Recovery factor

In plain English

How many times over did the strategy earn back the worst hole it dug? It puts the reward and the pain on the same scale, so a large return bought with a large decline stops looking like a large return.

Why it matters

Net profit and maximum drawdown are usually quoted side by side and read separately, which lets a seller lead with the flattering one. Recovery factor forces them into a single ratio, so a strategy cannot buy its return with a decline the account could not have survived.

• ✓It is the sanity check on a headline return. A 200% gain through a 50% drawdown scores 4.0; a 40% gain through a 5% drawdown scores 8.0, and the second is the more deployable product despite the smaller number on the poster.
• ✓It is scale-free in the same way profit factor is, so it survives comparison across account sizes and lot settings — provided both terms are measured on the same basis.
• ✓It exposes leverage dressed up as skill. Doubling the lot size roughly doubles both the profit and the drawdown, so recovery factor barely moves — which is exactly the point. A metric that improved would be measuring the leverage, not the strategy.
• ✓It reorders a catalogue. Rank the fourteen EAs published here by net profit, then by recovery factor: the two orders agree only loosely, at a rank correlation of 0.73. The smallest net profit of the fourteen, 386.05 USD, lands fifth on recovery factor — the return alone would have put it last. The second-largest, 10,280.42 USD, lands third.

How it is calculated

Both terms must be in the same units — either both in account currency or both as percentages — and both must come from the same test run. Mixing a percentage return with a currency drawdown produces a number that means nothing.

What counts as high or low

• under 2 The return did not justify the hole
• 2–3 Thin — workable only if the drawdown itself is shallow
• 3–5 Normal for a tested retail EA over several years
• over 5 Strong — check the sample is long enough to have found a real trough

Read it together with

No single number describes a strategy. These metrics change how this one should be read.

Example

Two strategies both finish the test window in profit. Reading the returns alone ranks them one way; dividing each return by the decline it required ranks them the other way.

Strategy B needs roughly four times the capital to be run at the same risk of ruin, which makes it a different product rather than an equal one.

How to interpret it

Recovery factor is a ratio of two point-in-time figures, so it inherits the weaknesses of both. Establish the window and the measurement basis before reading it.

• The numerator grows with time and the denominator does not have to, so a longer test tends to raise recovery factor for free. Compare figures over comparable spans, or use Calmar instead.
• It is path-blind. A strategy that earned everything in year one and then went flat scores the same as one that earned steadily, provided the drawdown matched.
• Read it beside worst losing streak. Recovery factor says the hole was repaid; it does not say how long the account sat in it.
• Watch the denominator stop moving. In all fourteen runs published here the deepest trough sits inside the backtest window, so trades added afterwards lift only the numerator. One listing went from 3.37 to 5.73 on 15 extra trades. Two went the other way, because their added trades lost money.

Deriving the ratio for the fourteen runs published here gives 1.67 to 15.66 on balance drawdown. One sits under the 2.0 line this page calls unjustified, and six sit above 5.0, where the table above says to go and check the sample length. Swapping the denominator to equity drawdown lowers the ratio in thirteen of the fourteen, and the widest fall — 15.66 to 8.86 — belongs to the highest scorer. The exception is one of the two runs read on MT5's Relative basis: there balance and equity pick different moments, so their currency amounts are not comparable at all. No EA page here prints a recovery factor — both terms come from the same run and the division is left to the reader.

Common mistakes

Frequently asked questions

What is a good recovery factor for an EA?

Roughly 3 to 5 over several years of realistic testing is a normal, believable range for a retail EA. Below 2 the return did not justify the decline it required; above 5 is strong but worth checking against the length of the sample, since a short test can simply have missed the real trough.

How is recovery factor different from profit factor?

Profit factor divides gross profit by gross loss and is blind to the path — it does not know whether the losses arrived one at a time or all at once. Recovery factor divides net profit by the deepest peak-to-trough decline, so it is specifically about the worst stretch rather than the total.

Where do I find recovery factor on an EA listing?

Usually nowhere. No EA page on this site prints it, and most vendor pages do not either. Both terms it needs are published here from the same run and beside the data window, so the division takes a moment. Doing it for the fourteen runs on this site gives 1.67 at the low end and 15.66 at the high end — a spread far wider than the net profits alone suggest.

Should I use the Calmar ratio instead?

For comparing strategies tested over different lengths, yes. Calmar divides the annualised return (CAGR) by maximum drawdown, so it does not reward a strategy simply for having been tested for longer. Recovery factor remains the quicker read when both figures come from the same window.

Related terms

• Maximum drawdown The largest peak-to-trough decline in account equity over a measurement period, as a percentage of the previous peak.
• Profit factor Total gross profit divided by total gross loss over a given period. A profit factor above 1.5 indicates a profitable strategy; above 2.0 is strong.
• Sharpe ratio A risk-adjusted return metric: excess return divided by the standard deviation of returns. It rewards a smooth equity curve, which is why it can flatter a martingale EA right up until it blows up.
• Compound annual growth rate The annualised rate of return over a multi-year period that accounts for compounding, normalising returns across different backtest lengths so two strategies tested over different spans can be compared.

Related resources

• MT5 Parameter Optimisation: 116 Combinations, 252 Runs, 0 Settings We ShippedGuides
• Value at riskGlossary
• LeverageGlossary