Algo Trading Strategy: Types, Risks and How to Choose
An algo trading strategy is a complete set of written rules that software follows to place orders: what to trade, when to enter, when to exit, how much to size, and when to stop. In India, a retail investor can run one only through a stock broker and within SEBI and stock exchange rules. No type of strategy is guaranteed to make money, and this guide names no strategy as a good one.
Risk note, up front: trading carries a risk of loss. This article is education, not investment advice. It contains no signals, no expected returns and no recommendation of any strategy, broker or product.
What is an algo trading strategy, exactly?
Algo trading means trading in which orders come from automated logic. The strategy is that logic, written down in enough detail that a program can follow it without a person deciding each step.
A few terms used below:
- Signal: the condition that tells the program to act, for example "price closed above its recent average".
- Entry and exit: the rules for opening and closing a position, including the exit you use when the trade goes against you.
- Position size: how much each order is for, and how much you hold in total.
- Backtest: running your rules on past data to see what they would have done. It describes the past. It does not predict the future.
What are the main types of algo trading strategy?
The table below groups strategies by the idea behind them. This grouping is our own general explanation, not a ranking, and it is not taken from any exchange or regulator.
| Family | The idea in one line | The main thing to watch |
|---|---|---|
| Trend following | Rules that act in the direction price has been moving, using conditions such as price against a moving average | Sideways markets can trigger repeated entries and exits that lose to costs |
| Mean reversion and pair trading | Rules that assume the gap between two related prices, or between a price and its usual level, will narrow | The gap can keep widening, and the rule must say when to give up |
| Arbitrage (cash-futures, index, calendar spread, conversion-reversal) | Rules that trade price differences between related instruments | Differences can be small and short-lived, so costs, margin and speed matter |
| News-based | Rules that react to a data release or a news event | A late or misread feed makes the rule act on stale facts |
| Execution (for example, splitting a large order into slices over time) | Carries out a decision that was already made | It does not decide what to buy or sell, so it is not a source of returns |
Most real strategies mix parts of these.
For the broker, static IP and exchange rules that apply to retail algo users, see is algo trading legal in India.
How do you choose an algo trading strategy to study?
Start with what you can explain and check, not with what sounds profitable.
- Can you state it in two sentences? If not, you will not know why it fails.
- Do you have the data it needs? A rule that uses data your broker does not supply cannot run.
- Does your broker allow the order types you need? Ask in writing.
- Have you counted costs? Brokerage, taxes and slippage reduce results, and simple tests often leave them out.
- What happens when it fails? Set the loss limit and the stop method before you start.
Then test in stages. How to backtest a trading algorithm without fooling yourself explains why a good backtest is weak evidence, and paper trading vs live trading for algo strategies explains what practice with no real money can and cannot show.
What should you write down before you code a strategy?
This template turns an idea into something a program, or a developer, can follow. Fill in every row before writing any code.
| Field | What to write |
|---|---|
| Market and instrument | The segment and instrument, and why it trades enough to fill your order size |
| Signal | The exact condition, in words, using data your broker supplies |
| Entry | When to enter and which order type your broker allows |
| Exit | The exit for a gain, the exit for a loss, and a time-based exit |
| Size | The maximum per order and per day |
| Stop conditions | Daily loss limit, data-feed failure, rejected orders, and a kill switch you can reach |
| Costs | Brokerage, taxes and slippage assumed in the test |
| Evidence | Which past periods you tested and how long you practised on paper |
For structure only, here is a rule with a blank left on purpose: "Buy when the closing price is above its N-day average, and exit when it closes below it." N is a number you would have to choose and test. This is an illustration of phrasing, not a signal.
What can go wrong with an algo trading strategy?
- Rule risk: the logic is wrong, or it only worked on the period of data you used to design it.
- Technical risk: a lost connection, a bad data feed or an expired login.
- Runaway risk: a loop that keeps sending orders, which is why brokers apply risk checks.
- Vendor risk: a seller who promises returns. SEBI has issued a circular on performance and return claims made by unregulated platforms offering algorithmic strategies (circular SEBI/HO/MIRSD/DOP/P/CIR/2022/117, 2 September 2022). Treat a promised return as a warning sign.
- Credential risk: anyone with your API key can place orders in your account. Keep keys private.
Frequently asked questions
Which algo trading strategy is best for beginners? There is no best strategy, and none is guaranteed to make money. A rule you can explain in two sentences is easier to test and to stop.
Can I use a strategy I bought from someone else? Only through your broker and with care. Ask which broker the orders go through and what the provider's status is.
Do I have to register my own strategy with the exchange? Registration rules for retail algo users depend on broker and exchange requirements; see is algo trading legal in India.
Is a strategy the same as a trading signal? No. A signal is one condition inside a strategy. A strategy also covers sizing, exits, costs and what to do when things break.
Where a custom build fits
Some traders want software built to their own written rules, connected to their broker's API. The home page describes the kinds of systems built, and the FAQ answers common questions. A developer can run your rules consistently. A developer cannot make a strategy profitable.
Limits of this article
Based on the SEBI circular listed below. The family table is a general explanation, not a ranking, and says nothing about how any strategy has performed. It does not cover taxation, later circulars or your broker's terms. Read the current rules before you trade.
Sources
- SEBI, circular SEBI/HO/MIRSD/DOP/P/CIR/2022/117, 2 September 2022: https://www.sebi.gov.in/legal/circulars/sep-2022/performance-return-claimed-by-unregulated-platforms-offering-algorithmic-strategies-for-trading_62628.html
This article is general information, not legal or financial advice. Trading involves risk of loss.