How to Track EA Performance: The 7 Metrics That Actually Matter

Here’s something that still surprises me: most traders spend weeks picking an EA, drop it on a $10,000 account, and then… never look at it again. Maybe they glance at the balance once a month. Maybe.

That’s like buying a car and never checking the oil.

An EA is not a money printer that runs itself. Markets shift. Spreads widen. Liquidity dries up during holidays and long weekends. The strategy that crushed it during gold’s Q1 2025 rally might be slowly bleeding out in a February 2026 consolidation. You won’t know unless you’re watching the right numbers — and most traders aren’t.

I’ll be direct: tracking EA performance properly isn’t complicated. But it does require knowing which metrics actually matter (spoiler: total profit isn’t one of them) and building the habit of reviewing them consistently.

What does tracking EA performance actually mean?

Tracking an EA means systematically monitoring its trading results over time using quantitative metrics. Not just “am I up or down this month?” but a structured analysis of how your EA is making (or losing) money, how much risk it’s taking, and whether its behavior matches what you expected from the backtest.

Think of the difference between checking your bank balance and reading the full statement. One gives you a number. The other tells you a story.

Why does this matter more than most people think? Because EAs degrade. All of them. No exceptions. The edge that worked in 2024 weakens as market microstructure shifts, as more traders discover similar patterns, as central bank policy changes the volatility regime. The Fed’s September 2024 pivot didn’t just move markets — it fundamentally changed how gold, forex pairs, and indices behave on intraday timeframes. If your EA was calibrated for a high-rate environment, it’s now operating in different terrain.

Performance tracking catches this degradation early. Without it, you’re flying blind.

The 7 metrics every EA trader must watch

1. Profit Factor

Gross profit divided by gross loss. The clearest measure of whether your EA has a real edge or not.

Here’s the thing: many traders obsess over total profit, but Profit Factor tells you something more fundamental — for every dollar you lose, how much do you get back? A Profit Factor of 1.8 means you earn $1.80 for every $1.00 lost. That’s a real, quantifiable edge.

Profit Factor	Assessment
Below 1.0	Losing money. Stop the EA.
1.0 – 1.3	Marginal. Commissions and slippage may eat the edge.
1.3 – 1.75	Acceptable. Viable in live trading with strict cost control.
1.75 – 2.5	Strong. Real and sustainable edge.
Above 4.0 (with fewer than 100 trades)	Suspicious. Likely overfitted or too few trades to be statistically significant.

Check this weekly. If Profit Factor drops below 1.3 for two consecutive months on a strategy that historically runs at 1.8+, something has changed. Don’t wait for it to go below 1.0.

2. Maximum Drawdown

The largest peak-to-valley decline in your account equity. This is the metric that keeps you in the game — or takes you out of it.

I’d argue that drawdown matters more than profit. A 50% drawdown requires a 100% gain just to get back to breakeven. Most traders can’t psychologically handle watching $5,000 evaporate from a $10,000 account, even if the strategy eventually recovers. They close at the worst possible moment.

The benchmark: keep maximum drawdown below 20% for moderate strategies, below 10% for conservative ones. If your EA showed 15% drawdown in backtesting but you’re seeing 25% live, that’s a red flag — not a “temporary rough patch.”

A counterintuitive point: a strategy with lower returns but controlled drawdowns will almost always outperform a high-return, high-drawdown strategy over the long run. The math of compounding punishes deep drawdowns severely.

3. Sharpe Ratio

Risk-adjusted return. It answers the question: “how much return am I getting per unit of risk taken?” Calculated as (average return − risk-free rate) / standard deviation of returns, annualized.

Sharpe Ratio	Assessment
Below 0.5	Poor. Too much risk for the return generated.
0.5 – 1.0	Acceptable. Competitive with buy-and-hold strategies.
1.0 – 2.0	Good. Solid risk-adjusted performance.
Above 2.0	Excellent. Institutional-grade risk management.

Why not just look at total return? Because a strategy that returns 40% annually with a Sharpe of 0.3 is taking enormous risk to generate those returns. You’re going to experience stomach-churning volatility. A strategy that returns 15% with a Sharpe of 1.5 lets you sleep at night — and will probably compound more money over 5 years because you won’t panic-close during drawdowns.

4. Win Rate (with context)

Percentage of trades that close in profit. Easy to understand, easy to misinterpret.

Look, win rate in isolation is practically meaningless. A martingale strategy can show 95% win rate right before it blows up the account. What you need is win rate combined with the risk/reward ratio:

Expectancy = (Win% × Average Win) − (Loss% × Average Loss)

A 40% win rate with 3:1 ratio = (0.40 × $300) − (0.60 × $100) = $60 per trade. An 85% win rate with 0.3:1 ratio = (0.85 × $30) − (0.15 × $100) = $10.50 per trade. The “ugly” 40% win rate earns almost 6x more per trade.

Track win rate alongside average win size and average loss size. All three together. Never in isolation.

5. Expectancy

Average profit per trade, counting both winners and losers. This is the number that tells you whether your EA has a mathematical edge — and how strong it is.

Positive expectancy means the EA makes money over time. But here’s what most people miss: expectancy has to be large enough to cover real-world costs that backtests underestimate. On XAUUSD with a typical ECN broker, round-trip costs (spread + commission) run $5-15 per 0.10 lots. If your EA’s expectancy is $8 per trade, half that edge is getting eaten by costs.

I track expectancy relative to cost: Expectancy / Round-trip Cost. If that ratio drops below 2.0, the strategy is on thin ice. Your broker is required to disclose these costs transparently — use real numbers, not estimates.

6. Recovery Factor

Net profit divided by maximum drawdown. It tells you how efficiently your EA recovers from its worst moments.

Say your EA generated $8,000 in total profit with a maximum drawdown of $2,000. Recovery Factor = 4.0. That’s healthy — the strategy earns 4x what it risks in its worst-case scenario. Below 3.0, recovery is sluggish. Below 1.0 means the strategy hasn’t even recovered its worst drawdown yet, which is a serious red flag.

Recovery Factor increases over time if the strategy is genuinely profitable (profits accumulate while max drawdown stays bounded). If it’s decreasing, it means new drawdowns are deeper relative to profits generated. Bad trend.

7. Trade Frequency

How often the EA trades per day, week, or month. It’s the metric everyone ignores, and it’s critical for two reasons.

First, statistical significance. An EA averaging 2 trades per month needs years to generate enough data to evaluate it. You’re flying blind for the first 12-18 months. An EA that makes 5-10 trades per week gives you meaningful data in 2-3 months.

Second, frequency changes reveal regime shifts. If your EA normally makes 15 trades per week and suddenly drops to 3, either it’s not finding setups (market regime changed) or its filters are rejecting everything (possibly overfitted to a very specific condition). Both scenarios demand investigation.

What’s a good Profit Factor for a trading EA?

This deserves its own section because it’s the most-searched question about EA performance — and most answers online get it wrong.

The “right” Profit Factor depends on trade count and time horizon. Here’s what actually matters:

• 1.5 – 2.5 with 300+ trades over 2+ years — the sweet spot. Robust enough to survive real-world costs, consistent enough to indicate genuine edge.
• Above 3.0 with 500+ trades — exceptional. Very few strategies sustain this long-term. If yours does, you’ve found something real.
• Above 4.0 with fewer than 100 trades — almost certainly overfitted. The developer (or the optimizer) found a narrow window where everything clicked. It won’t repeat.
• 1.0 – 1.3 in live trading — break-even territory after costs. Time to rethink parameters or stop.

Something that gets lost in these discussions: Profit Factor on gold (XAUUSD) during a structural uptrend is inflated for buy strategies. A random buy-only strategy with optimized exits can achieve PF 1.5 on gold just from the trend. Your EA needs to beat that baseline, not just beat 1.0.

How to set up your performance dashboard

You have the metrics. Now you need a system to track them consistently. Here are your options, from basic to professional.

Option 1: Native MT5 reports

Open the “Trade” tab in MT5, right-click and select “Report.” You get a basic HTML or XML summary of closed trades. It’s free and always available.

Limitation: no Sharpe ratio, no equity curve over time, no per-EA breakdown if you run multiple strategies. Works for a single EA on a single account. Completely insufficient for anything more complex.

Option 2: Myfxbook / FX Blue

Myfxbook connects directly to your broker account and tracks metrics automatically. It shows equity curves, drawdown analysis, and monthly returns. FX Blue offers similar functionality with a focus on detailed trade analytics.

Pros: free, verified broker connections, public sharing for signal marketing.

Cons: limited per-EA breakdown (groups everything by magic number, but the interface is clunky), no Monte Carlo analysis, no backtest comparison tools, and the free version is plastered with ads.

Option 3: Dedicated analytics platforms

For traders running multiple EAs across different accounts, a dedicated analytics dashboard is the professional approach. These platforms typically offer per-strategy breakdown, institutional-grade metrics (Sharpe, Sortino, Calmar ratios), performance heatmaps by hour and day, Monte Carlo simulations for risk quantification, and the ability to compare backtest projections against live results.

BLODSALGO Analytics, for example, provides all of the above plus calendar views, trade duration analysis, and Telegram alerts for key events — built specifically for EA traders who want institutional-level monitoring without institutional prices. It even processes MT5 backtest Excel reports client-side, so your strategy data never leaves your browser.

Other platforms worth considering include TradesViz (solid visualization, free tier available) and TraderSync (good trade journal, better suited for manual traders).

Option 4: Custom spreadsheet

Export trades from MT5 and build your own tracking sheet. Total control, zero cost, maximum effort. Honestly, if you run a single EA on a single account and enjoy spreadsheets, this works. But the moment you scale to 2-3 strategies across multiple accounts, the spreadsheet becomes a maintenance nightmare.

Backtest vs. live results: why your numbers don’t match

Every EA trader hits this wall sooner or later. The backtest shows Profit Factor 2.1 and 12% max drawdown. Three months live, you’re at PF 1.4 with 22% drawdown. What happened? (If you’re not sure how to interpret your backtest results in the first place, read How to Read MT5 Backtest Results first.)

Several things, and they’re all predictable:

Spread and slippage divergence. Backtests use historical or fixed spreads. Live trading encounters spread spikes during news events (NFP, FOMC, CPI). Gold spreads can triple during the London open. Your backtest didn’t model that $50 widening at 8:30 AM EST when December 2025’s NFP came in 100K above consensus.

Execution latency. Your backtest assumes instant execution. Real execution takes 50-200ms on a VPS, more from your home connection. For scalping strategies on M1-M5, this latency costs 1-3 pips per trade. Over hundreds of trades, that drags performance significantly.

Regime shifts. The backtest may include data from 2020-2024. But if 80% of the profit came from gold’s 2024 rally (XAUUSD went from $2,050 to $2,790), your strategy might be exploiting a specific trend that’s already exhausted. The early 2026 consolidation market is a completely different animal.

Overfitting. The uncomfortable truth. If your EA was optimized on the same data it was backtested on (which is what happens with most MQL5 Market EAs), the backtest is measuring how well the optimizer found patterns in historical data — not how well the strategy predicts future price action. Walk-forward analysis and out-of-sample testing exist precisely to combat this problem.

Rule of thumb: expect live Profit Factor to be 20-40% lower than the backtest. If the backtest shows PF 2.0, plan for PF 1.2-1.6 in live trading. If that’s not acceptable, the strategy needs more edge — not more optimization.

How often should you review EA performance?

Reviewing too often leads to overreacting to normal variance. Reviewing too rarely means you catch problems after they’ve already cost you real money. Here’s the cadence I use and recommend:

Weekly (5 minutes): check net P&L, current drawdown, and number of trades. Is the EA trading? Is drawdown within expected bounds? Any failed trades or connection issues? This is a pulse check, not an analysis.

Monthly (30 minutes): calculate rolling Profit Factor, Sharpe ratio, and compare to historical averages. Review the equity curve shape. Check whether trade frequency has shifted significantly. Compare live metrics against backtest expectations. This is where you catch early-stage degradation.

Quarterly (1-2 hours): full strategy audit. Run a Monte Carlo simulation on the last 6 months of trades. Analyze heatmaps to see if profitable hours/days have shifted. Compare per-instrument breakdown if the EA trades multiple pairs. Decide whether to continue, adjust parameters, or retire the strategy.

Most degradation is gradual. A strategy doesn’t normally blow up overnight (unless it’s a grid or martingale, in which case… yes, it absolutely can). It loses edge slowly over 3-6 months. Monthly reviews are your early warning system.

When to stop an EA: red flags in your analytics

Knowing when to pull the plug on a strategy is just as important as knowing what metrics to track. These are the concrete signals that it’s time to cut — or at minimum, reduce position size.

1. Profit Factor drops below 1.0 for 2+ consecutive months. The strategy has lost its edge. It’s systematically losing money, not experiencing a normal drawdown. Continuing is hoping, not trading.

2. Live drawdown exceeds backtest max drawdown by 50% or more. If the backtest showed 15% max drawdown and you’re at 23%, the strategy is behaving outside its historical boundaries. The model it was built on may no longer apply.

3. Trade frequency drops by more than 60%. If the EA went from 20 trades/week to 8 with no parameter changes, the market conditions it was designed for have disappeared. It’s sitting on the bench — which may protect you from losses, but also means the edge has evaporated.

4. Sharpe ratio is negative for 3+ months. Negative Sharpe means the strategy is generating negative returns with high volatility. The worst possible combination. You’re paying for the privilege of maximum uncertainty.

5. Win rate and average trade size diverge simultaneously. If win rate drops from 55% to 40% while average loss grows from $80 to $140, the strategy is losing edge on both dimensions. This isn’t variance — it’s structural deterioration.

Retiring an EA isn’t failure. Markets evolve. The half-life of a retail trading strategy is 6-18 months before it needs recalibration or replacement. Building a system that detects this early is what separates professionals from gamblers.

Putting it all together

Performance tracking isn’t sexy. Nobody posts on forums “reviewed my Sharpe ratio this weekend.” But it’s the difference between traders who compound capital for years and those who blow accounts every 6 months wondering what went wrong.

Start with the 7 metrics. Build a weekly habit. Use a proper dashboard (whether it’s Myfxbook, a dedicated analytics platform, or even a spreadsheet). And be honest with yourself when the numbers say it’s time to stop.

The best EA in the world is only as good as the trader monitoring it.