1 tool
An Expert Advisor is a program that watches the chart and places trades by a fixed set of rules. It solves two problems: you cannot watch the market all day, and you do not always follow your own plan. If you are new to the idea, read what an Expert Advisor is before you download one.
Every EA here does three jobs. It decides when to enter. It decides where the stop and target go. It decides when to close.
The rules are written into the code. The EA does not get tired, bored or scared after a loss. That is its strength.
It is also its limit. An EA does not know that a central bank meets tomorrow. It does not notice that the market has changed character. It keeps doing what it was told until you stop it.
So treat an EA as a tool that executes a plan. It is not a plan by itself. You still choose the symbol, the timeframe, the risk per trade and when to switch it off.
The EAs are sorted by the type of market they expect. Picking the right group matters more than picking the right inputs.
| Group | Expects | Typical weak spot |
|---|---|---|
| Trend-following EAs | Long directional moves | Choppy, sideways weeks |
| Breakout EAs | A quiet range, then a burst | Fake breakouts that reverse |
| Mean reversion EAs | Price swinging around an average | Strong trends that keep going |
| Grid & basket EAs | Price returning to earlier levels | One-way moves with no return |
| Scalping EAs | Many small moves on low timeframes | Wide spreads, slippage, slow execution |
Notice the pattern. Each group wins in one market condition and struggles in the opposite one. Trend and mean reversion are near mirror images. No EA type works in every condition.
Start with your situation, not with the EA. Use this checklist before you pick a group.
If two groups still fit, start with the one whose losing periods you understand. You are more likely to keep running an EA when its bad weeks do not surprise you.
Testing is where most of the work is. Follow this order.
Example: you backtest an EA on two years of data and it shows a 12% maximum drawdown. Plan for live drawdown to be larger, often much larger. For planning, assume something like 1.5 to 2 times the backtest figure. If 24% of your account would be too painful, cut the risk input until that worst case is something you can sit through.
We describe our own checks on the how we test page.
Over-fitting the inputs. If you tune ten settings until the backtest looks perfect, you have fitted the past. A simple EA with few inputs usually holds up better on new data.
Running several EAs on correlated pairs. Three EAs buying EURUSD, GBPUSD and AUDUSD are close to one large bet against the US dollar. Add up the total risk, not each EA on its own.
Ignoring news. Most EAs do not read the economic calendar. If yours has no news filter, consider pausing it around major releases.
Judging too fast. A trend EA can lose for six weeks and still be working as designed. Compare live results to the backtest's worst period, not to its strongest month.
Pick the group that matches your schedule and risk tolerance. Then run one EA from it on a demo account for a month before you look at any other. One tested EA you understand is worth more than five you do not.