Forex EA Setup Guide: The First 90 Days That Actually Matter

You bought a forex EA. Now what? The purchase is the easy part. What you do in the next 90 days determines whether that EA generates returns or joins the graveyard of abandoned trading bots on your hard drive. Most EA traders make critical mistakes in the first month: going live too fast, using full position sizes on day one, or turning off the EA during its first drawdown. This step-by-step forex EA setup guide covers the complete 90-day process from installation to the scale-or-kill decision that separates successful EA traders from the ones who cycle through products endlessly. This applies to any EA from any vendor. Think of it as your post-purchase standard operating procedure.

Phase 1: Demo Testing (Weeks 1-2)

The only acceptable response to "I just bought an EA" is: install it on a demo account. Not a live account. Not even a cent account. A demo account with your intended broker's trading conditions. This costs you nothing and gives you everything you need to verify the EA works as advertised before real money enters the equation.

Week 1: Installation and Basic Verification

Your first week is purely mechanical. You are confirming the EA functions correctly in your environment.

• Install on a demo account that mirrors your intended broker's conditions (spread type, leverage, account currency). The demo should match what your live account will look like.
• Verify the EA actually places trades. This sounds obvious, but compatibility issues between MT4/MT5 versions, missing DLL permissions, and incorrect chart timeframes are common problems that prevent EAs from functioning.
• Check that parameters match the vendor's documentation. Default settings should correspond to what the vendor described. If they documented specific magic numbers, lot sizes, or risk percentages, verify those are loaded correctly.
• Monitor the Experts tab for errors. Open the Terminal window (Ctrl+T), click the Experts tab, and watch for red error messages. Common issues include "OrderSend failed," license validation errors, or connectivity problems.
• Note the first few trades: entry price, exit price, timing, direction, and lot size. You are building a reference point, not evaluating performance yet.

Week 2: Performance Baseline

With a week of trades logged, you can start building a baseline that will become critical later when you evaluate live performance.

• Compare demo results against the vendor's verified track record if they have one. Look for Myfxbook, FXBlue, or other third-party verification. If the vendor has no verified results, that is a data point worth noting.
• Do not expect identical results. Execution differs between accounts, brokers, and even VPS locations. You are looking for the same general behavior: similar trade frequency, similar direction bias, similar average trade duration. Not pip-for-pip matches.
• Track these baseline metrics:
  • Average trade duration (minutes, hours, or days)
  • Win/loss ratio across all trades
  • Typical position size relative to account balance
  • Daily and weekly trade count
  • Average profit and average loss per trade
• Document everything. A simple spreadsheet works. You will need this baseline in Month 2 when comparing demo behavior to live execution.
• If the EA does not trade for 2 weeks, that might be normal. Some strategies are highly selective and wait for specific market conditions. Check the vendor's documented trade frequency before assuming something is broken.

Forex EA Setup: Infrastructure (Weeks 3-4)

While your EA continues running on demo, use weeks 3 and 4 to build the infrastructure that will support live trading. Do not rush this. A poorly configured infrastructure will undermine even the best EA.

VPS Configuration

You cannot run a live EA on your home computer reliably. Period. One Windows update at 3 AM, one power flicker during a volatile session, one internet drop while a trade is open. A VPS eliminates all of these risks for a predictable monthly cost.

• Set up a VPS with proximity to your broker's servers. If your broker's servers are in London, your VPS should be in London or nearby. If they are in New York, your VPS should be on the US East Coast. Proximity directly affects execution speed.
• Typical cost: $15-$50/month depending on provider and specifications. For a detailed breakdown of all running costs, read the full cost analysis of running a Forex EA.
• Move the EA from your local demo to VPS demo. Install MT4/MT5 on the VPS, configure the EA, and verify it works identically to your local installation. Run both in parallel for a few days if needed.
• Test the VPS stability. Check the Journal tab daily for disconnection events. A good VPS should show zero or near-zero reconnections per week.

I have written extensively about what happens when VPS goes wrong. Read how a VPS mistake cost $2,300 to understand why this is not an area to cut corners.

Broker Account Setup

With your VPS running, it is time to prepare your live trading account. Prepare it, not deploy on it.

• Open a live account with the appropriate account type. For most EAs that trade frequently, a raw spread or ECN account type will save you money over time compared to standard accounts with wider spreads.
• Fund with your planned amount but do NOT deploy the EA yet. Let the account sit funded while you complete the remaining setup.
• Verify these settings match your demo:
  • Leverage setting (same as your demo test)
  • Account currency (mismatched currencies affect lot calculations)
  • Swap conditions (check the broker's swap table for your EA's traded pairs)

For small accounts just getting started with live EA trading, Pepperstone offers low barriers to entry with competitive conditions. For larger accounts where raw execution quality matters most, IC Markets provides institutional-grade spreads.

News Filter and Protection

High-impact news events can cause extreme volatility, slippage, and spread widening. If your EA does not handle these well, a single NFP release can wipe out weeks of gains.

• Configure the news filter if your EA supports one. Most quality EAs include a built-in news filter or at least a parameter to pause trading around scheduled events. For a deeper understanding of why this matters, read why EAs freak out during news events.
• Set up notifications: email alerts, push notifications to your phone, or Telegram alerts for every trade opened, closed, and any errors encountered.
• Establish daily loss limits if your EA or platform supports them. A hard stop on daily losses prevents catastrophic single-session damage while you are still learning the EA's behavior.

Phase 3: Micro-Live (Month 2)

The goal of Month 2 is not profit. It is verification. This is the most critical phase of your forex EA setup — you are confirming that live execution matches the demo behavior you documented in Phase 1. Profit is a side effect, not the objective.

Going Live with Minimum Size

• Deploy the EA on your live account with the smallest possible position size. If the EA normally uses 0.1 lots, start with 0.01 lots. If it uses risk-based sizing at 2%, set it to 0.2%. You want enough trades to gather data, with minimal financial exposure if something goes wrong.
• The purpose is verification, not performance. You are checking that live execution matches demo behavior. The lot size is almost irrelevant at this stage.
• Expected differences from demo: slightly wider spreads during volatile periods, occasional slippage on entries and exits, minor timing variations of a few seconds.
• Unexpected differences to investigate immediately: completely different trade entry prices, trades appearing on live that never appeared on demo (or vice versa), consistently large slippage exceeding 1-2 pips, or the EA behaving as if it has different settings.

The Deviation Log

Keep a simple log comparing your live trades to what the demo showed during the same period. If you are still running the EA on both demo and live simultaneously, this comparison becomes straightforward.

Track these for every trade:

• Entry price difference between demo and live
• Slippage amount and direction
• Spread at the time of execution
• Any trades that appeared on one account but not the other
• Time difference between demo and live execution

After 30 or more trades, you will have enough data to assess execution quality with confidence. In my experience onboarding dozens of EAs, a well-configured VPS with a quality broker should show less than 0.5 pip average deviation from demo. Anything consistently above 1 pip needs investigation: either the VPS latency is too high, the broker's live execution differs significantly from demo, or there is a configuration mismatch.

If your deviation log reveals consistent problems, do not scale up. Diagnose and fix the issue first. Common fixes include: changing the VPS location, switching to a raw spread account type, or adjusting the EA's slippage tolerance parameter.

Phase 4: The Scale or Kill Decision (Month 3)

After 60 or more days combining demo and micro-live data, you have enough information to make a real, data-backed decision. This is the moment where most traders either make the right call or begin an expensive cycle of switching EAs every few weeks.

The 3-Metric Scale or Kill Check

Metric	Scale Signal	Kill Signal	How to Check
Maximum Drawdown	Within expected range documented by vendor (or within 1.5x of stated max)	Exceeds vendor's documented maximum drawdown by more than 50%	Compare your account's max DD against vendor's stated or verified DD
Win Rate	Consistent with vendor's verified results (within 5-10% variance)	Significantly below vendor's claims over a minimum of 30 trades	Count wins and losses over at least 30 completed trades
Profit Factor	Above 1.3 over 30+ trades	Below 1.0 over 30+ trades (net losing)	Total gross profit divided by total gross loss

How to use this table:

• All three metrics in the Scale zone: gradually increase position size. Double it, do not multiply by 10. Run at the new size for 2-4 weeks before increasing again. Once you have consistent live results, consider scaling beyond personal capital through performance-based programs like Axi Select, which allocates capital based on verified live performance — no challenge fees, no time pressure (affiliate link at no extra cost).
• Any single metric in the Kill zone: stop live trading immediately. Move back to demo. Contact the vendor with your data. If the vendor cannot explain the discrepancy, you have your answer.
• Metrics are borderline or mixed: extend the evaluation for another 30 days at minimum size. More data resolves ambiguity. Fewer trades increases uncertainty.

For a deeper understanding of what these metrics mean and how to interpret them correctly, read reading EA performance like a pro.

If you want a structured framework for tracking your EA evaluation from start to scale, the Trading Agenda provides a daily and weekly tracking system designed specifically for EA traders managing this process.

The Mistakes That Destroy New EA Traders

Every mistake on this list is one I have seen repeatedly. Some of them I made myself before developing this 90-day process. Understanding these failure modes is as important as following the steps above.

• Going live on day one. You have verified nothing. You do not know if the EA works with your broker, your VPS, your account type, or your market conditions. Demo first. Always. No exceptions. If you want to reduce the learning curve risk, start with a free strategy module — it lets you practice the entire demo-to-live process without even buying an EA first.
• Full position size immediately. Even if the EA performed well on demo, live execution introduces variables you have not tested. Start micro. Scale into full size over weeks, not hours. The cost of patience is minimal. The cost of discovering a problem at full size is not.
• Turning off the EA during drawdown. This is the single most destructive mistake in EA trading. Every EA experiences drawdowns. They are not bugs; they are a statistical certainty for any trading system. If you turn off the EA during a drawdown and turn it back on during a winning streak, you systematically destroy the statistical edge the strategy was designed to capture. You take all the losses and miss the recoveries. Only turn off an EA if the kill criteria from the table above are met.
• Running on a home computer. One Windows update during London open. One power flicker during NFP. One internet drop while three trades are open. The $15-$50 per month VPS cost is not an expense. It is insurance against scenarios that will eventually happen.
• Ignoring swap costs. Open positions held overnight incur swap fees. On some currency pairs and directions, these fees can silently eat profits over weeks. Before going live, check your broker's swap table for every pair the EA trades. If the EA holds positions for multiple days, swap costs can significantly affect net performance.
• Not documenting. If you do not track your EA's performance systematically, you cannot evaluate it objectively. Without documentation, every decision becomes based on feelings and recency bias, not data. You will convince yourself the EA is failing during a normal drawdown, or convince yourself it is working during a lucky streak. Both conclusions are equally dangerous without data.

When to Walk Away From an EA

Knowing when to stop is as important as knowing how to start. These are clear, non-negotiable criteria for walking away. They are not suggestions. They are circuit breakers designed to protect your capital from situations where continuing does more harm than stopping.

• Three consecutive months of underperformance versus the vendor's verified results. One bad month happens. Two bad months can happen. Three consecutive months of results significantly below what the vendor demonstrates means the strategy may have stopped working for current market conditions. Move to demo and observe for another month before making a final call.
• Drawdown exceeds the vendor's documented maximum by 50% or more. If the vendor claims 15% maximum drawdown and your account hits 23%, something is fundamentally wrong. Either the documentation was inaccurate, market conditions have shifted beyond the strategy's design parameters, or there is an execution problem. In any case, stop live trading.
• The vendor disappears or stops updating. Markets evolve. Brokers change conditions. Platforms update. An EA without ongoing support is an EA with an expiration date. If the vendor stops responding to support requests or ceases development, this is a terminal signal regardless of current performance.
• You cannot sleep because of the EA. If the position size or risk level creates anxiety that affects your daily life, the size is too large for your personal risk tolerance. This is not a weakness. It is information. Size down until the EA runs without affecting your wellbeing, or stop entirely.
• The math does not work. If your total monthly costs (VPS, data feeds, swap, spread costs) exceed the realistic return expectations for your account size, the economics are against you regardless of EA quality. A $200 account generating $10/month net cannot sustain $50/month in infrastructure costs. Be honest about the numbers.

Walking away from an EA is not failure. It is risk management. The traders who lose the most money are not the ones who try EAs and stop. They are the ones who cannot walk away from a losing situation and keep funding accounts hoping things will change without changing anything.

For a comprehensive approach to understanding what different position management approaches mean for your risk, read about fixed lots versus smart position management.

Frequently Asked Questions

How long should I demo test an EA?

Minimum 2 weeks. Ideally, you want at least 20 completed trades to analyze before forming any conclusions. Some slower strategies that trade a few times per week may require 4 or more weeks on demo to accumulate enough data. The demo phase costs you nothing except time, and that time buys you information that prevents costly mistakes on live accounts.

Can I skip the micro-live phase?

Technically, yes. Practically, you should not. By skipping micro-live, you accept unverified execution risk on your full position size. The micro-live phase costs very little (you are trading minimum lots) and validates that live conditions match your demo expectations. The difference between demo and live execution can be significant with some brokers, and you want to discover that on 0.01 lots, not 0.1 lots.

What if the EA does not trade for weeks?

Some strategies are highly selective and may go days or even weeks without placing a trade. This is especially common with strategies that wait for specific technical confluences or volatility conditions. Check the vendor's documented trade frequency. If your EA's silence falls within the documented range, be patient. If it is significantly below what the vendor claimed, investigate: check the Experts tab for errors, verify the EA is attached to the correct chart and timeframe, and confirm the trading session hours match your broker's server time.

How do I know if my VPS is good enough?

Check the ping to your broker's trading server. In MT4/MT5, look at the bottom-right corner of the terminal for the connection speed indicator. Ideally, you want under 5ms. Under 10ms is acceptable for most strategies. Above 20ms starts to affect execution quality, especially for scalping EAs. Additionally, monitor the Journal tab daily for disconnection or reconnection events. If you see frequent "connection lost" and "reconnected" messages, your VPS either has unreliable networking or insufficient resources. Upgrade your plan or change provider.

For detailed broker and VPS testing data, the Testing Lab provides real-world comparisons to help you make informed decisions about your trading infrastructure.

Resources

• Testing Lab — Broker and VPS recommendations based on real testing data
• DoIt Trading Newsletter — Weekly insights on EA trading, performance updates, and market analysis
• DoIt MultiStrategy Pro — The next step after your first EA: 5 strategies, one platform, portfolio diversification

Broker recommendations:

• Pepperstone — Low barrier to entry for your first live EA account
• IC Markets — Raw spreads and institutional-grade execution for scaling
• Free USDJPY Strategy Module — Practice the demo-to-live process at zero cost before buying any EA
• Axi Select — Scale capital based on verified live performance, no challenge fees (affiliate link)

{ "@context": "https://schema.org", "@type": "HowTo", "name": "I Bought a Forex EA. Now What? (The First 90 Days That Actually Matter)", "description": "You bought a forex EA — now what? A complete 90-day setup guide covering demo testing, VPS, micro-live validation, and the scale-or-kill decision.", "totalTime": "P90D", "step": [ { "@type": "HowToStep", "name": "Demo Test (Weeks 1-2)", "text": "Install the EA on a demo account with your intended broker's conditions. Verify it places trades correctly, monitor for errors, and build a performance baseline by tracking win rate, trade duration, and trade frequency.", "position": 1 }, { "@type": "HowToStep", "name": "Set Up Infrastructure (Weeks 3-4)", "text": "Configure a VPS with proximity to your broker's servers, open and fund a live trading account with appropriate settings, and set up news filters, notifications, and daily loss limits.", "position": 2 }, { "@type": "HowToStep", "name": "Go Micro-Live (Month 2)", "text": "Deploy the EA on your live account with the smallest possible position size. Keep a deviation log comparing live trades to demo behavior, tracking entry price differences, slippage, and missed trades over 30 or more trades.", "position": 3 }, { "@type": "HowToStep", "name": "Make the Scale or Kill Decision (Month 3)", "text": "Evaluate three key metrics: maximum drawdown versus vendor documentation, win rate consistency over 30-plus trades, and profit factor. If all metrics are positive, gradually scale position size. If any metric fails, stop live trading and investigate.", "position": 4 } ] }
{ "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "How long should I demo test an EA?", "acceptedAnswer": { "@type": "Answer", "text": "Minimum 2 weeks. Ideally, you want at least 20 completed trades to analyze before forming any conclusions. Some slower strategies may require 4 or more weeks on demo. The demo phase costs you nothing except time, and that time buys you information that prevents costly mistakes on live accounts." } }, { "@type": "Question", "name": "Can I skip the micro-live phase?", "acceptedAnswer": { "@type": "Answer", "text": "Technically yes, but practically you should not. By skipping micro-live, you accept unverified execution risk on your full position size. The micro-live phase costs very little and validates that live conditions match your demo expectations. The difference between demo and live execution can be significant with some brokers." } }, { "@type": "Question", "name": "What if the EA does not trade for weeks?", "acceptedAnswer": { "@type": "Answer", "text": "Some strategies are highly selective and may go days or even weeks without placing a trade. Check the vendor's documented trade frequency. If your EA's silence falls within the documented range, be patient. If it is significantly below what the vendor claimed, investigate: check the Experts tab for errors, verify the EA is on the correct chart and timeframe, and confirm trading session hours match your broker's server time." } }, { "@type": "Question", "name": "How do I know if my VPS is good enough?", "acceptedAnswer": { "@type": "Answer", "text": "Check the ping to your broker's trading server in MT4/MT5. Ideally you want under 5ms, under 10ms is acceptable for most strategies, and above 20ms starts to affect execution quality. Also monitor the Journal tab daily for disconnection events. Frequent connection lost messages mean your VPS has unreliable networking or insufficient resources." } } ] }