Skip to content
Paper tradingSince Jun 2026

quantsys

A rules-based trading system for Indian stocks

My role
Research, design and development
Made for
A personal research project in rules-based trading. It is not a product and not investment advice.
quantsys performance screen with an equity curve and figures such as total return, Sharpe ratio and maximum drawdown, using demo data
Performance view: the account value over time and the figures behind it (demo data).

Not investment advice. Every screen here shows demo data, not real trades.

What it is

quantsys trades the Indian stock market (NSE) by fixed rules instead of gut feeling. It follows 48 large companies and the NIFTY and BANKNIFTY index futures.

The same decision engine runs the tests on past data, the paper trading and live trading, so a result from one holds for the others. Every decision keeps a full record of why it was made and which risk limits changed it.

So far, ten years of tests (2016 to 2026) found no rule that beats simply holding the index after costs. That is why it trades only on paper, in a study that has run since July 2026.

What it does

  • One engine for every mode

    Tests on past data, paper trading and live trading all run the same decision code.

  • Several strategies side by side

    Trend following, pairs of similar stocks, momentum, month-end and expiry-week patterns, each given money according to its measured edge.

  • Risk limits come first

    Each trade risks at most 0.6% of the account, trading stops for the day after a 2.5% loss, and a 20% fall stops everything until a person restarts it.

  • A reason for everything

    Every blocked trade and every change in size goes into an audit trail, so any decision can be explained later.

  • Live trading stays locked

    Real money is only possible after a strategy passes strict statistical tests on real data, the broker is connected and the operator signs in again.

  • A terminal view

    A read-only monitor shows decisions, positions, risk and performance, each opened with a short code, like a trading terminal.

Screenshots

The terminal screens are rendered from demo data.

quantsys decision screen listing the capital split, signals, targets, orders and audit trail for one decision, using demo data
One decision with its signals, targets, orders and full audit trail (demo data).
quantsys monitor screen showing the engine running in paper mode with account, risk and recent events, using demo data
The monitor shows the engine, the account, risk and recent events (demo data).
quantsys strategies screen listing each strategy with its share of capital, results and number of trades, using demo data
Each strategy's share of capital, results and trade count (demo data).
quantsys positions screen listing open positions with profit and loss, using demo data
Open positions, with why each one was entered (demo data).

How it is built

The engine and the research tools are written in Python with NumPy, pandas, SciPy and statsmodels. A FastAPI service and a Next.js dashboard sit on top, backed by PostgreSQL with TimescaleDB, and Redis.

It connects to the Angel One SmartAPI broker, and every setting lives in one configuration file.

  • Python
  • NumPy
  • pandas
  • FastAPI
  • PostgreSQL
  • TimescaleDB
  • Redis
  • Next.js
  • TypeScript
  • Docker

Have a project in mind?

Tell me what you want to build, or ask me about this one.

Email me