The essentials
- A trading bot is software that places orders when pre-set conditions are met; it follows rules, it does not predict.
- Backtests are easily flattered by over-fitting and optimistic cost assumptions; expectancy after costs is the number that matters.
- Bots sold with guaranteed or very high monthly returns are a common scam format.
A forex trading bot is a program that watches prices and places, manages and closes trades according to rules written in advance. It removes hesitation and fatigue from execution. It does not remove risk, and it does not know anything about the future that its rules do not already contain.
How a forex trading bot runs
Most retail bots run inside a trading platform. In MetaTrader they are called Expert Advisors; other platforms have their own equivalents or allow connections through an API. The bot receives each new price, checks its conditions (for example, a moving-average crossover, a breakout of a range, or a time of day), and sends orders to the provider when they are met. It can also set stop-losses, move them and size positions.
Because the program only works while it is running, many users host it on a virtual private server (VPS), a rented computer in a data centre that stays online around the clock. A VPS reduces the risk of a home internet or power cut leaving a position unmanaged. It does not make a strategy better.
What a bot cannot do
A bot cannot adapt to conditions its rules never anticipated, read the meaning of unexpected news, or decide that today is a day to stay out. It cannot make a losing strategy profitable by trading faster or more often; with costs on every trade, frequency usually makes a weak edge worse. And it cannot remove the need for the account holder to understand what is being done in their name.
Types of automated strategy
- Trend-following: buys strength and sells weakness, accepting many small losses in exchange for occasional large gains.
- Mean reversion: bets that stretched prices snap back; tends to win often but can suffer large losses when a trend persists.
- Grid and martingale systems: add to losing positions at fixed intervals or in growing sizes. They can produce smooth equity curves for months, then a single sustained move can wipe out the account.
- Scalping bots: aim for very small gains many times a day, which makes them extremely sensitive to spreads, commissions and slippage.
Why backtests flatter
A backtest runs the rules over past prices. It is necessary but easily misleading.
- Over-fitting. Adjust enough settings and any rule set can be made to fit the past perfectly. The more parameters, the less the result means.
- Cost assumptions. A test with fixed tight spreads and no slippage overstates results, especially for short-term strategies.
- Data quality. Gaps, missing weekends or unrealistic fills in historical data change results.
- Regime change. A strategy tuned to a calm, trending period may fail when volatility or interest rates shift.
Testing on data the strategy was not tuned on, and then on a demo account in real time, catches some of these problems.
The number that matters: expectancy
A high win rate sounds persuasive but says little on its own. What matters is the average result per trade after costs:
Illustrative example: a bot wins 60% of its trades with an average win of $50, and loses 40% with an average loss of $90. Expectancy is (0.6 × $50) − (0.4 × $90) = $30 − $36 = −$6 per trade. Despite winning most of the time, it loses money steadily. Size each trade with the position size calculator and add realistic costs from our fees guide before trusting any figure.
Red flags in bot marketing
- Guaranteed returns, fixed monthly percentages or claims that the bot "never loses".
- Performance shown only as screenshots, or on an account you cannot verify independently.
- A requirement to deposit with a specific, often unregulated, provider for the bot to work.
- "AI" or "quantum" branding with no explanation of the actual rules.
- Requests to hand over account login details so someone else can "run the bot for you".
Many investment frauds now use automated-trading language as a hook. Our guide on how to spot trading scams describes the pattern.
Watching a bot that is running
Automation does not remove the need for oversight. Software can fail quietly: a lost connection, a platform update, a change in the provider's contract specifications or an unusual price spike can all make a bot behave in ways its author never tested. Check open positions and logs regularly, set a maximum loss for the account at which you will switch the bot off, and know how to close everything manually if the program stops responding.
If you want to experiment
Start on a demo account, keep a log of every trade and every change to settings, and run the bot long enough to see it through different market conditions. Understand every rule before it trades real money. If you cannot explain why a strategy should work, the backtest is not a reason to believe it will.
Reader questions
Do forex trading bots make money?
Some strategies have periods of profit, and many lose money after costs. No bot can guarantee returns, and past results, especially in backtests, often overstate what happens live.
Do I need a VPS to run a trading bot?
Not necessarily, but a bot only works while the platform is running. A VPS keeps it online if your own computer or connection goes down. It adds a cost and does not improve the strategy.
Is it legal to use a trading bot?
In most places, yes, subject to your provider's terms. Selling bots with misleading performance claims, or managing other people's money through them without authorisation, can break the law.