How to read an equity curve like a fund manager

Backtest software paints you a green line going up and to the right and your brain does the rest. "It works," you think. But the equity curve is a far richer document than its endpoint. A good reader can tell you in 30 seconds whether the strategy is a slow-grinder, a feast-and-famine outlier hunter, or a leveraged time bomb. Here's how to do it.

Step 1: read the shape, not the endpoint

Two equity curves can end at the same final value with completely different stories. One curve climbs in a straight diagonal from start to finish; the other spends 70% of the time underwater, then explodes upward in the last quarter. The endpoints match. The strategies are not the same — the second has a far higher chance of being abandoned in real life before the explosion arrives.

Look for monotonicity. The cleanest curves grind upward in a band. Curves with long flats, deep drawdowns, or single jumps are warning signs — the first two suggest the strategy doesn't actually work most of the time, the third suggests the result depends on one or two outlier trades.

Step 2: stare at the drawdowns

The maximum drawdown is the single most important number on the page. It's the largest peak-to-trough decline the strategy produced over the test period. Three rules:

• Whatever max drawdown the backtest shows, expect to see at least 1.5× that in live — your sample is finite and the worst case hasn't happened yet.
• Track time-to-recover, not just depth. A 20% drawdown that recovers in three weeks is very different from a 20% drawdown that takes 18 months.
• Run the trade-shuffled Monte Carlo. Re-order the trades randomly 1,000 times and look at the 95th-percentile drawdown of the bootstrap. That's your honest worst case.

Step 3: MAR ratio

MAR = CAGR / |max drawdown|. A strategy with 30% CAGR and 50% max drawdown has a MAR of 0.6 — you're being asked to live through a 50% loss to get 30% a year. Bad ratio. A strategy with 18% CAGR and 12% max drawdown has a MAR of 1.5 — much more livable.

Professional shops aim for MAR above 1.0 for the strategies they run unhedged. Retail traders often deploy strategies with MAR below 0.4 because the absolute return looks tempting, and then panic-pull at the first drawdown. The trick is to look at MAR before being seduced by CAGR.

Step 4: the Ulcer Index

Max drawdown is one number; the ulcer index summarises the entire drawdown experience. It's the root-mean-square of all the drawdowns over time, weighted by depth. A strategy with one 30% drawdown that immediately recovers can have the same max drawdown as one with five 30% drawdowns over the course of a year — but the second is psychologically far worse and the ulcer index reflects that. Low ulcer = pleasant to hold; high ulcer = expect to pause the bot mid-pain and never re-enable.

Step 5: break down by regime

Don't read the curve as one thing. Slice it: how did the strategy perform in 2020 (covid crash and bounce)? In 2021 (parabolic)? In 2022 (drawdown)? In 2023-24 (slow grind)? In 2025 (depends on your symbol)? A strategy whose entire P&L was earned in one regime is a strategy you don't actually know.

Regime          CAGR   Max DD   Trades   Win rate
Bull (2020-21)  +84%    -18%      89        58%
Top (2021-22)   -12%    -27%      53        41%
Bear (2022)     +14%    -19%      72        47%
Range (2023-4)  +31%    -11%     142        55%
Recent (2025)   +22%     -8%      97        53%

-> Profitable in 4/5 regimes, only loser is the topping phase.
-> Look at trade count in each regime — small samples lie.

Step 6: the trade distribution

Plot the histogram of per-trade returns. Three healthy shapes are common: a tight bell around a positive mean (steady-grinder, e.g. mean reversion); a fat right tail with many small losers and a few large winners (trend-follower); a bimodal distribution with stop-out losses and TP-fill wins (R-multiple structured exit). Unhealthy shapes: a single enormous winner doing all the work, or a left tail (rare large losers — the strategy has unbounded downside).

Step 7: the stitched curve, not the cherry-picked one

If the curve you're reading came from a single in-sample backtest, you're reading the upper bound of what the strategy can do. The honest curve is the stitched out-of-sample chain from walk-forward. Insist on seeing it before you trust any equity curve.

Next steps

Every backtest run on this platform — see the backtesting docs — emits the curve, the drawdown series, and per-trade statistics. The walk-forward docs add the stitched out-of-sample chain. Reading both, regime-sliced, in 60 seconds is the difference between deploying a real strategy and shipping a curve fit.