Forex Robot

What a forex robot is

Forex is the market where one currency is exchanged for another. A robot is a program that trades it for you by following written rules. It reads prices, decides by its rules, sends orders to a broker and manages them until they close. It repeats that every minute or every day with no tiredness and no hesitation.

The word "robot" says little about how it works. Sellers use it for very different things:

• An expert advisor (EA): a program that runs inside a trading platform and trades the account it is attached to. This is the most common retail form. See expert advisor.
• A standalone bot: a program outside the platform that talks to the broker through an interface.
• A script or indicator-based alerter: a program that only draws signals, and you press the button. Calling it a robot is a stretch.
• A signal or copy service: a service where somebody else decides the trades. Not a robot at all, although it is sometimes sold as one.

How it differs from a forex EA and a forex trading bot

The three terms overlap and none of them is a standard. In practice: "forex EA" is the narrow term and it names the platform format. "Forex trading bot" is the wider technical term for any automated trading program and the page on it covers what it needs to run. "Forex robot" is the marketing term, used by sellers and by people searching, and it can mean either. When a seller says robot, the first useful question is which of these four it really is, and where it runs.

How to judge claims about a robot

Most claims arrive as a few numbers: a win rate, a gain per year, a drawdown and a picture of an equity curve going up. Each number is incomplete alone. Ask for these:

1. The win rate and the payoff together. A win rate means nothing without the average win and the average loss. See expectancy for the arithmetic.
2. The costs. Was spread, commission and slippage included, on every trade? Costs are a fixed bite out of every trade, so they hit a robot with small wins hardest.
3. The test. Was it a backtest on past data, a test on data the settings never saw, or a live account? A backtest can be tuned until it looks good, see overfitting. A test on later data that the rules never touched is stronger, and live results with a statement from the broker are the strongest.
4. The biggest drop. A gain with no worst-case drop shown is half a picture. Look at max drawdown.
5. How it behaves when it is wrong. Does it use a stop, or does it add to losing trades, as martingale does? Hidden ladders are why the largest loss is so much bigger than the average one.

A worked example: a 90 percent win rate

This is a made-up robot, and every figure is invented. It is advertised with "90 percent winning trades". It wins 10 on an average winning trade, loses 100 on an average losing trade, and the spread plus commission cost 1.50 on every trade. The result per trade at each win rate is the win rate times the win, minus the loss rate times the loss, minus the cost.

Calculated by us on 29 Oct 2026, exact for the stated inputs. At 90 percent, the gross result is -$1.00 a trade, which is already a loss, and after the cost it is -$2.50. Over 500 trades that is -$1250.00. The break-even win rate is the loss divided by the win plus the loss, which is 100 / 110 = 90.9% before costs, and (100 + 1.5) / 110 = 92.3% after. The robot's 90 percent sounds high and is below that line.

Notice the shape. Ten wins of 10 earn 100, and one loss of 100 wipes them. Most days look fine. A long equity curve of small wins, then a single drop, is what this arithmetic produces, and a backtest that ended just before the drop would show only the smooth part. This shape is not unique to robots, but robots make it easy to build, because a program can tune its stop and target to give a high win rate on past data.

How robots fail

• Tuned to the past. The settings fit old data, then the market changes and the fit is gone.
• Costs ignored. The backtest has no spread, or a fixed one, while the live account has a wider one, especially when news is out. See spread and slippage.
• A rare large loss. The robot closes many winners and holds losers. One trending event wipes out many months.
• Rule changes by the broker, such as limits on order frequency, hedging or the size of positions. Rules vary by broker, so read yours.
• The machine goes off. A robot needs a platform, a connection and a computer that stays on. See forex VPS.

How to check before trusting a number

1. Rebuild the arithmetic yourself: win rate, average win, average loss, costs. Put them into the expected value calculator.
2. Run it on a demo account for long enough to see several different market conditions, and compare the live costs with the tested ones.
3. Check that the shown results are a list of trades you can count, not a picture.
4. Find the largest single loss in the record and compare it with the account size.

What this does not tell you

The example shows how to test one common claim. It does not say any real robot is good or bad, and it does not say which win rate or payoff is worth running. A robot with a break-even win rate of 40 percent and a real win rate of 45 percent can also fail, if costs rise or the market changes. The check is a minimum, not a verdict.