The Expert Advisor marketplace is crowded with impressive numbers. Equity curves that climb relentlessly upward. Profit factors above 3. Win rates over 80%. Annualised returns in triple digits. For a trader evaluating these products, the challenge is not finding impressive results - it is distinguishing genuine edge from a performance record that will not hold up when real money is deployed. This skill is especially relevant in Q4 2023, when the combination of heightened market volatility and a busy season for EA marketing creates significant pressure to act quickly on insufficient information.
The First Question: Live or Backtest?
The most basic distinction in evaluating any EA's track record is whether the results come from live trading or backtesting. These are not equivalent categories, and they should never be evaluated by the same standards.
A backtest is a simulation run on historical data. It tells you how the strategy would have performed under the specific assumptions programmed into the tester - spread model, slippage assumptions, data quality. It is a hypothesis about performance, not a measurement of it. Live results, verified through an independent platform with a real broker connection, are actual performance data.
This does not mean backtests are worthless - they are essential for understanding strategy characteristics and performing initial viability assessment. But a backtest alone, however impressive, should not be the primary basis for deploying real capital. The gap between backtest and live performance is one of the most consistently observed phenomena in retail systematic trading.
Evaluating Live Results
For live results, the key questions to answer before drawing conclusions:
How long has it been running? Statistical significance in trading performance requires a meaningful sample. A strategy running for 3 months with 50 trades is not statistically distinguishable from luck, even if the equity curve looks good. Generally, 6-12 months of live data with 200+ trades begins to provide a meaningful signal. Longer is better.
Is it verified independently? Results published directly by a developer - screenshots, PDF reports, unverified MetaTrader statements - can be manipulated or selectively presented. Results on platforms like Myfxbook, FX Blue, or MyFXBook with verified account connections cannot be altered retroactively. This verification is non-negotiable for serious evaluation.
What is the maximum drawdown, and how does it compare to the backtest? Live maximum drawdown significantly exceeding the backtest's maximum drawdown is a significant red flag. Some divergence is expected - live trading is harder than backtesting - but if the live account has already experienced 25% drawdown while the backtest showed 12%, the backtest assumptions were too optimistic.
What market conditions does the track record cover? A strategy that has only traded in trending conditions has not demonstrated it can handle range-bound markets, and vice versa. A 12-month track record that covers multiple distinct market regimes - trending, ranging, high volatility, low volatility - is significantly more informative than the same length of record in uniform conditions.
Statistics That Matter
Beyond the headline return, these are the numbers that provide genuine insight into strategy quality:
- Profit factor: Total gross profit divided by total gross loss. A profit factor of 1.3-1.8 in live conditions is realistic and sustainable. Above 2.5 on live results is exceptional and warrants scrutiny of whether conditions will continue.
- Average win to average loss ratio: Does the strategy win more per winning trade than it loses per losing trade, or does it rely on high win rates with small winners and large losers? High-win-rate strategies with poor risk-reward ratios are fragile - a few large losing trades can erase many small winners.
- Consecutive losses: The maximum number of consecutive losing trades tells you a great deal about the strategy's worst-case psychological demand. A strategy with a maximum of 3 consecutive losses is much easier to stick with than one with a historical maximum of 12, even if the overall win rates are similar.
- Sharpe ratio: Return relative to volatility of returns. A Sharpe above 1.0 in live conditions is good; above 1.5 is excellent. Sharpe ratios above 3 should be viewed with scepticism unless supported by a very large trade sample.
Questions to Ask the Developer
Legitimate EA developers with genuine edge are typically transparent about the mechanics of their strategy, the conditions under which it performs best and worst, and the risks involved. Questions worth asking:
- What market conditions does this strategy perform worst in, and why?
- How was the strategy developed - was out-of-sample testing used?
- What drawdown should a user expect to experience over a 12-month period?
- Has the strategy ever been updated, and if so, what triggered the update?
Evasive or vague answers to these questions are informative. Transparent, specific answers - even when they acknowledge limitations - suggest a developer who understands their own product and has thought carefully about its real-world deployment.
At Dollar Robber, both Black Tie and Gold Dwarf Scalper come with verified live account tracking, documented backtest parameters, and clear statements about the conditions each strategy is designed for. The goal is to provide enough transparent information for traders to make informed decisions rather than impulse purchases based on impressive-looking equity curves. That standard of disclosure is what responsible EA marketing looks like, and it is what you should require from any product you seriously evaluate.
The Bottom Line
Evaluating an EA honestly requires moving past the headline numbers and into the underlying data. A verified live track record of 12+ months, with a profit factor above 1.3, drawdown consistent with the backtest, and a Sharpe above 1.0, represents a meaningful signal of genuine edge. Anything less is hypothesis rather than evidence, and should be deployed with proportionally smaller allocation until sufficient live data accumulates to upgrade your confidence.