Skip to content
Arthvyuh

What is algo trading for retail traders in India (2026)

Algo trading used to mean large firms with servers next to the exchange. Today a retail trader in India can write rules, test them on years of prices and, through a broker’s API, let software place the orders. This article explains what that really involves in 2026: the parts of an algorithm, the rules SEBI has set, what you need to connect, and what an algorithm cannot do for you.

On this page
  1. Algo trading in one sentence
  2. The four parts of any algorithm
  3. What SEBI’s rules say in 2026
  4. What you need to start
  5. Kinds of rules retail traders automate
  6. The risks no algorithm removes
  7. A sensible way to begin

Algo trading in one sentence

Algorithmic trading means a computer program sends orders according to rules written in advance. The rules say what to trade, when to enter, when to exit and how much to trade. Once the rules are set, the program follows them without waiting for a person to press a button.

That is the whole idea. An algorithm does not know anything a person cannot know. It is faster and more consistent at following instructions, and it does not get tired or nervous. It is also exactly as good or as bad as the rules it was given.

A simple example of a rule set for NIFTY options: at 09:20 on a Tuesday, sell one lot of the at-the-money call and one lot of the at-the-money put; exit both if their combined premium rises 30% above the entry value; otherwise exit at 15:15. Every part of that sentence can be written as a setting. Whether it is a good idea is a separate question that only testing and your own judgement can address.

The four parts of any algorithm

Instrument: what is traded. For an options strategy that means the index, call or put, buy or sell, the expiry and how the strike is chosen, for example at the money, 100 points away, or the strike whose delta is nearest to 0.3.

Entry: what has to be true before a trade starts. It can be a time (enter at 09:20), an indicator condition on a completed candle (a moving-average cross on 5-minute candles), or a price level. The information must be available at the moment of entry; a rule that peeks at the day’s closing price at 10 a.m. is not a rule a program can follow live.

Exit: a stop-loss, a target, a trailing stop and a square-off time. Exits need the same precision as entries, including what the stop measures: the option’s premium and the index are different prices.

Size: how many lots. NIFTY options trade in lots of 65 units today, so one lot of an option priced at ₹100 is ₹6,500 of premium. Sizing decides how large every win and every loss is.

What SEBI’s rules say in 2026

SEBI’s circular of 4 February 2025, “Safer participation of retail investors in algorithmic trading”, brought retail algo trading through broker APIs under a common framework. A second circular of 30 September 2025 set the timeline: from 1 April 2026 it applies to all stock brokers.

The main points a retail trader will notice: brokers may not allow open APIs. API access needs an API key specific to you and a static IP address that your broker has whitelisted, with OAuth login and two-factor authentication. Each order placed by an algorithm carries an identifier issued by the exchange, so it can be traced. The broker is responsible for the algorithms that reach the exchange through its API.

In practice this is why brokers such as FYERS now ask for a separate trading app (on FYERS the App ID ends in -200) and a Primary IP. It is also why you log in to your broker yourself every day: the session belongs to you, not to the software.

What you need to start

A trading account with a broker that offers an API, with the F&O segment active. A way to write rules: code, or a no-code strategy builder. Historical prices to test the rules on. And, for live orders, the broker’s API app, a static IP and a daily login.

Arthvyuh puts these steps in one place. The Strategy builder turns your rules into settings without code. Backtests run them on one-minute option prices since 2019 for NIFTY and BANKNIFTY and since 2023 for SENSEX, with estimated charges. Practice replay lets you step through a past day with an option chain. Paper trading runs a strategy with virtual money on live prices. Live trading sends real orders through your own FYERS API app.

Kinds of rules retail traders automate

Time-based rules enter and exit at fixed clock times, for example selling an at-the-money straddle at 09:20 and closing it by 15:15. They are simple to state and to test, and they make the effect of the time of day easy to see.

Indicator-based rules enter when a condition on completed candles is met, such as a moving-average cross on the index or on the at-the-money straddle’s premium chart. They need care about when a candle is complete, and about how many settings are tuned.

Positional rules hold a position for days, exiting on a stop, a target or the contract’s expiry day at the latest. Each type has different costs, different overnight risks and different ways of failing, and a backtest should be read with that in mind.

The risks no algorithm removes

SEBI’s study of individual traders in equity derivatives (FY25–FY26, published August 2026) found that in FY26, 82.1% had net losses before transaction costs and 87.7% after them. Automating a trade does not change that arithmetic by itself.

Overfitting is the most common trap. If you change rules again and again until the past looks smooth, you have described the past, not found something that will repeat. A rule set that only works on the period it was tuned on is likely to disappoint on new dates.

Execution is not the same as a backtest. A live order can be rejected, partly filled, or filled at a worse price than the model assumed, especially in fast markets and in far out-of-the-money strikes. Internet, broker or exchange problems can delay orders. A configured stop is an instruction, not a promise of the exit price.

Costs add up. Every order pays brokerage and statutory charges, and an algorithm that trades often pays them often. Selling options can lose far more than the premium received.

A sensible way to begin

Write down your idea in plain words first. Turn it into precise rules. Backtest it on a long period and keep a later period aside that you do not tune on. Replay a few past days, including an expiry day and a day with a large move, to see how the position behaved minute by minute.

Then run it in paper trading for a while and compare the paper fills, charges and exits with what the backtest expected. Only you can decide whether and how much to trade with real money. The tools can show how a rule behaved; they cannot tell you what the market will do next.

Try it free in paper trading

Create a free account

Related on Arthvyuh

Education only, not investment advice. Examples are invented arithmetic, not anyone’s results, and do not predict any future result. Arthvyuh is not registered with SEBI as an investment adviser or research analyst.

Continue learning

The tools