Learn / Risk & Position Sizing
Reading an equity curve honestly
Two strategies can end the year at the same balance and be completely different things to hold. The ending number is the least informative part of an equity curve — the shape is where the information is.
What to look at, in order
1. The largest drawdown, and what recovery it required.
Not the drawdown percentage — the recovery it implied. A 40% drawdown needed 67%; a 60% drawdown needed 150%. Run it through the drawdown recovery calculator.
Then the question that matters: would you have kept executing through it? A curve containing a drawdown you would have abandoned is a curve you would not have.
2. Time underwater.
How long between a peak and its recovery. A 20% drawdown recovered in three weeks and one recovered over eight months are different experiences, and only one of them is survivable in practice.
Time underwater is the metric that best predicts whether a strategy gets abandoned, and it is almost never reported.
3. Whether returns are concentrated.
Does a handful of trades produce most of the gain? If so, the curve is flat with occasional jumps, and:
- The sample needed to evaluate it is much larger — see how many trades before you can judge a strategy.
- Missing a few of those trades changes the result completely.
- The flat stretches feel like failure while they are happening.
4. Whether the shape is stable.
Steady early and volatile later suggests something changed — market regime, position sizing drift, or discipline. A curve that changes character mid-series is two strategies plotted as one.
Three ways a curve flatters
Compounding hides early weakness. A strategy that was mediocre for a year and good for the next looks excellent, because the good period compounds on a base the mediocre period preserved. The visual is dominated by the recent part.
Plot it on a log scale. A constant percentage return is a straight line there, and the flattering curvature disappears.
Contributions look like returns. An equity curve including deposits shows a rising line that is partly your own money. Plot the return series separately from the balance, or the two are indistinguishable.
Sequence luck looks like skill. The same set of returns in a different order produces a different-looking curve — and with contributions, a different ending balance too. See sequence of returns risk. A smooth curve may be a lucky ordering of a rough distribution.
What a smooth curve does not prove
A suspiciously smooth curve deserves a specific question: what is being deferred?
Common answers:
- Losers held rather than cut. Unrealised losses do not appear in a realised equity curve. The curve is smooth because the losses have not been taken yet.
- Position sizes shrinking during difficulty, which flattens the curve by reducing participation rather than by improving decisions.
- A short sample that has not met its bad regime.
The first is the dangerous one, and it is why an equity curve on realised PnL alone is misleading. Mark positions to market.
The comparison worth making
Not “which curve ends higher” but:
| Strategy A | Strategy B | |
|---|---|---|
| Ending balance | Same | Same |
| Max drawdown | 15% | 45% |
| Recovery required | 17.6% | 81.8% |
| Longest time underwater | 6 weeks | 9 months |
| Return concentration | Spread | Top 5 trades |
Same result, and B is a far worse thing to have held — larger recovery, longer underwater, and more dependent on a handful of trades that might not recur.
Which means B’s result is also less repeatable, because more of it was determined by whether those few trades happened to land.
What to do with it
Size from the drawdown you can tolerate, not from the return you want. The drawdown is the binding constraint — see risk of ruin.
Check whether your risk per trade is stable across the curve. Rising risk during drawdowns is the most common path to ruin for accounts with a real edge, and it is visible as increasing volatility in the recovery legs.
Compare your curve against your journal. Periods of unusual shape should correspond to something recorded — a regime, a rule break, a change in sizing. If they do not, you do not know what produced them, which means you cannot repeat or avoid them. See what to record in a trading journal.
FAQ
What should I look for in an equity curve?
The largest drawdown and the recovery it required, how long the curve spent below a prior peak, whether returns are concentrated in a few trades, and whether the shape stays consistent. The ending balance is the least informative part, since very different experiences produce the same one.
Is a smooth equity curve a good sign?
Not necessarily, and it is worth asking what is being deferred. Losers held rather than cut do not appear in a realised curve, so smoothness can mean losses are pending. Marking positions to market is the check, along with asking whether the sample has met a bad regime yet.
Why does my equity curve look better than my results feel?
Often because compounding makes recent performance dominate the visual, or because deposits are included alongside returns. Plotting on a log scale and separating the return series from the balance both remove the flattery.
How much drawdown should I expect?
Enough that you should decide in advance what you would keep trading through, since that number is the binding constraint on position sizing. Work it backwards: take your expected longest losing streak at your risk per trade, and check the recovery it implies against what you would actually attempt.