Most trading journals track dozens of numbers and change no behavior at all. The problem isn’t too little data — it’s that the metrics people fixate on don’t answer the only question that matters: does this strategy make money per trade, and how confident can I be in that? A handful of KPIs answer it. The rest are decoration.
The vanity metrics to stop tracking
Some numbers feel like progress and tell you nothing actionable. They’re not wrong, they’re just not decisions.
- Total P&L — a scoreboard, not a diagnosis. It rises and falls for reasons the number itself can’t explain, and it’s dominated by your position size rather than the quality of your edge.
- Raw win count — a hundred small wins and one catastrophic loss can be a losing system. Counting wins ignores their size.
- Longest win streak — pure variance dressed up as skill. A coin-flip system produces streaks.
- Best trade ever — motivating, meaningless. You can’t repeat an outlier on demand.
The test for whether a metric is worth tracking: would a change in this number change what I do tomorrow? If not, it’s a vanity metric. Watching total P&L climb doesn’t tell you whether to size up, pause a strategy, or fix a leak. The metrics below do.
Expectancy: the one number that decides everything
Expectancy is the average amount you can expect to win or lose per trade, over many trades. It’s the single number that tells you whether a strategy is worth trading at all, because it folds win rate and reward-to-risk into one figure:
Expectancy = (Win rate × Average win) − (Loss rate × Average loss)
If it’s positive, the strategy makes money as trades accumulate. If it’s negative, no amount of discipline or position sizing saves it — you’re optimizing the rate at which you lose. Everything else on this page exists to sanity-check or contextualize expectancy.
Expressed in R-multiples — where 1R is the amount you risked on the trade — expectancy becomes portable across instruments and position sizes. An expectancy of +0.3R means you net roughly a third of your risk per trade on average, whether that risk was $50 or $500. Work an example through an expectancy calculator, and read the expectancy explainer and the R-multiple primer if the formula feels abstract. For a prop trader, expectancy in R is the cleanest way to know whether your edge is real before you scale into a firm’s limits.
Profit factor and average R as sanity checks
Expectancy can be gamed by memory and by outliers, so two companion metrics keep it honest.
Profit factor is gross profit divided by gross loss. Above 1.0 means your winners outweigh your losers in total; the further above, the more cushion you have. It’s a fast gut-check on the whole book — but on its own it hides how you got there, which is why it pairs with the next one.
Average R (and its cousins) breaks the edge into its parts:
- Average winning R — how big your winners are relative to risk.
- Average losing R — how well you cut losers. If this drifts past −1R, you’re not honoring stops.
- Average R across all trades — this is your expectancy in R terms, arriving from the other direction.
The value of looking at profit factor and average R alongside expectancy is that they reveal shape. Two strategies with identical expectancy can behave completely differently — one wins often and small, the other rarely and large — and that shape determines whether you can psychologically trade it and whether it survives a prop firm’s consistency rule.
Win rate in context, not in isolation
Win rate is the most over-weighted number in trading. A high win rate feels like skill, but it’s meaningless without the average win and average loss beside it. A 30% win rate can be wildly profitable if winners run several times your risk; an 80% win rate can bleed the account if the occasional loser wipes out many small wins.
Win rate only becomes useful when read together with reward-to-risk:
| Win rate | Reward:risk needed to break even |
|---|---|
| 30% | roughly 2.3 : 1 |
| 40% | roughly 1.5 : 1 |
| 50% | 1 : 1 |
| 60% | roughly 0.7 : 1 |
Those breakeven pairings show why win rate is a slider, not a target. Any point on the curve can be profitable; what matters is that your actual reward-to-risk clears the bar your win rate sets. A risk-reward calculator lets you check where a setup lands before you take it. Chase win rate alone and you’ll cut winners short to book more of them — quietly destroying the very reward-to-risk that made the system work.
How many trades before a metric means anything
Every metric on this page is a sample estimate, and a small sample lies. Ten trades tell you almost nothing; the number you compute could swing wildly on the next handful. Reacting to a metric built from a dozen trades is how good traders abandon good strategies and how bad strategies get scaled up right before they blow up.
There’s no magic threshold, but the discipline is the same: treat every metric as provisional until the sample is large enough to trust, and treat it as a range rather than a point. A confidence interval around your expectancy tells you how much the number could move with more data — and whether a recent dip is signal or noise. Shibiki computes a live Wilson confidence interval on your edge as trades post automatically, so you see not just the metric but how much to believe it yet. Until the range tightens, hold your conclusions loosely and let the sample grow.
Related: Expectancy calculator · Expectancy explained · R-multiple explained