03 · Forex Automation (EAs) and Copy Trading: Holy Grail or Meat Grinder
Forex has two of the most seductive narratives: "EA auto-profits" and "copy the pros". What they share is charging you for the promise that "you don't need to trade yourself". This chapter covers the technical truth about MT4/MT5 and EAs, how backtest scams operate, the traps of copy-trading communities, the market maker bucket-shop model — ending with a platform-selection red-line checklist and a conclusion for ordinary people.
1. What MT4 / MT5 Are
MetaTrader 4 (MT4) and MetaTrader 5 (MT5) are forex/CFD trading terminals developed by Russia's MetaQuotes, and they are the most widely used software in retail forex:
| Item | MT4 | MT5 |
|---|---|---|
| Positioning | The veteran forex terminal (released 2005) | Its successor (2010), supporting more instruments and tools |
| Language | MQL4 | MQL5 |
| Adoption | Standard on legacy platforms; highest forex market share | Increasingly the default offering |
| Instruments | Forex, some CFDs | Broader: forex, stocks, futures, options, crypto CFDs, etc. |
Core components:
- Market terminal: live quotes, candlestick charts, technical indicators;
- Order panel: market/limit/stop-loss & take-profit orders;
- EA (Expert Advisor): an automated trading program written in MQL that, once attached to a chart, executes an entire trading logic automatically;
- Strategy Tester: backtests EAs/strategies on historical data — note: this feature is genuinely useful but also a breeding ground for scams (see Section 3).
MT4/MT5 themselves are neutral tools: they don't create profit, only execution. When a platform offers you a "free EA" or a community shares a "guaranteed-profit EA", the problem was never the software — it's who wrote it and who's selling it.
2. What an EA Can Do
An EA is essentially "trading rules written as code, executed automatically by a program". It can:
1. Execute a Strategy by Rules
Using the classic dual-MA EA as an example, its logic fits in a few lines:
| Step | Logic | Code Sketch |
|---|---|---|
| Entry | Fast MA (e.g., MA5) crosses above slow MA (e.g., MA20) → buy; cross below → sell | if (fastMA > slowMA && prevFastMA <= prevSlowMA) Buy() |
| Exit | Close when any of reverse signal / stop loss / take profit triggers | if (price <= stopLoss) Close() |
| Loop | Recalculate on every new bar | OnTick() / OnBar() |
2. Stop-Loss/Take-Profit Management
- Automatically attaches stops/targets/trailing stops after entry, removing the human brain's tendency to "wait a bit longer" while losing;
- This is an EA's most useful feature — and the only one recommended for ordinary users: treat the EA as a "discipline executor", not a "trading brain".
3. 24-Hour Monitoring
- Forex runs 24 hours; humans cannot watch it all, EAs can;
- But "being able to watch" isn't "knowing how to trade" — the program merely executes your rules more diligently.
An EA in One Sentence
An EA is "the automated version of your trading rules": if your rules lose, the EA loses automatically; if your rules win, the EA wins automatically. It creates no edge out of thin air.
💀 Iron Rule: If Your Rules Lose, the EA Loses Automatically
If your rules lose, the EA loses automatically; if your rules win, the EA wins automatically. An EA is "the automated version of your trading rules" — it creates no edge out of thin air. That's why no publicly sold EA delivers "a stable 30% monthly": if one did, its owner would quietly compound it rather than sell it to you for $999. Ask three questions of any EA ad: Can the live account be watched in real time? How does performance look across random parameter sets? If it prints money, why sell it?
3. The Truth About EAs: The "Holy Grail" Scam
Why Every "30%-a-Month EA" Is a Scam
EA sales pitches are remarkably uniform: "stable 20%–30% monthly returns", "perfect three-year backtest curve", "source code included / installation included". Breaking down the tricks:
| Scam Tactic | Mechanism | How to Detect |
|---|---|---|
| Backtest cosmetics | Picking "the best-performing parameters" from the tester — one in ten thousand | Demand the random-parameter distribution and live results; reject anything showing only backtest screenshots |
| Future functions | The EA computes indicators with data that doesn't exist until the current bar closes (e.g., entering intrabar using close-derived values), distorting backtests | Check whether the EA trades before the current bar closes; re-run the backtest on other timeframes |
| Curve fitting (Overfitting) | Parameters tuned to precisely match historical prices; fail instantly out of sample | Backtest across 5–10 different year ranges; profit in only one range = fitting |
| Missing Monte Carlo | Showing one optimal path, hiding the effects of slippage, disconnections, requotes | Check whether the backtest accounts for spread, slippage, and network interruptions |
| No in-sample/out-of-sample split | "Training" and "validating" on the same stretch of data | Demand out-of-sample testing (optimize 2005–2020, validate 2021–2024) |
Realistic Expectations for Real EAs
- No fundamental difference from manual trading: an EA is just an execution machine; long-term expectancy depends on the strategy itself — whose win rate and payoff ratio are exactly what manual research would find;
- A real EA offers only three advantages: execution discipline, 24-hour operation, verifiable backtesting — not "win rate";
- No public strategy exists that makes "a stable 30% a month": if it did, its owner would quietly compound it instead of selling it to you for $999.
A Simple Falsification Test
Ask any EA ad three questions:
- "Can I watch your live account (not demo) in real time?" — No = 90%+ likely a scam;
- "Are the best-backtested parameters the only choice? How do random parameters perform on average?" — Dodging = fitting;
- "If it always profits, why not just trade it with max leverage yourself instead of selling?" — Internal consistency is the first principle of scam detection.
4. The Traps of Copy-Trading Communities
Copy trading: the platform replicates a "star trader"'s signals into your account, marketed as "profit alongside the masters".
Why "He Wins, You Lose"
| Factor | Mechanism |
|---|---|
| Different capital size | He has $1M, you have $1k; at the same lot ratio your distance to a blow-up is far shorter — the drawdown he can survive, you can't |
| Slippage and execution differences | His signals execute late on your account; slippage differs in big moves, so outcomes diverge completely |
| Different withdrawal timing | He screenshots at profit peaks for marketing; you enter during his drawdown — what you see is the past; what you bought is the future |
| Survivorship bias | Communities show only "surviving stars": equally many traders a year ago blew up and were hidden; 80%-win-rate scalping "stars" mostly trade high risk for small gains and can go to zero at any moment |
| Volatility of returns | "200% annualized" often comes from one extreme month (+60%) while other months lose — average returns and lived experience are two different things |
The Essence of Forex Signal Groups
- "Teacher-led trading" groups: signals → you deposit at a designated platform → teacher collects rebates (the platform shares revenue based on your losses) — your losses are the teacher's salary;
- "Profit screenshots": backend tools generate them in one click, and most are demo accounts or Photoshop;
- "Signals + shady broker" is an assembly line: teacher and platform are the same crew — when you win they alter quotes; when you lose they share the proceeds.
The single standard for judging whether copy trading / signal providers are legit: look at the complete long-term track record (including every losing trade), never at screenshots or "win rates". That no signal provider ever volunteers the full record is itself the answer.
💀 Iron Rule: The Only Credible Proof Is the Full Track Record
The single standard for judging copy trading / signal services: the complete long-term track record (including all losers), not screenshots or "win rates". That no provider shows the full record is itself the answer — the full record necessarily contains their worst losing streak, and survivorship bias displays only surviving stars. So the first line of defense isn't "how high is this trader's win rate" but "does he dare show you every losing trade".
5. Market Makers vs ECN/STP: Who Is Your Counterparty
Three Order-Routing Models
| Model | Full Name | Mechanism | Relationship to Clients |
|---|---|---|---|
| Market Maker (MM) | Market Maker | Broker sets its own book and quotes; client orders are not routed to market — the broker is the direct counterparty | Bets against clients: client losses = broker profits (especially where retail loss ratios are high) |
| STP | Straight Through Processing | Orders forwarded directly upstream to liquidity providers; broker earns only a spread markup | No betting (or minimal) |
| ECN | Electronic Communication Network | Orders matched in a liquidity pool with banks/institutions/other clients; quotes are true market price + commission | Pure conduit; clients face each other |
Why Some Brokers Bet Against Clients
- The MM model is not illegal per se (under proper regulation, FCA/ASIC etc. permit MM operations), but it creates a structural conflict of interest: the broker profits when you lose;
- Typical moves by betting brokers: artificial slippage / disconnects / "requotes" during data events, turning would-be winners into losers;
- Warning signs: quotes persistently deviating from mainstream markets; frequent "connection lost" during data releases; spreads widening dramatically.
The Structure of Spreads and Overnight Interest (Swap)
| Cost Item | Structure | Collected By |
|---|---|---|
| Spread | Bid - Ask, marked up by the market maker | Broker/market maker |
| Commission | Explicit ECN fee (fixed per lot) | Liquidity provider/conduit |
| Swap (overnight interest) | Charged/paid on positions held overnight per rate differentials: long high-yield currency earns swap, short high-yield currency pays swap; triple swap settles Wednesday (Thursday on some platforms) | Broker/market maker |
Key insight: spreads and swaps are neutral market structure, but a market maker can simultaneously be "your counterparty" and "your fee collector" — which is why regulatory licensing and order-routing model must be verified together.
The Full Picture of Forex Market Participants
| Participant | Role | Trading purpose |
|---|---|---|
| Central banks | The ultimate heavyweights | Exchange-rate policy, reserve management, market intervention |
| Commercial banks | Market makers | Quote for clients, earn the spread, manage their own exposure |
| Hedge funds / asset managers | The speculative main force | Macro bets, arbitrage, volatility trading |
| Multinationals | Hedgers | Settle payments, hedge FX risk |
| Retail traders (you) | The weak side | Speculation — mostly losing money under high leverage |
Key insight: retail traders are at a comprehensive disadvantage in information, speed, cost, and capital size. Banks see client order flow; hedge funds employ teams of PhDs — the retail trader's only edge is "flexibility", and leverage turns that flexibility into a disadvantage too. In this market, 9 out of 10 retail traders ultimately losing is not a joke, it is data (public figures differ across platforms; this is a common-sense conclusion only).
6. Red-Line Checklist for Choosing a Platform
| Red-Line Item | Concrete Check | Pass Criteria |
|---|---|---|
| Regulatory license | Verify the license number on the regulator's website (never trust screenshots of their own site): UK FCA, Australia ASIC, Cyprus CySEC are mainstream; US CFTC/NFA is strictest with the highest bar | License number real; status "Active/Authorised" |
| Withdrawal test | Deposit a small amount ($100–500) → request withdrawal immediately → record arrival time | Arrives within 3–5 business days, no "fees"/"frozen account" excuses |
| Leverage cap | Mainstream regulators cap retail leverage hard: EU ESMA major pairs 1:30, UK FCA 1:30, US around 1:50; brokers offering 1:500+ are almost always offshore-regulated | Licensed offshore ≠ illegal, but a whole risk tier higher |
| Negative-news search | Search "platform name + withdrawal / scam / complaints"; check FX complaint sites and regulator enforcement records | No widespread withdrawal complaints; no regulatory penalties |
| Deposit/withdrawal channels | Bank/licensed payment rails only; private transfers, crypto deposits, demands for cash via corporate accounts = instant blacklist | Everything through licensed channels |
| Client fund segregation | Client funds held in segregated accounts, audited as required by regulators | Verifiable on regulator sites / in terms |
7. Demo vs Live: The Truth
"The EA prints money in demo/backtest but loses live" — the most common real experience. The EA didn't change; demo and live differ systematically:
| Dimension | Demo/Backtest | Live |
|---|---|---|
| Quotes | Ideal prices, instant fills | Slippage, latency, requotes (severe during data events) |
| Spread | Fixed default value | Floating in real time, multiplying instantly on data releases |
| Psychology | Losses painless; the EA can pyramid endlessly | Real losses trigger human intervention (manually killing the EA, changing parameters) |
| Servers | Local/test environment | Network drops, platform failures, overnight maintenance |
| Fill confirmation | Assumes everything fills | Partial fills, rejected quotes, stop slippage |
The correct validation process (demo first, then live):
① Out-of-sample backtest: optimize on 2015–2020 data, validate on 2021–2024
↓
② Run 2–3 months on demo: check performance under slippage/drops versus backtest
↓
③ Micro-live (0.01 lots) for 1–2 months: confirm execution and mindset
↓
④ Only after ①②③ pass, consider small-size live trading at 1%–2% risk per tradeThe vast majority of paid EAs die at step one (out-of-sample backtest) — because their "perfect curves" hold only in specific historical windows.
8. Conclusions for Ordinary People
Common-Sense Risk/Reward Comparison
| Approach | Long-Term Expectancy | Risk |
|---|---|---|
| Manual trading | Matches ability (negative for most) | High |
| EA automated trading | Matches the strategy ("holy grail EA" ≈ 100% negative expectancy) | High |
| Copy trading | Strongly tied to stars' survivorship bias (most lose) | High |
| Not trading (learning by watching) | 0% | None |
If You Must Play
- Micro account + low leverage: start at 0.01 lots, cap leverage at 1:30 (not the platform-default 1:500);
- Trade EUR/USD only: tightest spread, most regular behavior (see 01 · Forex Trading Practice);
- Use EAs only for "discipline execution" (auto stop-loss/take-profit); never buy a paid EA;
- Copy-trade only strategies you already understand, with light size, funded at "losing it all is fine" level;
- The words "guaranteed profit" mean "leave immediately".
✅ Conclusion: "Guaranteed Profit" Means "Leave Immediately"
Any claim of "guaranteed profit" means "leave immediately". The conclusion for ordinary people is simple: the long-term expectancy of manual trading, EA automation, and copy trading is negative for most people; only not trading (learning by watching) has 0% expectancy — don't spend real money validating "probabilistic thinking"; save the tuition for books and a demo account.
Risk Warning
⚠️ Risk Warning
The forex industry contains many shady brokers and betting-shop models: unlicensed or clone-firm platforms can manipulate quotes, apply artificial slippage, and block withdrawals; under the MM model the broker bets directly against clients, and much of the "teacher-led signals", "copy-trading rebates", and "paid EA" chain ultimately monetizes retail losses. Buying paid EAs, joining signal groups, or chasing "30% a month" marketing will most likely zero your capital fast. Content here about regulation, leverage, and broker models is educational — defer to the latest regulations and platform terms. This article is not investment advice.