mean()
    df['MA50']=df.close.rolling(50).mean()
    df['pos']=(df.MA20>df.MA50).astype(int)
    df['ret']=df.close.pct_change()
    return (df.pos.shift(1)*df.ret).fillna(0)
df=get_ohlc('NIFTYBEES','2020-01-01','2025-10-14')
stats=vbt.Portfolio.from_returns(backtest_macross(df)).stats()
print(stats[['Sharpe Ratio','Total Return']])

  2. Paper Trade Logic

    python
    today_df=get_ohlc('NIFTYBEES',today,today)
if today_df.MA20.iloc[-1]>today_df.MA50.iloc[-1]:
    print("Go Long ETF")
else:
    print("Go Flat")

  3. Live Execution

    python
    def rebalance_etf(sym,capital):
    df=get_ohlc(sym,today,today)
    pos= int((capital/df.close.iloc[-1])) if df.MA20.iloc[-1]>df.MA50.iloc[-1] else 0
    # Send order to match pos
    # Retrieve current qty via u.get_positions()

  4. Risk Controls

    • Maximum leverage 1.5×

    • Stop-trail at 1% daily drawdown

3. Sectoral ETF Mean Reversion (Pairs)

Source: Chan’s ETF pairs with Bollinger Bands
Why It’s Easy: One ETF pair, simple z-score; low trade frequency.
Target Capital: ₹5–10 Lakh
Expected Metrics: APR ~12–18%, Sharpe ~1.4

Strategy Details

  • Pair: BankBees (banking ETF) vs AutoBees (auto ETF)

  • Spread = Price(AutoBees) – β × Price(BankBees) (β via simple regression)

  • Entry: Spread > mean + 2 σ (short) or < mean – 2 σ (long)

  • Exit: Spread returns to mean

Implementation Steps

  1. Spread Calculation & Backtest

    python
    from sklearn.linear_model import LinearRegression
df1=get_ohlc('BANKBEES','2023-01-01','2025-10-14')
df2=get_ohlc('AUTOBEES','2023-01-01','2025-10-14')
lr=LinearRegression().fit(df1.close.values.reshape(-1,1),df2.close)
beta=lr.coef_[0]
spread=df2.close - beta*df1.close
mu,σ=spread.mean(),spread.std()
z=(spread-mu)/σ
pos = np.where(z>2,-1,np.where(z< -2,1,0))
ret=pd.Series(df2.close.pct_change().fillna(0))*pos[:-1]
stats=vbt.Portfolio.from_returns(ret).stats()
print(stats[['Sharpe Ratio','Total Return']])

  2. Paper Trading

    python
    # Same spread logic on today’s bar; generate pos

  3. Execution

    python
    def trade_pair(sym_long,sym_short,qty):
    u.place_order(u.get_instrument_by_symbol('NSE_EQ',sym_long),qty,TransactionType.Buy,ProductType.Delivery,OrderType.Market)
    u.place_order(u.get_instrument_by_symbol('NSE_EQ',sym_short),qty,TransactionType.Sell,ProductType.Delivery,OrderType.Market)

  4. Risk Controls

    • Max capital per pair: ₹2 Lakh

    • Weekly rebalancing; close positions if no mean reversion in 5 trading days

Deployment Plan & Safety Measures

  1. Local Testing & Backtesting

    • Run all backtests on historical data; verify Sharpe >1.2.

    • Use vectorbt for performance reports.

  2. Paper Trading Phase (1–2 months)

    • Auto-generate signals; log P&L without real orders.

    • Validate live data feeds and latency.

  3. Live Deployment (Start Small)

    • Allocate ₹1–2 Lakh initially per strategy.

    • Scale capital gradually only upon consistent performance (≥2 months).

  4. Monitoring & Kill Switches

    • Daily P&L dashboard (Streamlit/Grafana).

    • Automatic full-stop if drawdown >5% in a week.

  5. Regulatory Compliance

    • Register algorithm with SEBI-registered broker (Upstox).

    • Maintain audit logs for 3 years.

    • Adhere to position limits (5% free float) and margin requirements.

By focusing on Buy-on-Gap, MA Crossover, and ETF Pair Mean Reversion, retail traders with ₹5–10 Lakh can implement robust, low-complexity quant strategies using Upstox API—with clear steps for data, backtesting, paper-trading, and live deployment under SEBI guidelines.

References
Ernie P. Chan, Algorithmic Trading (2013)
Rishi K. Narang, Inside the Black Box (2009)

Comments

Post a Comment

Popular posts from this blog

Fyers Automate

What Is Fyers Automate? Fyers Automate is a workflow-based, no-code trading automation platform built natively within the FYERS brokerage ecosystem. Unlike external algo tools that rely on third-party integrations or require coding, Automate runs directly on FYERS' own infrastructure, making it the first of its kind among Indian discount brokers. It was built in alignment with SEBI's recently finalised framework for retail algorithmic trading. expresscomputer +1 How It Works Automate uses a drag-and-drop visual canvas where you build workflows using three core building blocks : [ fyers ]​ Triggers — define when the workflow wakes up (e.g., price breakout, range breakout, webhook signal, order status, portfolio P&L level) Actions — define what to do (place order, square off, modify order, cancel order, send alerts) Conditions — define the logic path the workflow follows (IF/AND/OR/ANY branches for multi-leg or conditional strategies) You can combine these into sophisti...
Read more

Options Income Strategies System

  Executive Summary & Implementation Roadmap Executive Overview This comprehensive guide delivers a complete options income generation system for Indian markets, designed for ₹5-10 lakh capital. The framework prioritizes  regime-adaptive strategy selection ,  risk-first execution , and  scalable implementation  from manual trading to full automation. Top 3 Recommended Starting Strategies: Short Iron Condor  (Sideways/Low Vol) — High win-rate, defined risk, capital-efficient Credit Put Spreads  (Neutral-Bullish) — Directional bias with controlled risk Short Strangles with Protective Wings  (Sideways/Medium Vol) — Higher premium, managed tail risk Capital Deployment Philosophy: Start with 10-15% of capital per strategy (₹50k-₹1.5L per position) Maximum 3 concurrent strategies initially Reserve 40% capital for margin buffer and opportunity reserve Per-trade risk cap: 1.5-2% of total capital (₹7.5k-₹20k max loss per trade) Expected Timeline to Pro...
Read more

Complete Roadmap to Become a Successful Profit-Making Full-Time Quant Trader in India

Based on Ernest Chan's framework and Indian market expertise, here's your optimized 18-month roadmap to quant trading success: Phase 1: Foundation Building (Months 1-6) Educational Foundation 1 Essential Skills to Master : Mathematics & Statistics  (Months 1-2) Linear Algebra : Matrix operations, eigenvalues, PCA Statistics : Probability theory, hypothesis testing, regression analysis Time Series Analysis : ARIMA, cointegration, Kalman filters 2 Resources : Khan Academy, MIT OpenCourseWare, "Statistics and Machine Learning Toolbox" 2 Programming Mastery  (Months 2-4) Primary Language: Python 3 # Core Libraries to Master import pandas as pd # Data manipulation import numpy as np # Numerical computing import matplotlib.pyplot as plt # Visualization import scipy.stats as stats # Statistical analysis import sklearn # Machine learning import vectorbt as vbt # Backtesting framework Chan's Recommendation : "MATLAB for resea...
Read more