How To Avoid
EA Scams
Strong EA selection relies on evidence quality, execution realism, and risk architecture. This guide gives a practical audit sequence for evaluating automated systems.
Updated:
Key takeaways
- Ask four things of any backtest: data source, spread, commission, and the dates covered.
- A real edge survives a nudge to its settings. An over-tuned one breaks on a sensitivity run.
- Test in three market types — trending, ranging and wild. One good run proves nothing.
- Run live at two brokers. Numbers that hold at both belong to the strategy, not the feed.
- Recovery logic that raises exposure after losses needs a hard risk cap before anything else.
How To Spot
A Faked Backtest
Audit settings, assumptions, and parameter stability before trusting historical curves.
A backtest is only as good as the inputs you can repeat. Before you trust the curve, ask four things. Where did the price data come from? What spread did it assume? What did it charge in commission? Which dates did it cover?
A robot you can trust does not fall apart when you nudge a setting. It should behave much the same one step either side of its inputs. The same holds in calm markets and wild ones.
Test it in three kinds of market: trending, range-bound, and wild. One good run proves nothing. Three of them tell you whether the edge is real.
Integrity Checklist
- One fixed set of price data, with the time zone spelled out.
- Real costs: spread, commission, and swap.
- Walk-forward runs that keep the test data apart from the data used to tune.
- A Monte Carlo reshuffle, to see if the order of trades was luck.
Curve-Fitting Markers
An over-tuned robot breaks when you change a setting or raise the cost by a hair. A quick sensitivity run shows it.
Live
Verification
Validate execution behavior and risk consistency on controlled real-account samples.
A live test shows how the robot behaves on a real account. It is the only way to see slow fills and server lag. Also prices that shift from one session to the next.
Use a small account and a fixed risk per trade. That keeps the runs comparable. It also keeps the account alive long enough to teach you something.
Run it at two brokers, not one. If the numbers hold at both, the edge sits in the strategy. If they do not, it sat in one broker's feed.
Forward Test Protocol
Decision Rule
Put money in only when the evidence earns it. Keep going while drawdown, trading cost, and behaviour stay inside the limits you set before the test began.
Frequently Asked Questions
What is the first filter for an EA performance claim?
Start with data integrity: tick quality, spread and commission assumptions, and reproducible test settings.
Why is live verification required after a backtest?
Live verification confirms behavior under real execution conditions, including slippage, latency, and broker routing differences.
Which risk pattern deserves immediate attention?
Aggressive recovery logic with rising exposure after losses deserves strict control through hard risk caps.
What evidence quality standard improves decision accuracy?
Versioned strategy rules, complete trade logs, and consistent reporting cadence build high-quality evaluation evidence.
What is the role of this guide?
This guide provides an educational evaluation framework for automated trading systems.
Continue Your EA Due Diligence
Backtest vs Forward Test
Apply validation sequence before allocation: historical test, then controlled live verification.
Slippage and Latency
Execution friction metrics complete EA evaluation and improve fill-quality expectations.
Position Sizing Playbook
Risk architecture defines account survivability during strategy drawdown periods.
Trading Journal Framework
Structured logs capture parameter drift, broker sensitivity, and stability across market regimes.
The Truth About Trading Signals
Investigative dataset on conversion-first marketing and performance-claim patterns.
Robots Rankings
Compare live-track records and consistency markers across automated systems.