White Box vs Black Box Algos: What’s the Difference?
Blog by Aliceblue News Desk
30 October 2026
9 minutes
You come across an algo that promises steady monthly returns. You ask how it picks its trades, and the answer is “that’s proprietary.” Do you go ahead?
That is the question behind white box vs black box algos. A white box algo shows you its rules, so you can read and repeat them. A black box algo does not. SEBI treats the two differently, and this guide explains how, with Indian examples.
Table of Contents
Quick Answer
A white box algo is an automated trading strategy with visible rules, like “buy when RSI falls below 30.” A black box algo keeps its rules hidden from the user. SEBI’s February 2025 circular sorts retail algos into these two types. Black box providers must register as Research Analysts and keep a detailed research report for each algo.
What Is Algo Trading?
An algo is a set of instructions. You, or a provider, write the rules, and a computer places orders when those rules are met. Even “buy this stock if it crosses ₹500” counts.
In India, retail algo trading goes through your broker. SEBI’s circular makes the broker the principal and the algo provider its agent, so the broker handles complaints and monitors the API.
What Are White Box and Black Box Algos?
SEBI asks one question: can the user see the logic and repeat it? If yes, it is a white box algo, which SEBI also calls an execution algo. If the user cannot see the logic and cannot repeat it, it is a black box algo.
A moving average crossover is a typical white box example. It follows a written rule, so you can check every order against that rule. A black box gives you the trade but not the reason for it.
You may hear sellers use the word “grey box.” It is not a SEBI category. Ask the provider which of the two official types the algo is, and get the answer in writing.
How Does It Work?
- Live prices arrive from the exchange.
- The algo checks its rules. In a white box, you can read them. In a black box, you cannot.
- A signal appears: buy, sell, or wait.
- The order goes to your broker with a unique identifier from the exchange, so it can be traced back to the algo.
- Risk controls apply, such as a stop loss and an off switch. SEBI’s circular also keeps a kill switch with the exchange, which can halt orders from a particular algo.
Example With Numbers
This is a made-up example with ₹1,00,000. It leaves out brokerage and taxes.
White box. The published rule says buy Infosys when its 20-day average price moves above its 50-day average. Stop loss is 3% below entry and target is 4% above. The signal fires at ₹1,500, and the algo buys 66 shares for ₹99,000.
Result | Exit price | Per share | Total |
Target hit | ₹1,560 | +₹60 | +₹3,960 |
Stop loss hit | ₹1,455 | −₹45 | −₹2,970 |
Suppose five trades in a row hit the stop loss. You can open the rule and ask whether the stock has stopped trending.
Black box. A seller on Telegram posts screenshots of 18% returns, with no rules and no registration number. You invest ₹1,00,000 and are down ₹4,200 after a month. You cannot tell if that is normal for the strategy or a sign it has broken. Ask for the Research Analyst registration number first, and treat any guaranteed return claim as a warning.
SEBI Algo Trading Rules for Retail Traders
SEBI first set these rules in its circular of 4 February 2025, and the full framework applies to all brokers from 1 April 2026. If you trade through a broker API, this is what it means for you:
- Broker route only. The broker is the principal, and open APIs are not allowed. Access works only through a client-specific API key and a static IP that your broker whitelists.
- Static IP. This is an internet address that does not change. NSE’s standards make it mandatory for API access, so orders from any other address are not accepted.
- Two-factor login. Brokers must use two-factor authentication for API access. Alice Blue’s guide describes it as a daily login.
- Algo ID. Every algo order must carry a unique identifier from the exchange.
- Order speed limit. NSE sets the threshold at 10 orders per second per exchange, and exchanges can change it after notice. Below it, you do not register your own algo. Above it, you must register it with each exchange, through your broker.
- Family use. An algo you built yourself and registered can be used for your family only, meaning self, spouse, dependent children, and dependent parents.
- Charges. Providers and brokers can share subscription charges and brokerage, but all charges must be disclosed to you clearly.
For providers, black box algos need Research Analyst registration and a research report for each algo. Any change in the logic means registering it again as a fresh algo. One explainer adds that a white box algo is registered with the exchange once and can then be offered to many users.
Key Features
- White box: visible rules that you can repeat and test.
- Black box: hidden rules, a provider with Research Analyst registration, and a research report for each algo.
- Both: exchange tagged orders, a broker who is responsible, and your own choice of capital and risk limit.
Advantages
- A white box teaches you something. You can test the rules on old data, see why each trade happened, and notice when the rules stop working.
- A black box suits people who want trades placed without managing the logic. Keeping it secret does not make it better, though. It only means you cannot check it.
Risks and Limitations
SEBI’s FY26 study found that 87.7% of individual traders in equity derivatives lost money, and options made up about 92% of the losses. The study covers F&O trading in general. It does not compare white box with black box, so neither type can claim a better record.
Money risks to know:
- Overfitting. A strategy tuned so closely to old data that it fails on new data.
- Costs. Brokerage, taxes, and slippage (a worse price than expected) eat into small gains.
- Sideways markets. Trend rules keep taking small losses when prices go nowhere.
A black box adds a problem of its own, which is that you cannot judge its risk before you start. You also still decide how much capital goes in. And then there is behaviour: switching strategies after two losses, or adding money after one good week.
White Box vs Black Box Algos: Differences
Parameter | White box | Black box |
Can you see the logic? | Yes | No |
Can you repeat it yourself? | Yes | No |
SEBI category | Execution algo | Black box algo |
Provider duties | Exchange empanelment and registration | Same, plus Research Analyst registration and a research report for each algo |
If the logic changes | Follows the exchange’s approval process | Registered as a fresh algo |
Can you check why a trade happened? | Yes | No |
Main worry | Rules may stop working | Risk cannot be judged or verified |
Suits | Beginners | Experienced users who can verify the provider |
Tips for Beginners
- Start with a white box strategy.
- Use the smallest size you can, such as one lot or 5 to 10% of your trading capital, and judge it on your own results over several weeks.
- Look at maximum drawdown before returns. It is the biggest fall from a peak.
- With a black box provider, ask for the Research Analyst registration number and search it on SEBI’s website.
- If a provider cannot explain how the strategy decides, treat it as a black box.
- Ask what period the backtest covers and whether it includes a bad market.
- Check which charges the provider and your broker share, since SEBI requires clear disclosure.
- Keep a one-step way to switch the algo off.
Common Mistakes to Avoid
- Trusting screenshots. They say nothing about risk. Rely on the registration number and your own small-size results.
- Using an algo for people outside your family. A self-built algo registered with the exchange can be used for self, spouse, dependent children, and dependent parents, and no one else.
- Running from a changing IP address. Orders from an address your broker has not whitelisted are not accepted.
- Assuming a registered broker makes every algo approved. The broker needs the exchange’s permission for each algo.
- Changing rules during a losing streak. Pick a review point, such as every 30 trades, before you start.
- Leaving out costs. Add brokerage and taxes before you judge a strategy.
Who Should Use White Box or Black Box Algos?
White box suits beginners, swing traders, and anyone who wants to understand each trade. Black box suits experienced traders who can verify a provider and are fine with not seeing the logic. If you are still learning order types and stop losses, learn those first.
Conclusion
White box algos show their rules, and black box algos keep them private. SEBI adds extra duties for black box providers, such as Research Analyst registration and a research report for each algo. Neither type promises profit, and the risk stays with you. For most beginners, a white box is easier to follow, test, and stop. If you try one, start with a small amount, check the provider’s registration, and treat any promise of guaranteed returns as a warning. A small start shows you how an algo behaves with real costs and sideways markets.
Frequently Asked Questions
Blog by Aliceblue News Desk
