Recovery Factor: How Much Profit a Strategy Earns per Unit of Its Worst Drawdown

The recovery factor is the ratio of a strategy's net result to the depth of its maximum drawdown over the same stretch of history. It answers a single question: did the strategy pay for its worst slump, and by what margin? If the profit is merely comparable to the drawdown that had to be endured, the strategy was operating on a knife-edge: at one point the capital was knocked back by almost the entire amount earned. If the profit covers the worst slump several times over, the drawdown was an episode rather than the substance of the equity curve.

What exactly goes in the numerator and the denominator

The numerator is the cumulative result over the test period or over the history of live trading: net profit after costs, not the "gross" sum of winning trades. The denominator is the maximum drawdown, that is, the largest decline in capital from a peak to the subsequent trough. Both values must be taken from the same period; otherwise the metric loses its meaning.

This leads to a property that is often overlooked: the recovery factor is dimensionless. It does not tell you how quickly the capital returned to its peak; it tells you how many times what was earned covers the worst decline. The speed of the rebound is measured by a different metric, the recovery period. Two strategies can have similar recovery factors and yet differ radically in how long the capital spent below its previous high. For a trader who lives off that capital, the difference is fundamental.

Why the denominator is the most fragile part

The maximum drawdown is a figure built on a single episode in the history. The entire reliability of the recovery factor rests on how representative that episode is. As a result, the metric has predictable weaknesses.

The longer the test, the greater the chance that a genuinely bad market phase falls within the sample, and the lower the recovery factor will be, all else being equal. Values calculated over periods of different length cannot be compared: a short track record almost always looks better, not because of the quality of the strategy, but because its worst day has not yet arrived.

The numerator, by contrast, grows over time almost mechanically: the longer a profitable system runs, the larger the cumulative total, while the maximum drawdown is updated only rarely. Over a long history, the recovery factor therefore tends to rise on its own; this is an artefact of the calculation, not an improvement in the strategy.

Finally, the metric knows nothing about the shape of the drawdown. A sharp plunge followed by a rebound and a slow slide lasting many months produce the same denominator if their depth is equal, even though they are experienced in completely different ways. This distinction is captured by recovery asymmetry.

How to use the metric in practice

Step 1: fix the calculation window and do not change it when comparing systems. Step 2: calculate the recovery factor separately for sub-periods. If it stays within the same order of magnitude across different market phases, the metric carries real information; if it holds up only over the history as a whole, it has been assembled from one lucky year. Step 3: look at the recovery factor together with the recovery period and the trading frequency. Infrequent trading yields few independent observations, and both figures become close to random.

There is no universal threshold for this metric, and I will not put a number on it: the acceptable level depends on the horizon, the leverage and what the portfolio consists of. The only honest benchmark is a comparison with the alternative you are giving up. On the Russian market, this is usually the yield to maturity on government bonds and the result of simply holding equities with payouts reinvested according to the dividend calendar. If the strategy does not beat such an alternative with a margin to spare for the drawdown, the recovery factor merely confirms that the risk was taken for nothing.

A separate caveat applies to portfolio strategies: the recovery factor measures the final equity curve and says nothing about which exposure produced it. A drawdown caused entirely by a broad market move and a drawdown caused by concentration in a narrow set of securities produce the same denominator. The sources need to be analysed separately, through factor risk and the other definitions in the glossary.

In short, the recovery factor is a good filter and a poor argument. It quickly weeds out systems whose profit does not pay for the decline they went through, but a high value on its own proves nothing until you know the length of the sample, how the parameters were selected and the shape of the worst drawdown.