Quant investing in India, for retail investors

What quant investing means

A quant strategy has three parts: data, a rule and execution. The data might be prices, financial statements or index membership. The rule turns that data into a list of holdings: the 50 stocks with the strongest 12 month momentum, say, at equal weight. Execution places the trades on a fixed schedule. Nothing in the chain depends on how anyone feels that morning.

That is the whole difference from discretionary investing. A good discretionary manager may use plenty of data. But the final decision, buy or not, sell or hold, is a judgement, made fresh each time. A quant rule makes the judgement once, when it is written and tested, and then applies it mechanically. You can still be wrong. You are just wrong in a consistent, measurable way you can study.

Quant investing is not the same as algo trading

The words get mixed up, and the difference matters for risk.

Both use computers. One builds a portfolio and holds it; the other trades in and out. The rest of this guide is about the first, with a section on the rules that now govern the second.

Three routes for a retail investor

A quant fund is the simplest: you hand the model to a fund house. The catch is that you usually cannot see the rule in enough detail to know what you own, and the fund can change it. A rules based index sits in between: the rule is published, the shares are in your name, and the platform runs it. Your own algorithm gives you full control and full responsibility, including for the code.

SEBI's retail algo framework, in plain words

Until 2025, retail investors who connected their own code or a third party's strategy to a broker's API sat in a grey zone. SEBI's circular of 4 Feb 2025 on safer participation of retail investors in algorithmic trading set the rules. Its start date moved from 1 Aug 2025 to 1 Oct 2025, then to a phased glide path for brokers through January 2026, with full application from April 2026.

1. 4 Feb 2025SEBI issues the retail algo framework

   Brokers become responsible for algos their clients run through APIs.

2. Jul 2025Start moved to 1 Oct 2025

   More time for brokers and exchanges to build the systems.

3. 1 Oct 2025Phased start

   Brokers that are ready go live; others meet milestones for registering algo products and strategies.

4. Jan 2026Mock sessions and onboarding cut off

   Brokers that missed the milestones cannot onboard new retail API algo clients.

5. Apr 2026Full application

   The framework applies to every broker offering retail algos.

What the framework says, in short:

• Brokers are accountable. Algo orders come through the broker, which must get exchange approval for each algo it offers and tag every algo order with an exchange issued ID, so there is an audit trail.
• A threshold for your own code. Orders above a set rate, 10 orders per second per exchange, need the algo to be registered; below it, your own API orders do not.
• White box and black box. An algo whose logic is disclosed and replicable is a white box. One whose logic is hidden is a black box, and its provider must be registered with SEBI as a research analyst and keep a research report for each strategy; any change in logic means registering it again.
• API access is controlled, with static IPs and authentication, to stop unknown programs trading on a client's account.

The detail is in the systematic execution module, and the exchanges publish their own operating rules. Check the current circulars on sebi.gov.in before relying on any one date.

The F&O warning

Much of what is marketed as "algo trading" to retail investors is short term futures and options trading. SEBI's study of individual F&O traders found that 93% lost money over the three years FY22 to FY24, and SEBI raised STT on futures and options again from 1 Apr 2026, with futures at 0.05% and options at 0.15% on sale. Automation does not change the arithmetic of a leveraged, short horizon game with costs on every trade. If someone sells you an algo with a monthly return figure, ask for the full record, the worst month and the leverage.

What to check in any quant product

1. Can you read the rule? At least: universe, signal, number of holdings, weights, rebalance calendar.
2. Is the back test honest? Costs deducted, the worst fall and its length shown, and the dates stated.
3. How much of the record is live? A back test is a model. A record of lists published on their dates, before the results, is stronger evidence.
4. Who holds the money and the shares? Your own demat and bank account is the safest answer.
5. What happens when it goes wrong? Can you exit on any trading day, and at what cost?

The how backtesting works module covers the second point in depth, and the machine learning for alpha module explains why a more complex model is not automatically a better one.

Common mistake: trusting a back test that was tuned to look good

A quant rule can be adjusted until its past curve looks perfect: a different lookback, one more filter, a better start date. Each tweak fits the past a little more and predicts the future a little less. Signs of a tuned test: very smooth curves, no losing years, short histories, and rules that changed after the results were known. Prefer rules with simple logic, long histories and a public record of every change.

How this shows up in RupeeCase

RupeeCase is the middle route in the table above: long only rules based indices, each rule published on its page, the shares held in your own demat at your own broker, and every rebalance list published on its date in the rebalance record. It is not an F&O algo platform; hedging with F&O is a feature that is coming, for protection only. The technology page explains how the engine turns research into a frozen, published rule. Dynamic Asset Allocator shows a quant rule that does more than pick stocks: a fixed equity core with regime switches into debt and gold, rebalanced every 3 weeks, with a model CAGR of 45.18% and a deepest fall of −10.57% as of 27 Apr 2026, on a shorter history than the other rules, which starts in December 2022.

Questions investors ask