Grid EA Settings for MT4: All 96 Setups Won Over 96% of Baskets, and 40 Blew the Account on at Least One Start Date

A grid EA looks simple in the MT4 inputs window: a lot size, a step, a multiplier, a maximum number of trades and a take profit. The defaults are usually chosen to make the backtest look good. What the inputs do not tell you is how much of the account each combination puts at risk when the market keeps going one way.

So I ran the whole grid of settings on the same bars, from 33 different start dates, and counted which ones survived. I tested the logic behind free robots in my five-robot test. This is the practical follow-up: which grid settings to avoid, which to consider, and how to check any grid EA in MT4 before it touches a live account.

What do grid EA settings in MT4 mean?

The names differ from robot to robot, but almost every grid or averaging EA has these five inputs:

A multiplier of 1.0 is a plain grid: every order has the same lot. Anything above 1.0 is a grid with martingale inside it, and the later orders are the big ones. I compared that lot rule on its own in martingale vs fixed lot.

How were the 96 settings tested?

• Data: 16,925 EURUSD M5 bars, June 3 to August 26, 2026.
• Robot: the first order goes on the side of the 50 EMA. Each time price moves one step against the last order, another order opens with the multiplied lot, until the maximum is reached. The whole basket closes at the target. There is no stop loss, as in most free grid EAs.
• Account: $1,000, 1:100 leverage, stop-out when equity falls below 50% of the margin in use, checked at the worst price inside every bar.
• Cost: 1 pip per order, from my spread and swap test. No overnight swap was charged, so real results would be somewhat worse for baskets held for days.
• Start dates: every setting was run from 33 start dates, every second trading day, plus one full run from June 4.

Which setting mattered most?

The lot multiplier. Each row below covers 24 settings (4 steps × 3 order limits × 2 targets):

$1,000 start. “Average median end” is the median end balance of each setting across its 33 starts, averaged over the 24 settings.

This is the trade every grid EA makes for you. A bigger multiplier made more money on the typical start date and lost more on the bad ones. With 2.0, the best setting in the test, step 10 with a 10-pip target, had a median end of $1,598.41 and won 99.4% of its baskets. It was also stopped out on 5 of 33 starts and its worst start ended at $318.59.

The flat-lot grid was the opposite: two of its 24 settings blew up, and the other 22 made very little. With a 10-pip target the 11 flat grids that never blew up ended with medians between $1,003.32 and $1,021.03, about 1% in twelve weeks, and their worst starts still dropped to between $463 and $876.

What grid step is safest?

Wider. Each row covers 24 settings:

A narrow step fills the basket on a small move. Even the flat-lot grid with a 5-pip step and no order limit was stopped out on 17 or 18 of 33 starts, depending on the target, because it kept adding orders all the way down: up to 61 of them in the full run.

Does a maximum-orders limit help?

Yes. Each row covers 32 settings:

The limit works because it caps the total lot. It does not stop the loss: once the basket is full, the robot just holds it with no stop and waits for price to come back. In this sample price came back often enough. The 5-pip basket target also did slightly better than 10: 31 of 48 settings never stopped out against 25 of 48.

How far can price move before the account is cut in half?

This is the number to work out before running any grid EA. For each setting, I calculated how many pips price has to move against the first order, with no pullback, before the open basket loses $500 of a $1,000 account:

0.01 first lot, $10 per pip per lot, 1 pip cost per order, margin ignored. With large baskets the stop-out comes sooner than this.

Now compare those numbers with the market. In this data the median EURUSD day moved 48.4 pips from high to low, one day in ten moved 95.3 pips or more, and the largest day moved 137.4. Over five trading days the median range was 101.9 pips and the largest 201.8.

So a 10-pip grid with a 2.0 multiplier and 10 orders loses half the account on a move smaller than an ordinary week. A flat-lot grid with 5 orders needs a move of over 1,000 pips, which did not happen in these twelve weeks. That is the whole difference between the rows of the first table.

Which settings survived every start date?

6 of the 96 never stopped out and ended in profit on all 33 starts. They still had bad moments in the full run from June 4:

Two of the six “safe” settings dropped below half the account at some point and recovered. They were not safe. The market simply came back before the broker closed the basket. All six use a step of 20 or 30 pips, and five of the six use the 5-pip basket target. None has a step of 5 or 10.

How do you check a grid EA’s settings in MT4?

1. Open the inputs before you switch anything on. In MT4 press Ctrl+N to open the Navigator, drag the EA from Expert Advisors onto a chart, and open the Inputs tab. If you have not installed a file in MT4 before, my MT4 install guide shows where the files go; EAs go in MQL4/Experts instead of MQL4/Indicators.
2. Write down the five numbers: start lot, step, multiplier, maximum orders and basket target. If there is a stop loss input, note whether it is set to 0. Zero usually means no stop.
3. Work out the full basket. Add up the lots: start lot × (1 + m + m² + … ) up to the maximum orders. With 0.01, ×2.0 and 8 orders that is 2.55 lots. Multiply by $10 per pip and you know what each further pip costs when the basket is full.
4. Compare with the table above. If your setting loses half the account on a move smaller than a normal week, the win rate on the sales page does not matter.
5. Backtest from several start dates. In the Strategy Tester (Ctrl+R), tick “Use date” and run the same settings from at least five different start days, using “Every tick”. One date can be the lucky one, as the grid in my five-robot test showed. My guide to backtesting an EA in MT5 covers the same checks on the newer platform.
6. Read the maximal drawdown, not the profit. In the report, “Maximal drawdown” and “Relative drawdown” show the lowest point. That is the number that decides whether the robot survives the next trend.
7. Run it on demo first. Then press the AutoTrading button in the toolbar only when you have seen it handle at least one strong one-way day.

What should you watch out for?

• Swap. Baskets can stay open for days. Every order pays overnight swap, which this test did not charge.
• News. The spread widens around releases, exactly when grids add orders. My news release test shows how large those moves are.
• Different periods. Twelve weeks of EURUSD is one market. A setting that survived here can meet a stronger trend next month. Settings tuned to one period often fail the next, as my in-sample vs out-of-sample test showed.
• Bigger accounts do not fix it. A larger deposit with the same start lot moves the stop-out further away. It also means more money is lost when it arrives.

Frequently asked questions

What are the best settings for a grid EA in MT4?

In this test, a grid step of 20 or 30 pips, a 5-pip basket target and a maximum of 5 orders were the most robust. Only 6 of 96 settings survived and made money on all 33 start dates, and all six used a step of 20 or 30 pips.

What lot multiplier should a grid EA use?

1.0, if survival matters most. 22 of 24 flat-lot settings never stopped out, against 5 of 24 with a 2.0 multiplier. The 2.0 settings made more on the typical start date and had the worst single result, $163.45 from $1,000.

Why does my grid EA have a 99% win rate?

Because it closes baskets at a small profit and never closes them at a loss. All 96 settings in this test won at least 96.3% of baskets, including the 16 that were stopped out in the run from June 4.

What does MaxTrades do in a grid EA?

It caps how many orders the basket can hold, and with it the total lot. With a 5-order limit, 27 of 32 settings never stopped out. With no limit, 13 of 32.

How much money do I need for a grid EA?

Work it out from the settings. Add up the lots of a full basket, multiply by $10 per pip, and check how many pips against the basket cut your account in half. With 0.01 lots, a 10-pip step and a 2.0 multiplier, a $1,000 account loses half on a 79-pip move.