Drawdown Control at the Portfolio Level

A strategy that loses 10% is down 10%. To get back to breakeven, it needs to gain about 11.1%, not 10% — losses and the gains needed to recover them aren't symmetric, and the gap widens fast: a 20% loss needs a 25% gain to recover, a 50% loss needs a 100% gain. Portfolio-level drawdown control is a set of rules, mechanical rather than discretionary, that cut a portfolio's risk as losses accumulate specifically to avoid sliding into that steep part of the recovery curve — accepting a worse outcome in scenarios that would have recovered on their own, in exchange for a much better outcome in the scenarios that wouldn't have.

The mechanics of a drawdown rule

A typical rule sets thresholds against a portfolio's high-water mark — its highest cumulative value to date — and cuts gross or net exposure by a set amount at each one: full risk up to a 5% drawdown, half of target risk between 5% and 10%, flat (or close to flat) beyond 10%. Some versions scale continuously rather than in discrete steps, reducing exposure smoothly as a function of distance from the high-water mark. Either way the logic is the same: the deeper the hole, the less risk the strategy is allowed to take digging further, on the theory that a strategy already struggling is statistically more likely to keep struggling than a fresh one — whether from a genuine regime change the strategy hasn't adapted to, or from the psychological and career pressure on a team managing a large loss to make it back quickly, which tends to produce worse decisions, not better ones.

Worked example

A pod is running $200m of risk capital targeting 12% annualized volatility, with a drawdown rule: full risk to -5%, half risk from -5% to -10%, flat below -10%. The pod loses 6% over six weeks — worse than a single bad month but not catastrophic — and the rule cuts its target volatility from 12% to 6%, roughly halving its gross exposure. Two scenarios from there: if the loss was noise (the underlying signals were fine, the market just moved against the pod temporarily), the pod recovers more slowly than it would have at full risk, giving up some of the snap-back gain — a real cost. If instead the loss was the first sign of a genuine regime break (the pod's signals stopped working), the halved exposure means the drawdown that follows is roughly half of what it would have been, so a scenario that might have run to -18% at full risk instead runs to roughly -10% to -11% at half risk before the second threshold cuts exposure further. The rule doesn't know in advance which scenario it's in — it pays the cost of scenario one on every drawdown, in exchange for capping scenario two whenever it happens.

What this means in practice

Multi-strategy platforms build drawdown rules into pod risk budgets as a matter of course, with thresholds tight enough (often single-digit percentages) that a struggling pod is cut back or shut down well before it could meaningfully damage the platform (see The Multi-Strategy Platform Model). Standalone funds use looser versions of the same idea, sometimes as an internal risk discipline rather than a hard contractual rule, precisely because a looser rule leaves more room for a genuinely temporary drawdown to recover without being cut off early.

What erodes it

A drawdown rule that's too tight turns ordinary volatility into forced selling at the worst possible time — cutting exposure right as a mean-reverting strategy hits its statistically most attractive entry point, locking in a loss the strategy's own logic says should have been an opportunity. This is the core tension: the tighter the rule, the better it protects against genuine regime breaks, and the more it costs in whipsawed, prematurely-cut trades during ordinary drawdowns that were never going to be catastrophic. Calibrating the threshold is a bet on how the strategy's typical drawdown compares to its rare catastrophic one, and getting that calibration wrong in either direction has a real cost.