Recovery Factor: Return Versus Pain in One Number

Two systems both made $10,000 this year. One never dropped more than $1,000 on the way; the other went $6,000 underwater first. They look identical on a profit statement — and recovery factor is the single number that tells them apart.

The formula: net profit / max drawdown

Recovery factor (RF) is exactly what it sounds like:

RF = net profit / maximum drawdown

Both in the same units. A system that netted $10,000 against a worst peak-to-trough hole of $2,000 has an RF of 5 — it earned back its deepest drawdown five times over. You can compute it on dollars or on percentages; either way it’s a dimensionless ratio, and higher is better.

What makes RF useful is that it forces the cost of the return into view. Net profit alone tells you the system made money. RF tells you how much punishment you absorbed to get it — and on a prop account, the punishment is what breaches you.

What recovery factor tells you Sharpe doesn’t

Sharpe ratio measures return per unit of volatility — the general dispersion of your returns. It penalizes big up swings as much as down ones, and it says nothing directly about the one event that actually ends prop accounts: the deepest hole.

Recovery factor cares about exactly that. It ignores day-to-day chop and focuses on the single worst peak-to-trough decline — the thing that trips a drawdown floor and breaks a trader’s nerve. A system can post a mediocre Sharpe (choppy, noisy) yet a strong RF (never a deep hole), or a beautiful Sharpe wrecked by one catastrophic tail. For a trader whose survival depends on not touching a floor, RF is often the more relevant lens. The two answer different questions; neither replaces the other.

Benchmarks for tradable systems

Treat what follows as rules of thumb from a full track record, not laws — RF is only meaningful over a sample big enough to have suffered a real bad patch.

Below ~2: the system spends heavily relative to what it makes. Thin margin for a prop account, where the drawdown floor is close.
Roughly 3–5 over a full cycle: generally considered solid for a discretionary or systematic strategy you’d fund.
Very high on a short sample: usually luck or an untested tail, not quality. The representative bad streak simply hasn’t happened yet.

The most important word in all of that is sample. A dazzling RF built from a few months of favorable conditions is a mirage, and the next section is why.

Why it’s fragile on short samples

RF’s denominator is a single extreme observation — your one worst drawdown. Single extreme values are the least stable statistic you can build on. Early in a track record you may not have hit your representative bad streak yet, so the denominator is artificially small and RF flatters you.

The consequence is harsh: one new, deeper drawdown can halve the ratio overnight. A system showing RF 8 after a smooth run can drop to RF 4 the first time it has a normal bad month — nothing about the edge changed, you simply finally sampled the tail. So never trust RF computed on a handful of trades, and expect your number to drift down as history accumulates, not up. This is the same logic as “historical max drawdown is a floor, not a ceiling,” applied to the denominator.

Recovery factor across market regimes

A system’s RF measured only in the regime that suits it is worthless as a forecast. Trend systems post gorgeous recovery factors in trending years and collapse in chop; mean-reversion does the opposite. A lifetime RF of 4 built from one great year and one ugly one hides the ugly year’s RF of 0.8 — and the ugly year is the one that governs whether you survive on a funded account.

So compute RF per regime or per year, not just lifetime, and pay attention to the worst regime’s figure. That number is the honest one: it tells you how the system behaves when conditions turn against it, which on a prop account is precisely when a low RF turns a rough patch into a breach.

Pairing it with drawdown duration

RF tells you how deep the hole was. It says nothing about how long you were in it. Two systems with identical recovery factors can differ enormously — one climbs back to a new high in a week, the other stays underwater for months.

For a prop trader, drawdown duration (time to a new equity high) is its own risk:

Time in drawdown is time you might still breach the floor.
It’s time a challenge time-limit keeps ticking.
It’s time your discipline frays and sizing creeps.

Pair RF (depth efficiency) with maximum drawdown duration (time efficiency) for the full picture. On a trailing account especially, a long shallow drawdown can be as dangerous as a short deep one, because the moving floor keeps pressing — confirm the exact mechanics with your firm.

Getting RF right depends on measuring the real equity curve, and that’s where continuous tracking helps: Shibiki builds your curve from auto-journaled fills, so max drawdown, its duration, and RF are computed from actual trades rather than a hopeful backtest. Live edge health flags when the number is drifting, the hard risk limits enforced at the broker cap the drawdown that sits in RF’s denominator, and the prop-firm drawdown calculator and drawdown recovery calculator turn the ratio back into the account terms you actually trade against.