It’s a natural shortcut to estimate a portfolio’s likely drawdown by looking at its member strategies’ individual drawdowns — averaging them, or simply eyeballing the worst one. It’s also a shortcut that can be badly wrong in either direction, because a portfolio’s real drawdown comes from how its strategies interact over time, not from any simple summary of their separate historical worst cases.
Strategies interact — that’s the whole point of a portfolio
The entire premise of combining strategies is that they don’t all do the same thing at the same time. That interaction is exactly what determines a portfolio’s combined drawdown, and it can move it in either direction relative to a naive average of the members’ individual drawdowns.
Simultaneous drawdowns: the direction everyone worries about
If several member strategies tend to draw down at the same time — because they share exposure to a similar condition, even if their individual backtests never showed it clearly — the portfolio’s combined drawdown can be worse than any single member’s own historical worst case. This is the scenario diversification is meant to guard against, and it’s exactly why checking correlation and behaviour, not just individual quality, matters before combining strategies — see Correlation in Algorithmic Portfolios.
Offsetting periods: the direction people forget about
The reverse is also genuinely possible, and it’s the outcome good diversification is aiming for: if one member strategy is drawing down while others are performing normally or well, the portfolio’s combined equity curve can be considerably smoother than any individual member’s own curve. A portfolio’s combined drawdown can, in this case, be meaningfully better than a simple average of its members’ individual drawdowns would suggest.
This is why averaging member drawdowns is not a reliable estimate of portfolio drawdown in either direction — it discards the timing information that actually determines the outcome.
The combined equity curve is the real evidence
The only way to know a portfolio’s genuine historical drawdown is to build its actual combined equity curve — every trade from every member strategy, placed in chronological order, treated as one continuous account — and measure the drawdown from that combined curve directly. This is a materially different calculation from looking at each strategy’s separate equity curve and its separate drawdown figure.
This combined-trade approach is consistent with the broader idea explored in What Is Portfolio Trading: a portfolio’s genuine performance, in every dimension, comes from its own combined history, not from summarising its parts.
What this means for evaluating candidates
When comparing two candidate portfolios built from strategies with similar average individual drawdowns, don’t assume the portfolios themselves will behave similarly. Their combined drawdown depends on how their respective member strategies’ trades actually line up in time — a detail that only shows up once the combined curve is calculated, not before.
A boundary worth being honest about
None of this reveals exactly how any specific portfolio’s combination is calculated or searched for at scale — that’s a separate, more technical question about implementation. What matters for a trader building or evaluating portfolios is simply the principle: combined portfolio drawdown is an emergent property of the whole combination, and it needs to be measured directly from the combined trade stream, not estimated from the individual pieces.
Where this fits in the SIPS workflow
Every combined-portfolio metric inside SIPS — including drawdown — is calculated from the portfolio’s own combined chronological trade stream, never derived by averaging member strategy figures. Portfolio Analysis presents this combined equity curve directly, alongside each member strategy’s own individual contribution, so both views are available side by side.
The practical takeaway
Don’t estimate a portfolio’s drawdown from its members’ separate drawdown figures — the interaction between strategies can make the combined result meaningfully better or meaningfully worse than that average suggests. The only reliable evidence is the portfolio’s own combined equity curve, measured directly.