Your EA has 6 parameters. Each has 20 candidate values. That's 20^6 = 64 million combinations.
Your laptop runs one backtest in about 0.03 seconds. 64 million backtests = roughly 22 days of continuous computation. Three weeks with the fan screaming and the computer doing nothing else.
With the MQL5 Cloud Network, you distribute those 64 million backtests across thousands of computers worldwide. Completion time: about 2 hours.
That's not theoretical. That's the actual efficiency gap. And most EA traders don't even know this tool exists.
What the cloud network actually is
A distributed computing network built directly into the MT5 strategy tester. Thousands of computers worldwide contribute idle processing power. You pay a small fee to use that power for parameter optimization.
No extra software to install. No server to configure. No distributed computing knowledge required. If you have an MQL5 community account and the MT5 strategy tester, you can use it.
When cloud computing makes sense
Not every optimization needs the cloud. Local testing handles small jobs fine.
Use cloud when: parameter combinations exceed 500,000, you're running "every tick based on real ticks" mode (10x slower than open prices), you need walk-forward analysis across multiple time windows, or you're testing across multiple currency pairs simultaneously.
Stay local when: under 50,000 combinations, using open price mode, and your EA logic isn't complex. Start it before bed, results in the morning.
Setup in five steps
-
Open the Strategy Tester (Ctrl+R in MT5). Select your EA, symbol, timeframe. In Input Parameters, set start value, step size, and end value for each parameter. Change test mode from "Single Test" to "Optimization."
-
Click the Agents tab at the bottom. Check "MQL5 Cloud Network." Log into your MQL5 community account if prompted.
-
Set a budget cap. Start small — $5 for your first cloud run. Your MQL5 account needs funds loaded in advance.
-
Click Start. Tasks are automatically split into thousands of packages and distributed globally. You can monitor progress, active agents, and spending in real time.
-
Results appear in the same format as local optimization. Analyze them the same way you would any backtest report.
What it costs
Pricing is per computation unit — essentially, how much processing power you consume.
| Scenario | Combinations | Mode | Approximate cost |
|---|---|---|---|
| Simple dual MA EA, 3 params | 8,000 | Open prices | $0.50–1.00 |
| Medium complexity, 4 params | 100,000 | Every tick | $3–8 |
| Complex multi-indicator, 5 params | 3,000,000 | Every tick | $10–15 |
For most traders, a few dollars in cloud fees is cheaper than three days of electricity running your computer at full load. And the time saved is worth more than the money.
How to save money: coarse first, fine second
The biggest cost driver is testing too many combinations at full resolution from the start.
Round 1: use wide step sizes. MA period from 10 to 200 with step 20 = 10 values. Find the approximate sweet spot (say, 60–100 performs best).
Round 2: narrow the range. MA period from 60 to 100 with step 2. Now you're testing 21 values instead of 95.
Two rounds combined: far fewer total combinations than running step 2 across the full range.
Even better: use the genetic algorithm first ($1 or less). It intelligently searches the parameter space using roughly 5–10% of total combinations. Once you see which regions are promising, switch to complete algorithm for a precise search within that region.
A real example from our testing: 5-parameter trend EA on EURUSD H1. Total combinations: 636,000. Genetic algorithm on cloud: 30,000 backtests, $0.80, 12 minutes. Found the optimal region. Local fine optimization of 17,820 combinations in the narrowed range: 45 seconds. Total: $0.80 and under 15 minutes vs 3+ hours running everything locally.
Important considerations
Your EA source code is safe. Only the compiled .ex5 file is sent to cloud agents, not your source code. Each task package is cleared after execution. MetaQuotes provides explicit security guarantees on this.
Results are identical to local runs. Cloud agents use the exact same MT5 backtesting engine. Same algorithm, same data, same numerical output. The only difference is speed.
Cloud computing doesn't prevent overfitting. Making it easy to test millions of combinations also makes it easy to find parameters that perfectly fit history but fail live. After cloud optimization, always run out-of-sample testing and walk-forward verification. Finding the "best" historical parameters is not the goal. Finding parameters with reasonable logic that hold up on unseen data is.
Internet connection matters. Task distribution and result collection happen over the network. Unstable connections can cause interruptions. MT5 supports resume-after-disconnect, but frequent drops reduce efficiency. Use a stable connection, ideally the same VPS that runs your EA.
FAQ
Will I be charged if my balance runs out mid-optimization?
Optimization stops automatically. Completed results are kept. You won't go into debt. This is why setting a budget cap before starting is important.
Can I contribute my own computer to the network and earn money?
Yes. Install the MetaTester agent in MT5 and register as a cloud computing node. Your computer accepts optimization tasks from other traders when idle. The income is modest — enough to cover electricity on an old machine or VPS — but it's passive.
Genetic algorithm or complete algorithm?
Under 100,000 combinations: complete algorithm. It's exhaustive and won't miss the optimal solution. Over 100,000: genetic first for coarse screening, complete algorithm for fine search in the promising region. This is the most cost-effective approach for any serious EA optimization.
Does this work with MT4?
MT4 has limited cloud optimization support compared to MT5. If you're doing serious multi-parameter optimization, MT5 is the right platform. This is one of several reasons we develop and test on MT5 at FXTool. Browse the FXTool marketplace for EAs that come with pre-optimized parameters verified through this process.
About the author: The FXTool team builds and tests MetaTrader trading tools daily. We run every EA we sell on live accounts and publish the results. This guide reflects what we've learned from building 50+ EAs and working with thousands of retail traders.
Forex trading involves significant risk and may result in total loss of capital. This article is for educational purposes only and is not investment advice. Understand the risks and consider your financial situation before trading.