Drawdown and profit factor: a strategy's risk

Two strategies can finish a season at the same units and be very different to follow. One climbs steadily. The other falls a long way, recovers and finishes level with the first. Most people stop following the second one at the bottom, so they never see the recovery. Risk numbers describe the path, not just the destination.

Maximum drawdown is the largest fall in units from a high point of the running total to a later low point. If a strategy reaches +20 units, falls to +4 and later climbs to +30, its drawdown on that stretch is 16 units, even though it finished higher.

Compare it with your own stake. At 1% of your bankroll per unit, a 16-unit drawdown is a 16% fall in your bankroll. Ask whether you would have kept betting through it. If not, the strategy's final number is not one you would have collected.

Drawdown grows with time. A longer record has had more chances to hit a bad stretch, so a short record with a small drawdown has simply not been tested as hard. Compare drawdowns over similar numbers of picks.

Profit factor is the units won on winning picks divided by the units lost on losing ones. Above 1, the strategy won more than it lost. At 1.2, it won 1.2 units for every unit it gave back.

It hides the order of results, so read it next to drawdown. A profit factor of 1.2 with a deep drawdown and one with a shallow drawdown are different experiences. With no losses yet the ratio cannot be calculated, so we show it as not yet measurable rather than as infinite.

The Strategies board shows maximum drawdown and profit factor for every strategy, beside its units, ROI and CLV. Each strategy page adds an equity curve, the running total of units pick by pick, so you can see where the drawdown happened and how long it lasted.

These numbers describe the past at the published prices. Your own drawdown would differ if you bet different prices, skipped picks or staked unevenly.

Take two hypothetical strategies, each with 300 graded picks and +15 units. Strategy A has a maximum drawdown of 9 units and a profit factor of 1.08. Strategy B has a maximum drawdown of 31 units and a profit factor of 1.07.

On units alone they are the same. On risk they are not. Following B at 1% of bankroll per unit meant sitting through a 31% fall at some point. Following A meant a fall of 9%. Most bettors would have stopped following B at its worst, taking a loss on a strategy that later recovered. A is the one you could actually have followed.

Strategy pages also show the average win and average loss in units. Their ratio explains a profit factor. A strategy that wins small amounts often and loses large amounts rarely, such as one that backs heavy favourites, can have a good win rate and a thin profit factor. A strategy on underdogs wins less often and more per win.

Neither shape is better. They feel different. Favourite-heavy records tend to have shorter, sharper drawdowns; underdog-heavy ones tend to have longer, slower ones. Choose the shape you can live with, and size to it.

A rough rule: size so that a drawdown twice as deep as the worst one on record would still leave you betting. Records have not seen their worst stretch yet, and a strategy's next drawdown can be deeper than any so far. Flat stakes of 1% to 2% of bankroll per unit are the usual way to stay inside that.