# The Trading Performance Metrics Worth Tracking

> Trading performance metrics grouped by the question each one answers, from edge to drawdown, sizing and execution. Formulas, good ranges and a worked example.

Canonical: https://www.journalx.io/blog/trading-performance-metrics

Trading performance metrics answer five questions: do I have an edge (expectancy, profit factor), what do my wins and losses look like (win rate, average win and loss), how rough is the ride (drawdown, Sharpe), can my sizing survive it (risk per trade, risk of ruin), and am I executing the plan (MAE, MFE, slippage, rule adherence).

Most trading dashboards show twenty numbers at once, and most traders read the two that feel best that week. The fix is not more trading performance metrics. It is knowing which question each one answers, so you can read them in order and stop when one fails. This post groups them that way, with the formula for each, then runs all of them on one example journal.

![The five questions trading performance metrics answer, read in order: do I have an edge (expectancy, profit factor, SQN), what do wins and losses look like (win rate, average win, average loss, win/loss ratio, R-multiple), how rough is the ride (max drawdown, equity curve, recovery factor, Sharpe, Sortino), can I survive my sizing (risk per trade, position sizing, risk of ruin, Kelly criterion, max daily loss), and am I executing the plan (MAE and MFE, slippage, holding period, results by setup, plan followed %).](https://assets.journalx.io/marketing/blog/trading-performance-metrics/trading-performance-metrics-map.avif)

## Measure everything in R first

Before any metric, put every trade on the same scale. An [R-multiple](/glossary/r-multiple) is a trade's result divided by the amount you planned to risk, the distance from entry to your [stop-loss](/glossary/stop-loss) times position size. A $300 win on $100 of risk is +3R, a full stop-out is −1R.

Dollar figures mix up how well the strategy works with how big you happened to trade. In R, a strategy that kept working after you doubled your size looks the same in both periods. Every metric below works in R.

## 1. Do I have an edge?

- **[Expectancy](/glossary/expectancy)** is the average result per trade: (win rate × average win) − (loss rate × average loss). Above zero means the strategy made money per trade over the sample. It is the single most important number in the journal. Our guide on [how to calculate expectancy](/guides/how-to-calculate-expectancy) works it by hand and in a spreadsheet.
- **[Profit factor](/glossary/profit-factor)** is gross profit ÷ gross loss. Above 1.0 is profitable, and 1.5 to 2.0 is generally considered strong. It tells the same story as expectancy from a different angle: dollars made per dollar lost.
- **[System Quality Number (SQN)](/glossary/system-quality-number)** is average R ÷ standard deviation of R × √trades, usually capped at 100. Van Tharp's scale calls 2.0 to 2.4 average and 2.5 to 2.9 good. It answers a sharper question than expectancy: is the edge big compared with the noise around it?

None of these mean much on a small sample. A typical setup needs [100 to 400 trades](/blog/how-many-trades-to-evaluate-a-strategy) before a positive expectancy stands clear of zero.

## 2. What do my wins and losses look like?

These show how the edge is built, which tells you what to protect.

- **[Win rate](/glossary/win-rate)** is winning trades ÷ total trades.
- **[Average win](/glossary/average-win)** and **[average loss](/glossary/average-loss)** are the typical size of each. Their ratio, average win ÷ average loss, is often called the payoff ratio.
- **[Win/loss ratio](/glossary/win-loss-ratio)** is the count of wins ÷ the count of losses. It is not the payoff ratio, and mixing the two up is common.
- **[Risk-reward ratio](/glossary/risk-reward-ratio)** is the planned version of the payoff: stop distance against target distance, set before entry. Comparing planned reward with realized average win shows whether your exits deliver what the plan promised.

The key relationship is the breakeven win rate: 1 ÷ (1 + payoff ratio). With 2R winners and 1R losers you break even at 33% winners. That is why a high win rate on its own proves nothing, which we work through in [why a 70% win rate can still lose money](/blog/win-rate-vs-expectancy).

## 3. How rough is the ride?

Two strategies with the same expectancy can feel completely different to trade. These metrics measure the path.

- **[Maximum drawdown](/glossary/drawdown)** is the largest peak-to-trough fall in equity. It is the number that tests whether you will keep following the rules.
- **[Equity curve](/glossary/equity-curve)** is the picture behind every number here. Look at it before the ratios: long flat stretches, a single trade carrying the whole result, or a slope that changed halfway through all show up there first.
- **[Recovery factor](/glossary/recovery-factor)** is net profit ÷ maximum drawdown: how many times the strategy earned back its worst fall. It grows with time, so only compare it over equal periods.
- **[Sharpe ratio](/glossary/sharpe-ratio)** is average return above the risk-free rate ÷ standard deviation of returns, usually from daily returns and annualized. [William Sharpe's own definition](https://web.stanford.edu/~wfsharpe/art/sr/sr.htm) is the reference. Above 1 is commonly treated as good, but only over a long record.
- **[Sortino ratio](/glossary/sortino-ratio)** is the same idea using only downside deviation, so a strategy with big winning days isn't penalized for them.

Losing streaks belong here too. A strategy that wins 40% of the time typically sees a longest streak of about 11 losses in 500 trades, and one trader in ten sees 14. Our post on [losing streaks](/blog/losing-streaks) has the table by win rate.

## 4. Can I survive my sizing?

An edge only pays if you are still trading when it shows up.

- **[Risk per trade](/glossary/risk-per-trade)** is the share of the account lost if the stop is hit, usually 0.5% to 2%. It turns a drawdown in R into a drawdown in percent: at 1% risk, an 8R drawdown costs about 8% of the account.
- **[Position sizing](/glossary/position-sizing)** is how that risk becomes a quantity: risk budget ÷ stop distance. Track it to catch trades where size drifted from the rule.
- **[Risk of ruin](/glossary/risk-of-ruin)** is the probability that a normal losing run takes the account past a limit you can't trade through. It is driven far more by risk per trade than by the strategy.
- **[Kelly criterion](/glossary/kelly-criterion)** is W − (1 − W) ÷ R, the growth-maximizing risk per trade. It is useful as a ceiling, not a target, since it assumes you know your true win rate and payoff exactly.
- **[Max daily loss](/glossary/max-daily-loss)** is a rule rather than a metric, but logging how often you hit it, and what happened on the days you ignored it, belongs in the review.

## 5. Am I executing the plan?

The last group separates the strategy from the person running it.

- **[Maximum adverse excursion (MAE)](/glossary/maximum-adverse-excursion)** is how far each trade went against you. It shows whether stops are too tight or too loose.
- **[Maximum favorable excursion (MFE)](/glossary/maximum-favorable-excursion)** is how far each trade went your way. Compared with the exit, it shows how much of the move your exits keep.
- **[Slippage](/glossary/slippage)** is fill price against intended price. Measured in R, it shows what execution costs per trade, which can be most of a thin edge.
- **[Holding period](/glossary/holding-period)** is time in the trade. Closing winners fast while letting losers run long is a pattern worth catching early.
- **Results by [setup](/glossary/setup)** split every metric above by the pattern you traded. A blended expectancy often hides one setup paying for another.
- **Plan followed %** is the share of trades that matched your [trading plan](/glossary/trading-plan). Then compare expectancy on followed trades with expectancy on rule breaks.

## One journal, every metric

Here is a hypothetical 100-trade sample run through every group. Each trade risked 1% of the account.

| Question        | Metric                     | Result        | Reading                                    |
| --------------- | -------------------------- | ------------- | ------------------------------------------ |
| Edge            | Expectancy                 | +0.40R        | Positive, about 40R over 100 trades        |
| Edge            | Profit factor              | 1.72          | 94.5R won against 55R lost                 |
| Edge            | SQN (SD of 1.55R)          | 2.5           | Good on Tharp's scale                      |
| Wins and losses | Win rate                   | 45%           | 45 wins, 55 losses                         |
| Wins and losses | Average win / average loss | +2.1R / −1.0R | Payoff ratio of 2.1                        |
| Wins and losses | Win/loss ratio             | 0.82          | More losses than wins, and that is fine    |
| Wins and losses | Breakeven win rate         | 32%           | 13 points of room below the actual 45%     |
| Ride            | Maximum drawdown           | 8R            | About 8% of the account at 1% risk         |
| Ride            | Recovery factor            | 4.9           | 39.5R net ÷ 8R drawdown                    |
| Ride            | Longest losing streak      | 7 trades      | Inside the normal range for a 45% win rate |
| Sizing          | Full Kelly                 | 19%           | Far above the 1% actually risked           |

Read in order, this is a clear story. The edge is real and decent for its sample size. It comes from a payoff ratio of 2.1, not from being right often, so the thing to protect is the size of the winners: cutting them early would hurt more than a lower win rate. The drawdown is modest at 1% risk, and full Kelly sits far above the risk taken, so sizing is not the problem. That is not a reason to size up.

What the table cannot say is whether the trader followed the plan. That is the job of the fifth group, and it is where most of the useful changes come from.

Read the groups in order and stop when one fails. A negative expectancy makes the Sharpe ratio irrelevant. A drawdown you can't sit through makes a good expectancy irrelevant. Most review time is wasted polishing the later numbers while an earlier one is broken.

## Mistakes that make good metrics lie

- **Judging on too few trades.** Twenty trades can show a 70% win rate from a coin-flip strategy. Track the count next to every metric.
- **Blending setups.** One expectancy across four setups is an average of four different businesses. Split before deciding anything.
- **Leaving out costs.** Commissions and slippage come off every trade. A +0.15R edge before costs can be zero after them.
- **Watching one number.** Tight targets raise win rate and shrink payoff. Any single metric can be bought by making another worse.
- **Comparing different periods.** Recovery factor, net profit and streak length all grow with time. Compare like with like.

For a routine that uses these metrics in order every month, see [how to analyze your trading performance](/guides/how-to-analyze-your-trading-performance) and the [monthly trading review](/guides/monthly-trading-review). For a sense of how standard these figures are, the [MetaTrader 5 strategy tester report](https://www.metatrader5.com/en/terminal/help/algotrading/testing_report) prints profit factor, expected payoff, recovery factor and Sharpe ratio for every backtest.

JournalX calculates expectancy, win rate, profit factor and drawdown from your synced trades, and its filters let you split any of them by setup, session or whether you followed your Gameplan.