# How to Analyze Your Trading Performance

> How to analyze trading performance in the right order: confirm the edge, see how it pays, test if you can survive it, then find where it comes from and what leaks.

Canonical: https://www.journalx.io/guides/how-to-analyze-your-trading-performance

To analyze your trading performance, read your stats in a fixed order. First confirm there is an edge (expectancy in R, over enough trades). Then see how it pays, whether you can sit through its drawdowns, which setups and sessions produce it, what is leaking, and whether it still holds on recent trades.

Knowing how to analyze trading performance is mostly knowing what to read first. A trading dashboard shows twenty numbers at once, and the loudest ones (total P\&L, win rate) are the least useful to start with. Read in the wrong order and you end up fixing your entries when the real problem is a single setup, or a habit of moving stops, or fees eating a thin edge.

This guide gives you the order. Each step asks one question, names the few metrics that answer it, and tells you what to do with the answer before moving on. It is a deeper pass than the [weekly trade review](/guides/the-weekly-trade-review-process): run it monthly or quarterly, or every 50 to 100 trades, on a sample big enough for the numbers to mean something.

![The order to read your trading stats: is there an edge, how does it pay, can you sit through it, where does it come from, what is leaking, and is it holding up.](https://assets.journalx.io/marketing/guides/how-to-analyze-your-trading-performance/order-to-read-trading-stats.avif)

## Before you start: get the data in shape

The analysis is only as good as the trade log under it. Before reading a single metric, check three things.

- **Every trade is in.** Include the scratches, the revenge trades, and the ones on the second account. A sample that leaves out the trades you'd rather forget will flatter you.
- **Every trade has a planned risk.** You need the dollar amount you stood to lose at the stop, because that is what turns results into [R-multiples](/glossary/r-multiple). A $300 win on $100 of risk is +3R; on $600 of risk it is +0.5R. Without R, you are comparing trades of different sizes as if they were the same.
- **Every trade has a setup tag and a "followed the plan" flag.** Steps 4 and 5 depend on both. If your journal doesn't capture them yet, [what to track in a trading journal](/blog/what-to-track-in-a-trading-journal) lists the fields worth adding.

Net of fees, too. Commissions and exchange fees belong in every trade's result, not in a separate line you look at once a year.

## Step 1: Is there an edge?

Start with [expectancy](/glossary/expectancy), the average result per trade across wins and losses together. In R, it reads as "on average, each trade made or lost this fraction of what I risked." A strategy that wins 45 percent of the time with 2R winners and 1R losers has an expectancy of (0.45 × 2) − (0.55 × 1) = +0.35R. If you want to work it by hand, the [expectancy calculation guide](/guides/how-to-calculate-expectancy) walks a 20-trade example.

Then ask how many trades that number rests on. Expectancy from 25 trades can swing by half an R either way through luck alone. A useful rule from our post on [how many trades you need to judge a strategy](/blog/how-many-trades-to-evaluate-a-strategy) is that you need roughly (2 × standard deviation ÷ expectancy)² trades before a positive result sits clearly above zero. With a 1.5R standard deviation and +0.25R expectancy, that is about 144 trades.

**What to do with the answer.** If expectancy is clearly positive on a decent sample, carry on. If it is negative or the sample is too small, the rest of the analysis is about diagnosis, not confirmation, so read the next steps looking for where the money goes.

## Step 2: How does the edge pay?

The same expectancy can come from very different trading. Break it into its parts: [win rate](/glossary/win-rate), [average win](/glossary/average-win), [average loss](/glossary/average-loss), and [profit factor](/glossary/profit-factor) (gross profit divided by gross loss).

|               | Trader A | Trader B |
| ------------- | -------- | -------- |
| Win rate      | 62%      | 35%      |
| Average win   | 1.0R     | 2.6R     |
| Average loss  | 1.0R     | 1.0R     |
| Expectancy    | +0.24R   | +0.26R   |
| Profit factor | 1.63     | 1.40     |

Both traders make about a quarter of an R per trade. They will live through it very differently. Trader B loses almost two trades in three and depends on a few large winners, so cutting those winners short would wipe out the edge. Trader A wins often with small, even payoffs, so a single oversized loss does the damage.

This is also the step where a high win rate stops fooling you. A 70 percent win rate still loses money when the average loss is more than about 2.3 times the average win, which is the whole argument of [why a 70% win rate can still lose money](/blog/win-rate-vs-expectancy).

**What to do with the answer.** Write down which kind of strategy you have. It tells you which mistake would hurt most: for a low win rate system, exiting winners early; for a high win rate system, letting a loser run past the stop.

## Step 3: Can you sit through it?

An edge is only useful if you keep trading it through the bad stretches. Read three things together.

- **[Maximum drawdown](/glossary/drawdown):** the largest peak-to-trough fall in your account, in percent and in R. Ask whether you could have kept following the rules at the bottom of it.
- **Longest losing streak:** Trader B from the table, over 200 trades, should expect a run of about 10 straight losses. Trader A's longest is likely around 5. Our post on [how long a good strategy can keep losing](/blog/losing-streaks) has the estimate and a table by win rate.
- **The [equity curve](/glossary/equity-curve):** its shape, not just its end point. A curve that climbs on two or three huge days and drifts sideways the rest of the time is fragile, because missing one of those days changes everything.

**What to do with the answer.** If the drawdown or the streak is bigger than you can stomach, the fix is usually size, not strategy. Lowering your [risk per trade](/glossary/risk-per-trade) shrinks every drawdown in proportion without touching the edge.

## Step 4: Where does the edge come from?

Now split the results. A blended expectancy of +0.15R can be one setup at +0.5R carrying another at −0.3R. Slice by one dimension at a time and read expectancy and trade count for each group:

- **[Setup](/glossary/setup):** the most useful cut, and the reason every trade needs a setup tag.
- **Market or symbol:** some instruments suit your style; others just add noise.
- **Time of day and day of week:** the first 15 minutes after the open and the last hour often behave differently from the rest.
- **Long vs short:** many traders have an edge on one side only.
- **Holding time:** quick scalps and multi-day holds rarely share the same numbers.

Watch the sample size in every slice. If you cut 200 trades by setup, then by session, then by side, each group might hold a dozen trades, and a dozen trades can show you anything. The more ways you slice, the more likely one slice looks great by chance alone. Treat a standout slice as a question to test going forward, not a conclusion.

**What to do with the answer.** Look for a slice that is clearly negative on a decent sample. Cutting a losing setup is usually the cheapest improvement you'll find, because it takes nothing away from the trades that work.

## Step 5: What is leaking?

This step compares what you planned with what you did. Three leaks are common and each has its own metric.

**Rule breaks.** Split all trades into "followed the [trading plan](/glossary/trading-plan)" and "broke a rule," then read expectancy for each group. The gap between the two is what your discipline costs you per trade. A trader whose plan-followed trades make +0.4R and rule-break trades lose −0.6R doesn't need a new strategy.

**Exits.** Compare each trade's result with its [maximum favorable excursion](/glossary/maximum-favorable-excursion), the furthest it went in your favor. If winners routinely reach +2R and close at +0.8R, you are giving back more than you keep. The mirror metric, [maximum adverse excursion](/glossary/maximum-adverse-excursion), shows whether your stops sit where your working trades tend to go. Taking profits quickly while holding losers is a well-documented habit called the [disposition effect](https://en.wikipedia.org/wiki/Disposition_effect), so check your own numbers for it.

**Costs.** Add up fees and [slippage](/glossary/slippage) and express them in R. Paying $5 round trip on trades that risk $50 costs 0.1R a trade, which is most of a thin edge before you have done anything wrong.

**What to do with the answer.** Rank the leaks by cost per trade multiplied by how often they happen. The biggest one is your next change.

## Step 6: Is the edge holding up?

An all-time average can hide a strategy that stopped working months ago. Plot expectancy over a rolling window, for example the last 50 trades, and lay it against the all-time number.

![Rolling 50-trade expectancy over 300 simulated trades: the all-time average stays at +0.15R while the rolling line falls below zero after trade 180.](https://assets.journalx.io/marketing/guides/how-to-analyze-your-trading-performance/rolling-expectancy.avif)

In this simulated record, trades 1 to 180 average +0.37R and trades 181 to 300 average −0.18R. The all-time figure still reads a comfortable +0.15R, and anyone checking only that number would keep trading a strategy that has been losing for 120 trades.

Read the rolling line with some patience. A 50-trade window swings a lot even when nothing has changed, as the early part of the chart shows. What deserves attention is a drop that stays down, ideally lining up with something you can name: a new market regime, a rule you quietly changed, a different instrument.

**What to do with the answer.** If recent results are clearly weaker, compare the recent trades with the earlier ones setup by setup before you change anything. Often one setup has faded while the rest are fine.

## Step 7: Pick one change and test it

Finish by writing down one change, the largest leak or the weakest slice you found, phrased so you can check it later: "stop trading setup C," "no entries in the first 15 minutes," "risk 0.5% instead of 1% until drawdown recovers." Then give it enough trades to judge before you analyze again.

One change at a time is the point. Change three things at once and the next analysis can't tell you which one helped. If the change is a new rule rather than a cut, test it on past data first; the [backtesting guide](/guides/how-to-backtest-a-trading-strategy) covers how to do that without fooling yourself.

## Putting it together

Analyzing trading performance is a sequence of questions, each answered before the next one makes sense. Confirm there is an edge on enough trades. Understand how it pays and what that makes you vulnerable to. Check that you can live through its worst stretches. Find which slices produce it, which leaks drain it, and whether it is still there on recent trades. Then change one thing. Run the same sequence every month or quarter and the numbers stop being a scoreboard and start telling you what to do next.

JournalX shows expectancy, win rate and profit factor for any filter you build (by setup, session or symbol), so each step of this analysis takes a few clicks.