Trade smarter. Forex & gold.Grid and Basket EAs Explained: How They Win, Why They Blow Up, and How to Size One Safely
Grid and basket EAs are some of the most popular robots for sale, because their equity curves often look smooth for months. They are also the type of EA most likely to wipe out an account in a single bad week. Both facts come from the same maths. This guide shows that maths with real gold numbers, so you can see exactly what a basket EA is doing with your money, how big it can get, and which safety limits actually protect you.
Key takeaways
- A basket EA adds trades when price moves against it, then closes them all together at one target.
- It wins often because the average entry follows price, so a small bounce is enough to close in profit.
- Its losses are rare but large: risk grows with every trade added, and fast with a lot multiplier.
- A basket with 6 trades starting at 0.02 lot and growing ×1.4 is 0.33 lot when full: over 16 times the first trade.
- The protection that matters is a hard money stop per basket, a daily loss limit and an account drawdown stop, all set before you start.
- Size the EA from its worst basket, not its average month. Use the calculator below with your own settings.
Grid, basket, averaging, martingale: what the words mean
These words get mixed up in EA adverts, so here is what each one means on this site:
| Word | What it means | Risk level |
|---|---|---|
| Grid | Orders placed at fixed distances (for example every $30 on gold), in one or both directions. | Medium to very high, depends on lot growth |
| Basket | A group of trades in one direction that are managed and closed together, at one combined target or stop. | Depends on how many trades and how big |
| Averaging (cost averaging) | Adding to a losing position so the average entry price gets closer to the current price. | High if there is no hard stop |
| Martingale | Doubling (or more) the lot size after each loss or each new step, hoping one win recovers everything. | Very high: the classic account killer |
Most real products mix these. A “basket EA” usually means: open one trade on a signal, add more trades at a grid distance if price goes against it, make each new trade a bit bigger (a lot multiplier), and close the lot together at one money target. Many gold EAs sold today are basket EAs of this kind. That's why we wrote this guide: we want traders to understand the risk before they buy one, not after.
How a basket wins so often
The trick is the average entry price. Say the EA buys gold at $4,100. Price falls to $4,070 and it buys again, a little bigger. The average entry is now somewhere between the two, closer to the bigger trade. Each new trade pulls the average further down. So when gold finally bounces, it doesn't need to climb back to $4,100. It only needs to reach the average entry plus a small target.
That's why basket EAs often show win rates of 70% to 90%. Most baskets close in profit after a small bounce. The catch is what happens when there is no bounce.
Average entry after four buys
Gold buy basket, start lot 0.02, a new trade every $30, lots grow ×1.4 (rounded to 0.01).
Gold only has to win back about $34 of the $90 fall for the whole basket to break even. Add a small target on top and the basket closes in profit.
The maths of a growing basket
Now the other side. Every trade the basket adds makes the next $1 move against you more expensive. Here is the same basket filling up, step by step, with a new trade every $30 against the first one:
| Trade | Opened at (from first) | Lot | Total lots | Average entry (from first) | Bounce needed to break even | Floating loss at that point |
|---|---|---|---|---|---|---|
| 1 | $0 | 0.02 | 0.02 | $0.00 | $0.00 | $0 |
| 2 | $30 | 0.03 | 0.05 | $18.00 | $12.00 | $60 |
| 3 | $60 | 0.04 | 0.09 | $36.67 | $23.33 | $210 |
| 4 | $90 | 0.05 | 0.14 | $55.71 | $34.29 | $480 |
| 5 | $120 | 0.08 | 0.22 | $79.09 | $40.91 | $900 |
| 6 | $150 | 0.11 | 0.33 | $102.73 | $47.27 | $1,560 |
Two things stand out. The first trade is 0.02 lot, but the full basket is 0.33 lot, over 16 times bigger. And when the 6th trade opens, $150 below the first, the basket is already $1,560 down on paper. If gold keeps falling from there, every extra $1 costs another $33 (0.33 lot × 100 oz).
Why basket EAs blow up
A basket EA is betting that price will come back. Usually it does, which is why the equity curve looks smooth. But gold sometimes moves $100 to $300 in one direction in a day or two: around a central bank decision, a jobs report, a war headline, or a big change in interest-rate expectations. In those moves there is no bounce, and the basket keeps adding size into the move.
| Danger | What happens | What protects you |
|---|---|---|
| One-way trend | The basket fills up and the loss grows by total lots × 100 oz per $1 | A hard money stop per basket |
| News spike or gap | Price jumps past the next steps and past the stop, filled at a worse price | Smaller lots, news filter, wider account buffer |
| Several baskets at once | Two or three baskets open in different sessions all lose together | Max baskets or one basket at a time |
| Bigger spread | Every trade in the basket pays the spread, so costs multiply | A RAW or ECN account with low spreads |
| Hiding the loss | The EA holds losing baskets for weeks, so the balance looks fine while the equity sinks | Watch equity drawdown, not balance |
Here is the pattern to recognise: months of small, steady gains, then one loss that wipes out several months or the whole account. That isn't bad luck. It's the payoff shape of the strategy:
That test shows the shape honestly: 72% of trades won, but the average loss ($92.49) was about twice the average win ($44.54). The strategy still made money in that window because the losses were capped at the basket stop. Remove the cap and the same EA becomes a time bomb.
The safety limits that actually matter
Every basket EA should have these limits, set to real numbers before it ever runs on real money. If an EA doesn't have them, don't run it.
| Limit | What it does | Good sign | Red flag |
|---|---|---|---|
| Max trades per basket | Stops the basket adding forever | A small number (for example 5 to 8) | 20, 50 or 'unlimited' |
| Lot multiplier | How fast each new trade grows | 1.0 to 1.5 | 2.0 or more (martingale) |
| Max lot | Caps the size of any single trade | Set, and small for the account | Not offered |
| Hard money stop per basket | Closes the whole basket at a set $ loss | A fixed $ amount you can afford | 'No stop loss, it always recovers' |
| Daily loss limit | Stops all trading for the day | A few % of the account | Not offered |
| Account drawdown stop | Shuts the EA off at a set % from the high | Set below your pain point | Not offered |
For reference, the basket EA in the test above used all six: a ×1.4 multiplier, at most 6 trades, a 0.20 max lot, a $100 hard basket stop, a 4% daily loss limit and a 20% account drawdown stop. Those caps are what made its losses survivable in testing, and they are the reason it isn't a martingale.
How to size a basket EA for your account
Size a basket EA from its worst case, not from its average month. A simple method:
- Find the worst basket. From the seller's test report or your own backtest, find the largest single basket loss and the maximum equity drawdown. Use the bigger of the two, and check that the test used real ticks.
- Add a safety margin. Live trading has bigger spreads, slippage and gaps. Assume the next worst loss is 1.5 to 2 times the worst one in testing.
- Compare with what you can afford to lose. If that number is more than about 10% to 20% of your account, use a smaller preset, a smaller start lot, or a bigger account.
- Check the full basket against margin. Use the total lots of a full basket in the margin calculator. If a full basket would use more than about a third of your free margin, it's too big.
Sizing check on a $5,000 account
A seller's report on a $10,000 test account shows a 12% maximum equity drawdown ($1,200). You want to run it on $5,000 with half the lots.
24% is more than most people can sit through calmly. Halve the lots again, or start on a cent account until you've seen how it behaves live.
Basket exposure calculator
Put in your EA's settings to see how big a full basket gets, how far price must bounce to break even, and where your basket stop would close it.
For learning only. Brokers differ in contract sizes and rounding, so always check your broker's contract specification.
How to test a basket EA before you trust it
- Backtest with real ticks over at least 12 months that include strong trends, not just ranges.
- Look at the equity line, not the balance line. A basket EA's balance can rise while equity dips sharply under it. The dips are the real risk.
- Find the worst week. Check the tester results for the largest losing trade and the maximum consecutive loss, then look at what gold did that week.
- Test a higher spread. Run the same test with a wider spread. If the result turns negative, the EA needs a low-cost account to work.
- Forward-test on demo or cent for at least a month before using real money.
Our guide to reading an MT5 Strategy Tester report explains every number you'll see in those tests.
Common mistakes
- Judging the EA by its win rate. A 90% win rate means little if the 10% of losses are four times bigger than the wins.
- Turning off the basket stop to 'let it recover'. That's how a capped loss becomes an account loss.
- Raising the multiplier to make it profit faster. Profit grows a little; worst-case loss grows a lot.
- Running two basket EAs on the same symbol. Their baskets can fill up together in the same move.
- Sizing from the average month. Size from the worst basket and the maximum equity drawdown, with a safety margin.
- Trusting a backtest without real ticks. Grid distances and stops depend on exact prices; low-quality data hides the worst fills.
Printable checklist
- I know the EA's multiplier, max trades, max lot and basket stop.
- The multiplier is 1.5 or less, and the basket has a hard money stop.
- There is a daily loss limit and an account drawdown stop, and both are switched on.
- I know the total lots of a full basket and what $1 against it costs me.
- I've sized from the worst basket in testing, times 1.5 to 2.
- That worst case is a loss I can afford.
- I've checked the equity line, not only the balance line.
- I've tested with real ticks, and with a higher spread.
- I've forward-tested on demo or cent before using real money.
Frequently asked questions
Is a basket EA the same as a martingale?
Not always. A martingale doubles (or more) the lot size each step. A basket EA may use fixed lots or a gentle multiplier like 1.4. The lower the multiplier and the tighter the basket stop, the less it behaves like a martingale.
Why do basket EAs have such high win rates?
Because the average entry follows price, so most baskets close after a small bounce. The win rate tells you how often it wins, not how much it loses when it doesn't.
Can a basket EA lose my whole account?
Yes, if it has no hard stop or the stops are set too wide for the account. With a hard money stop per basket, a daily loss limit and an account drawdown stop, the loss is capped, but gaps and fast moves can still go past a stop.
Why does the balance look fine while the account is losing?
Losing baskets stay open, so their loss shows in equity, not balance. Always watch equity and the maximum equity drawdown.
What multiplier is safe?
No multiplier is safe on its own. Lower is safer: 1.0 means fixed lots, 1.2 to 1.5 is gentle, 2.0 or more is martingale. The basket stop and max trades matter as much as the multiplier.
Sources and further reading
- MetaTrader 5 Help: Strategy Tester report
- MQL5 article: evaluating a signal (drawdown and deposit load)
- Investopedia: martingale system
This guide is general education, not financial advice. Broker terms differ; always check your own broker's contract specifications.