Black Box Trading Strategy (Algo, Backtest, Rules, Settings)
In financial trading, as with most aspects of life, everything is getting automated, thanks to advances in artificial intelligence and machine learning. From vital signs in cardiology practice to price changes in the financial market, people are building systems that track relevant data and automatically effect the right actions based on changes in the tracked data. In financial markets, this is known as black-box trading. Wondering what is a black box trading strategy?
The black box trading strategy is a method of trading where a computer program monitors the markets to spot trade signals, initiates buy or sell orders, and manages the trades based on its pre-programmed logic. It is an automated form of trading whereby the trader has no direct input on the trade-to-trade decision-making, as the trading algorithms do that for them.
In this post, we take a look at the black box trading strategy, before we finish the post with a backtest.
Related reading:
- Many different trading systems for sale, and
- A free trading strategy list (The link contains access to hundreds of backtested trading strategies)
What is a black box in trading?
In technical terms, a black box is a device or system that uses data inputs to produce outputs without revealing any information about its internal workings. It uses computer programs or algorithms that are based on a given logic to process and execute its function. So, its workings are not known except to the person who set it up. This is where the name black box comes from — the method of operation is opaque.
A black box in trading refers to a system that monitors the market, executes trades, and manages them without revealing the rationale behind its trading decisions. In other words, the exact specifications of a black-box trading system are not known, except to the person who created it. In trading, a black box is also known as a trading algorithm or simply a trading algo.
The source file, often kept by the developer, consists of lines of intricate programming code that define its operations, and those are governed by specific trading rules and guidelines. But the executable file that runs on the trading servers and computers simply executes the logic coded inside of it, which an outsider cannot know.
Every black-box algorithm starts with a trading strategy, which is translated into computer coding language and loaded onto a trading platform that plugs it into the market. No matter how intricate the code or robust the platform is, a black-box system scans the market to generate trading signals. In addition to generating potential buys and sells, the system will also enter and close orders based on pre-programmed logic. Every black box system is unique, proprietary, and protected from public scrutiny.
Essentially, a black box in trading is a private trading system that uses computer programs to generate buy and sell orders and manage trades using pre-programmed logic. The main objective of every black-box system is to establish and maintain a quantifiable edge in the market. An edge is the method by which a strategy or system regularly wins market share. Depending on the type of trading and methods used in its design, it could be either very complex or very simple.
The black box trading system can be used to automate the entire trading process or just to manage already executed trade orders. It used to be available only to institutions and wealthy traders. Some hedge funds and pension funds use the black box system in order to help them manage their trades.
However, as technology evolved, black-box systems have come online to the masses, and no longer must one be a large institution to have access to a trading black box. Individual retail traders can now get and use them, and there are numerous options readily at their disposal — they can code it by themselves or pay programmers to code for them.
What does a black box trading strategy mean?
The black box trading strategy is a method of trading where a computer program monitors the markets to identify trade signals, initiates buy or sell orders, and manages the trades based on its pre-programmed logic. It is an automated form of trading whereby the trader has no direct input on the trade-to-trade decision making, as the trading algorithms do that for them.
The strategy got the black box name because the exact specifications of a trading system are typically shrouded in secrecy. The lines of code that govern its operations and the specific market situation it exploits are not known to an outsider — only the creator knows how it works and keeps that a secret, which is why each system is unique and proprietary. However, no matter who created a black box trading strategy, the system is usually designed to perform these three tasks:
- Identify trade setups: Every black box is based on a trading strategy with its unique way of identifying trade setups, which could be based on momentum oscillators, market reversal points, or trend-following breakouts. This is coded into a logic for the black box’s algorithm. In accordance with the pre-programmed logic, the black box algorithm scans a selected market or markets for trade setups, and upon the defined criteria being met, a trade signal is created. Each signal is a prompt for the automated system to enter a specific market.
- Execute trades: Once a trade setup is spotted and a signal created, the black box’s execution logic automatically places orders at the market in accordance with pre-defined parameters — position size and so on.
- Manage trades: If a trade order is executed, the newly opened position is managed by the system’s automated framework, with reference to the type of order, market direction, stop losses, and profit targets. There are different trade management strategies that can be employed, with a few examples being trailing stops, scalping, and scaling. The system often includes a risk management method, and can align risk to reward on a trade-by-trade or account equity basis. In more complex black boxes, portfolio optimization, position sizing, and hedging strategies may be integrated into trading operations.
One of the benefits of the black box trading strategy is that a computer isn’t subject to human emotions and cannot be affected by emotional biases that are inherent in discretionary trading.
In addition, a computer can do millions of calculations per minute, which a human being is just not capable of. Moreover, the computer doesn’t tire; it can stay online 24/5 and keep monitoring the markets and processing tens of thousands of trades per day. It does not miss any trades. A human being cannot watch the market all the time (they must sleep at some point), and as such, would miss a lot of trades. We have repeatedly said that automation is power, but you need to be careful.
The black box trading strategy is also referred to as automated trading, algorithmic or algo trading, or quantitative or quant trading. Quant trading often involves big data analytics and the use of complex mathematical models. Black box trading is the basis of high-frequency trading (HFT) that is employed by big prop trading firms.
Which algorithms are black boxes?
Algorithms are a set of complex instructions or rules that a computer needs to follow when solving a certain problem. In the case of a trading algo, the problem is to identify trade setups, execute trades, and manage them. Algorithms are extremely literal in the sense that they simply do exactly what they are asked while ignoring any other, important, consideration. You cannot influence how they work unless you have access to the codes.
Thus, any trading algorithm is a black box to an outsider. Whatever happens inside the trading algorithm is only known to the person that created it or a person who has the source code. In line with this, any trading algo that you buy and that does not come with the source file (comes with only the execution file) is a black box because you cannot know how it generates its trade signals and manages trades.
Related Reading: Program Trading
Why are algorithms called black boxes?
Algorithms are called black boxes because whatever happens inside an algorithm is only known to the person or persons that created them and those who have access to the source code. Any other person, including the users who do not have access to the source code, cannot know how it works and the logic it is based on.
An algorithm simply takes data input to execute an output. The user may know the data input and can see the output, but they cannot know how the algo generates the output from the input unless they get the source code to study the logic. Thus, on the basis of how they operate, algorithms are black boxes.
What is a black box machine learning algorithm?
A black box machine learning algorithm is the phrase used to describe machine learning models that offer you a result or make a choice without explicitly stating or demonstrating how they did it. What internal procedures were applied and how different considerations were weighted are not known.
To put it another way, it is a system that lacks transparency because no human — not even the programmers and admins of the machine or algorithm — knows or understands how the output was reached. Only the algorithm itself is aware of exactly how the decisions were made. This type of system is often used in fraud prevention. It can give the user a fraud score without telling or showing how that score was reached.
How do you trade with a black box?
You can create your own trading algo if you know how to code. But if you want a black box system that you cannot influence the logic, you can buy one that comes with only the execution file. When you get the algorithm, here are the steps to follow to trade with it:
- Set it up: You can set it up on your home computer, but you must make sure you have the computer powered all the time and your internet must be good. Any downtime means missing some trade setups. Alternatively, you can secure VPS hosting and have your black box stay online at all times.
- Trade with a demo account initially: It is best to trade with a demo account first before putting your hard-earned money on the line.
- Go live if good enough: If the black box system is consistently making money, you can go live, but risk less than 1% of your capital per trade.
- Monitor it: Check on things from time to time to be sure it does not malfunction and drain your account.
What is Black Box Algorithm – Definition and Examples
A black box algorithm is one where the user cannot see the inner workings of the algorithm. How the system works is kept a secret probably as a sort of security and safety measure to avoid data leaks and unfair competition. However, due to the secrecy they contain and the lack of transparency, some consider it controversial.
An example is an algorithm used in fraud prevention; it can give an output without showing how the output was generated. Some trading bot vendors also sell only the execution file, making such a black box system.
What is a black box model in AI?
Black box AI is any artificial intelligence system whose inputs and operations are not visible to the user or another interested party. A black box, in a general sense, is an impenetrable system, and most AI by nature are set up with the black box model.
For example, deep learning modeling is typically conducted through black box development, with the algorithm taking millions of data points as inputs and correlating specific data features to produce an output in a self-directed process. Often, the data scientists, programmers, and users don’t know how it arrives at its output.
What is black box calculation?
This refers to how a black box system arrives at its output. The calculation is often not known to the user, and in some models, even the developers that created the model may not know how the system does it. This is often the case in some machine learning systems that generate and use their own data to process further information which becomes the input data for yet further calculations.
Black box trading strategy backtest
We are not using any black box systems ourselves. The main reason is this: we are afraid of curve-fitted trading strategies. When you are using any form of computing power to generate signals you must be aware of the risks. If you are very experienced, such a tool can be very powerful. But in the hands of the amateur, we believe it can do more damage than gains.
Thus, because we are not using any black box trading strategies, we are not making any backtest.