The best EA software: how to choose (and where Nebula fits)

The best EA software is the one that can show you how its Expert Advisors performed on price data they were never built on. A backtest on the same history an EA was designed around tells you almost nothing about how it will trade next month. Nebula is EA software for MT4 and MT5 that builds trading systems without code and tests every one of them on in-sample, out-of-sample and blind data before it lets you deploy it.

We make Nebula, so weigh this guide with that in mind. The checklist below applies to any EA software you are considering, including ours.

What is EA software?

An Expert Advisor (EA) is a program that trades automatically inside MetaTrader 4 or MetaTrader 5. "EA software" covers three different things people buy:

• Ready-made EAs: a finished robot, usually one fixed strategy, sold on a marketplace or a website.
• EA builders: tools that let you assemble your own EA from rules or blocks instead of writing MQL code.
• Strategy generators: software that searches through many possible strategies, tests them, and exports the ones that pass as EAs.

Nebula is in the third group. You set the instrument, timeframe, direction and the risk you will accept, and its engine builds and tests candidate strategies from a library of strategy blocks, then compiles the survivors into a ready-to-run .ex4 or .ex5.

How to choose the best EA software: 7 questions to ask

1. How was it tested on data it never saw? This is the question that matters most. Ask for results on an out-of-sample period and, ideally, a blind period that was locked away while the strategy was built. If the only evidence is one long backtest, you cannot tell a real edge from a curve-fit.
2. Can you see the whole record, not just the best bit? You should be able to inspect the equity curve across every test window, the rules the EA trades, and how it was graded. No black boxes.
3. Does it account for real trading costs? Spread, slippage and commission quietly move live results away from the backtest. Good software models them and tells you which settings it used.
4. Can you check its numbers on your own broker? The EA should behave in your MetaTrader Strategy Tester the way the software said it would. Nebula's parity check gives you the exact Strategy Tester settings so you can confirm this yourself.
5. Is it one idea or many? A single fixed strategy is a single bet. Software that can build several systems and check how correlated they are lets you run a book instead of the same trade in different costumes.
6. What happens when the market changes? Every edge fades. Ask how the software finds replacements and retires systems that stop working.
7. Does it respect your risk limits? You should set the risk per trade and the maximum drawdown you will accept, and the software should build within them.

Ready-made EA, EA builder or strategy generator?

Why testing on unseen data matters so much

Give a computer enough history and it will always find a strategy that would have made money on it. That is curve-fitting: the strategy has memorised the past rather than found anything that repeats. It looks flawless on the data it was built from and has no edge anywhere else.

The only reliable defence is to hold data back. Nebula scores every candidate across three windows:

• In-sample: the history the strategy is built on.
• Out-of-sample: later data used to check the strategy, which it was not built on.
• Blind: the most recent stretch, locked away for the whole build and only opened at the end.

Anything that merely memorised the past tends to fail in the blind window, and Nebula drops it. What is left is graded and ranked, with its equity curve across all three windows, before it goes anywhere near a live account.

How many fall away? Across the Nebula network, 304,705 systems have been tested on sealed price history, and 45.0% held up on the sealed window. More than half did not.¹ That is why a single backtest is not enough evidence to buy any EA.

Where Nebula fits

• Builds strategies for you, to your instrument, timeframe, direction and risk limits, with no coding.
• Tests every candidate on unseen data: in-sample, out-of-sample and a blind window.
• Grades and ranks the survivors, with the full record and rules visible.
• Screens your portfolio for correlation, so you are not running one trade ten times.
• One click to MT4 or MT5: compiles each system into an .ex4 or .ex5 for your own broker, with a parity check you can run yourself.
• Runs on Windows and macOS. There is a free tier that needs no card and includes one real deploy.

It is not the right tool for everyone. If you already have a precise strategy and just want it automated, a simple EA builder may suit you better. If you want systems found and tested for you, and you want to see the evidence before you trade them, that is what Nebula is for.

Frequently asked questions

What is the best EA software for MT5?

The best EA software for MT5 is the one that shows how its EAs performed on data they were never built on, lets you inspect the rules, accounts for trading costs and lets you check its numbers in your own Strategy Tester. Nebula builds EAs for both MT4 and MT5 and tests each one on in-sample, out-of-sample and blind data before deployment.

Can I build an EA without coding?

Yes. EA builders and strategy generators such as Nebula create Expert Advisors without you writing any MQL code. In Nebula you set the instrument, timeframe, direction and risk, and it builds, tests and compiles the EA for you.

What is out-of-sample and blind testing?

Out-of-sample testing checks a strategy on data it was not built on. Blind testing goes further: the most recent period is locked away for the whole build and only opened once the strategy is finished, so it cannot have influenced any decision.

Does a good backtest mean an EA will make money?

No. A backtest only shows how a strategy would have behaved on past data, and live results are affected by costs, slippage and changing markets. Results on unseen data are better evidence than a single backtest, but they are still not a guarantee.

Does Nebula work with MT4 and MT5?

Yes. Nebula compiles each system into a ready-to-run .ex4 (MT4) or .ex5 (MT5) Expert Advisor for your own broker's terminal.

Is there a free version of Nebula?

Yes. The free tier needs no card and includes one real deploy. Download Nebula free or see how it works.

¹ Nebula network results, 22 September 2026: 304,705 systems tested on sealed price history; 45.0% held up, meaning a profit factor above 1 on the sealed window. Figures are from back-testing and blind tests on historical data, not live trading. Full results by grade and the method: our results page.

Screens show backtests and simulations. Past performance is not a reliable indicator of future results. Trading involves risk to your capital.