Trading Metrics That Actually Tell You Something

A 68% win rate looks great until you check the math. A trader running a scalping strategy on EURUSD can win seven out of ten trades and still lose money by the end of the month. The winners average 8 pips each. The losers average 22. The win rate is real; the account is shrinking anyway.

That gap between "this number looks good" and "this strategy is actually working" is where most retail algo traders get stuck. They track one or two metrics, feel reassured, and don't realize the full picture looks different from the fragment they're watching. This article breaks down the trading metrics that matter, what values to look for, and why none of them work alone.

Performance metrics: is your strategy actually making money?

These are the numbers most traders check first. They tell you whether the strategy is profitable, but each one hides something.

Win rate is the percentage of trades that close in profit. If you take 100 trades and 55 close green, your win rate is 55%.

Good range: depends entirely on strategy type. Trend-following systems often run 35-45% and still make money because the winners are large. Scalpers and mean-reversion strategies need 60%+ because each winner is small. A "good" win rate without context is a meaningless number.

The trap: win rate says nothing about magnitude. You can win 80% of trades and go broke if your average loss is five times your average win. Traders who optimize for win rate often cut their winners early to "lock in the green," ending up with a metric that looks impressive on paper while the equity curve slopes downward.

Profit factor is gross profit divided by gross loss. If your winners totaled $12,000 and your losers totaled $8,000, the profit factor is 1.5.

Good range: anything above 1.0 means the strategy is net profitable. 1.3-1.7 is solid for a live system. Above 2.0 on a backtest is worth investigating because it often doesn't survive contact with real spreads and slippage. Below 1.0, and you're giving money to the market.

The trap: profit factor doesn't tell you anything about consistency. A profit factor of 1.5 could come from steady small wins or from a single massive outlier trade surrounded by mediocre performance. It also doesn't account for risk. A profit factor of 1.4 on 10:1 leverage is a very different proposition from 1.4 on 2:1.

Expectancy tells you how much you should expect to make, on average, per trade. The formula: (win rate × average win) minus (loss rate × average loss). If your win rate is 55%, average win is $200, and average loss is $150, expectancy is (0.55 × $200) - (0.45 × $150) = $42.50 per trade.

Good range: anything positive. $10-50 per trade is typical for retail strategies on moderate sizing. The actual dollar amount matters less than the sign; a positive expectancy applied consistently over hundreds of trades is what builds an account. Consistently applied negative expectancy is what empties one.

The trap: expectancy assumes consistent future behavior. It's calculated from past trades, so it's only reliable if market conditions remain similar and execution stays clean. A strategy with $40 expectancy in a trending market can flip to negative expectancy in a range-bound environment, and nothing in the numbers will warn you.

Average reward-to-risk (R:R) is your average winning trade divided by your average losing trade, expressed as a ratio. If winners average $300 and losers average $150, your R:R is 2:1.

Good range: for trend-following, 2:1 or higher is standard. For scalping and high-frequency approaches, 0.5:1 to 1:1 is normal because the win rate compensates. The ratio always works in tandem with the win rate. A 1:3 R:R (you lose three times what you win) is sustainable only if you win more than 75% of the time.

The trap: averages hide outliers. An "average R:R of 2:1" might mean every trade is roughly 2:1, or it might mean most trades are 0.5:1 with one monster winner skewing the average. The distribution matters more than the mean.

Maximum drawdown is the largest peak-to-trough decline in your account equity, measured as a percentage. If your account peaked at $10,000 and dropped to $7,200 before recovering, your max drawdown is 28%.

Good range: below 20% is conservative. 20-30% is aggressive but survivable for a retail trader who understands the risk. Above 30%, and you need to question whether the strategy's risk profile fits your actual tolerance, because a 30% drawdown requires a 43% gain to recover.

The trap: max drawdown is backward-looking and always represents the worst case so far. Your real max drawdown hasn't happened yet. Also, a shallow drawdown might just mean the strategy hasn't been tested in enough market conditions. A strategy with a 12% max drawdown across two months of backtesting tells you almost nothing about what it will do across a full cycle.

Risk-adjusted metrics: are the returns worth the volatility?

Win rate and profit factor tell you what happened. Risk-adjusted metrics tell you whether the returns justified the ride. Most retail traders don't calculate these, but you'll encounter them in prop firm dashboards and any serious fund reporting.

Sharpe ratio measures excess return per unit of total volatility. The higher the number, the higher the return you get per unit of risk you take. It's calculated as (strategy return minus risk-free rate) divided by the standard deviation of returns.

Good range: above 1.0 is acceptable, above 2.0 is strong, above 3.0 is exceptional (and worth double-checking, because strategies with Sharpe above 3 over long periods are rare). Below 0.5 and you're taking a lot of risk for not much return.

The catch: Sharpe penalizes all volatility equally, including upside volatility. A strategy that occasionally produces very large wins will have a lower Sharpe ratio than a strategy with steady, modest returns, even if the first one makes more money. It also assumes returns are normally distributed, which trading returns rarely are.

Sortino ratio fixes the main problem with Sharpe: it only penalizes downside volatility. The formula replaces standard deviation with downside deviation, so big wins don't count against you.

Good range: roughly the same thresholds as Sharpe, but Sortino numbers will typically run higher for the same strategy because upside volatility is excluded. If your Sortino is significantly higher than your Sharpe, it means much of your volatility is coming from large wins, which is usually a good sign.

Both ratios matter more over longer periods. Calculating Sharpe on two weeks of trades is meaningless. You need months of data before these numbers stabilize. They're also useful for comparing different execution approaches. If you're deciding between copy trading and bot trading, running both for a month and comparing their Sortino ratios tells you more than comparing raw P&L.

Trading metrics prop firms actually evaluate you on

If you're running automated strategies through a prop firm evaluation, the metrics the firm tracks are not the same ones you'd look at to judge your strategy. Prop firms care about compliance, not alpha. You can have the best expectancy in the building and still fail the evaluation because your drawdown touched a floor you didn't know was moving.

Maximum drawdown works differently in prop firms than in personal trading. Most firms set a hard limit, typically 8-12% of the initial account balance. Cross it, and the evaluation ends. Some firms calculate this as a static floor from your starting balance; others trail it upward as your equity grows, which means every profit raises the floor you can't drop below. The math is specific to each firm, and getting it wrong by even a fraction ends the run. Understanding how drawdowns in prop firms are calculated under static and trailing variants is the difference between knowing your floor and guessing at it.

Daily loss limit caps how much you can lose in a single trading session, typically 4-5% of your balance. This is the rule that kills evaluations on news days. One XAUUSD spike, one oversized position, and you've eaten the day's entire allowance in a single candle. The limit resets every session, but the evaluation doesn't un-fail.

Consistency score is the metric most traders don't discover until it fails them. Some prop firms require that your profit distribution is reasonably even, so a single massive trading day can't carry the entire evaluation. If 40% of your profit came from one day and the other 29 trading days produced the remaining 60%, some firms will reject that, even if the total P&L hits the target. The rule is designed to verify that you have a repeatable process, not a lucky streak. Prop firm risk management rules vary: FTMO, Funding Pips, and others each have their own thresholds.

Profit target is straightforward: hit X% profit to pass the phase. Typical targets are 8-10% for Phase 1 and 5% for Phase 2. The catch is that chasing the target with bigger lot sizes near the end of a phase is the exact behavior most likely to trigger a drawdown breach. The metrics fight each other. You need to hit the target without crossing any of the floors, and sizing is the variable that links the two.

The broader point: your strategy might be genuinely profitable, but the prop firm doesn't care about your Sharpe ratio or your expectancy per trade. It cares about whether you stayed inside the rules every single day. This makes execution precision matter more than it does in personal trading, because every ghost position, every duplicate fill, every stop that lands a few points off the calculated price can be the difference between passing and failing.

Execution quality metrics: where your backtest meets the real market

Most of the metrics above are things you can calculate from a backtest. They look clean, precise, and reproducible in the simulator. Then you run the strategy live, and the numbers start drifting.

The drift happens for specific, measurable reasons. Your backtest assumed perfect fills at the signal price. Live fills include spread, slippage, and broker-specific requoting. Your backtest assumed instant execution. Live execution has latency from TradingView to your broker, through whatever bridge sits in between. Your backtest assumed every signal results in exactly one trade. Live execution sometimes produces duplicates, phantoms, or missed fills.

These differences compound. Across 200 trades, a consistent 1.5-pip slippage on entries alone can turn a profitable backtest into a breakeven live strategy. The problem is that most traders never see this data because their broker statement doesn't itemize slippage per trade, and their TradingView logs don't show what the broker actually did.

Average slippage per trade is the gap between the price your strategy intended and the price you actually got. If your strategy signaled a buy at 1.0842 on EURUSD and the fill came back at 1.0844, that's 0.2 pips of slippage on that trade. Tracked across hundreds of trades, the average tells you how much the market is "taxing" your strategy on top of the spread.

Good range: 0.1-0.5 pips on major forex pairs through a decent broker is normal. Consistently above 1.0 pip means the broker is slow, the strategy is trading in low-liquidity windows, or the signal latency is too high. On XAUUSD, wider slippage is expected because the spread itself is larger.

Signal-to-fill latency is how long it takes from the moment your strategy fires an alert to the moment the trade is open on your broker account. This includes TradingView processing time, webhook delivery, bridge processing, and broker execution. If you're using an automated trading setup with a webhook pipeline, every link in that chain adds time.

Good range: under 500ms total is good for webhook-based retail execution. 500ms-1s is acceptable. Above 1s and you're trading on stale prices, which inflates slippage. For most retail strategies trading on 5-minute or higher timeframes, latency in the hundreds of milliseconds won't materially affect your results. For scalpers on 1-minute bars, it can.

Fill accuracy is the question underneath all other metrics: did the trade that landed on your broker account match what your strategy actually intended? Right size, right direction, right stop-loss, right take-profit. If any of these differ from the signal, every other metric you compute from that trade is slightly wrong.

This is where the gap between backtest and live stops being abstract. If your stop-loss was supposed to be at 1.0820 but landed at 1.0817 because of rounding or requoting, your average R:R in the journal doesn't match your backtest's average R:R. If a phantom signal opened a position your strategy never intended, your win rate, drawdown, and expectancy all shift.

FillEdge's trade journal captures this automatically. Every signal is reconciled against the fill that followed: intended price vs. actual price, intended stop vs. actual stop, with per-trade slippage and latency. Each trade carries a status badge (✓MATCHED, 🎯LOCKED, 👻CAUGHT, 🛡️BLOCKED, 🔀REORDERED, 💀EXPIRED) so you can see at a glance whether the execution was clean.

When you calculate metrics from that journal, you know the underlying data is verified against the original signals, not reconstructed from a broker statement that doesn't know what your strategy intended.

Without this kind of verification, you're computing trading metrics from noisy data. The metrics look real. They might be off by enough to mislead you.

Why single trading metrics lie: read them in combinations

A single metric is a flashlight in a dark room. It shows you one wall. You need several to see the shape of the space.

Here are four diagnostic combinations that tell you more than any individual number:

High win rate + low profit factor. You're winning often, but your winners are small relative to your losers. This is the classic "cut winners early, let losers run" pattern. The fix is almost always on the exit side: your entries might be fine, but you're closing too soon when the trade goes your way and holding too long when it goes against you. A 72% win rate with a profit factor of 0.9 means you're net losing money despite winning most of your trades.

Low win rate + high average R:R. This is normal trend-following behavior. You lose more often than you win, but when you win, the trade covers multiple losses. If your expectancy is positive and your drawdown is inside your tolerance, don't panic after a losing streak; the strategy is working as designed. A 38% win rate with a 3:1 R:R produces a positive expectancy of about 0.52R per trade. The losing streak is the price of admission.

Positive expectancy + high max drawdown. The strategy makes money over time, but the path is rough enough to threaten the account. This is a sizing problem, not a strategy problem. The expectancy is real; the lot size is too large for the account's ability to absorb the inevitable losing sequences. Scale down, because the math still works at a smaller size, and destroying the account on a drawdown means the positive expectancy never gets the sample size it needs to play out.

Decent profit factor + rising average slippage. The strategy logic is sound, but execution quality is decaying. This can happen when a broker widens spreads during your typical trading hours, when your signal latency creeps up, or when market conditions shift and your entries are landing in thinner order books. The fix isn't in the strategy's code; it's in the execution pipeline. Check your latency trend, check your broker's spread history, and check whether your signals are still arriving fresh or getting stale.

Building a trading metrics habit that works

Knowing what to measure is the easy part. Most traders learn the definitions, check their metrics once after reading an article like this one, and then never open the spreadsheet again. The metrics that actually change your trading are the ones you review consistently.

A ten-minute weekly review beats a two-hour monthly deep dive. Pick a fixed time. Sunday evening before the week opens is a popular choice.

Look at three numbers first: expectancy over the trailing 30 trades, max drawdown over the same period, and average slippage per trade. If all three are stable, you're on track. If any one of them has shifted, that's your signal to dig deeper into the journal.

The advantage of automation is also the risk: the system runs without you. Traders who create trading bots for TradingView and deploy them across broker accounts can go weeks with degrading execution before anyone notices. The weekly review catches drift before it compounds. And if you're running automated trading across multiple accounts, per-account trading metrics expose which broker or which strategy is underperforming, something that's invisible when you only look at combined P&L.

Don't optimize for more metrics. Optimize for checking the right ones regularly. Expectancy tells you whether the strategy still works. Drawdown tells you whether you can survive long enough for it to play out. Slippage indicates whether execution is preserving the edge your strategy created.

Those three, tracked weekly, are worth more than a dashboard full of numbers you glance at once and forget.

FAQ

What is a good profit factor for a trading strategy?

A profit factor above 1.0 means the strategy is net profitable; 1.3-1.7 is a solid range for a live system trading real spreads and slippage. Above 2.0 on a backtest is worth questioning, because that number tends to shrink once you account for execution costs and fill quality in live trading. Below 1.0, and the strategy is losing money regardless of the win rate.

How do you calculate expectancy in trading?

Expectancy = (win rate × average win) minus (loss rate × average loss). For example, a strategy with a 55% win rate, $200 average win, and $150 average loss has an expectancy of (0.55 × 200) - (0.45 × 150) = $42.50 per trade. A positive number means the strategy should make money over a large enough sample of trades; a negative one means it will lose money no matter how many trades you take.

What trading metrics do prop firms track?

Prop firms evaluate you on rule compliance, not strategy quality. The metrics that end evaluations are maximum drawdown (typically 8-12% of account balance, sometimes trailing), daily loss limit (usually 4-5% per session), consistency score (profit can't be concentrated in a single day), and profit target (8-10% for Phase 1, 5% for Phase 2 at most firms). Your Sharpe ratio and expectancy per trade don't matter to the firm if you cross any of these floors.

What is the difference between Sharpe ratio and Sortino ratio?

Both measure risk-adjusted return, but they define "risk" differently. The Sharpe ratio divides excess return by total volatility, so it penalizes large wins the same way it penalizes large losses. Sortino ratio uses only downside deviation, so big winning trades don't drag the number down, making it a better fit for strategies that are profitable but volatile on the upside.