Academy
AdvancedRisk 8 min

Regime dependence: when context breaks a strategy

A strategy that prints money in a bull market can quietly bleed out in a flat one — using the exact same rules. The market didn’t break the strategy. The regime changed underneath it.

A “regime” is the prevailing character of the market over a stretch of time: a strong uptrend, a brutal downtrend, a calm sideways drift, a high-volatility panic. Most strategies don’t work equally well across all of them. When a strategy’s performance depends heavily on which regime it’s in, we call it regime-dependent.

Why this is dangerous

A strategy tested mostly during a bull market can look fantastic — but you’ve only learned that it works in bull markets. Deploy it into a different regime and it can fail completely, even though nothing about the rules changed. The strategy was never as good as it looked; you just hadn’t seen it in the wrong weather.

Why we care at Apex Quant

When we test a strategy, we look at how its results vary across market regimes, not just the overall average. A strong average can hide the fact that nearly all the gains came from one favorable period — a fragility that only shows up when you break the results apart.

Average performance hides the truth

A single headline number — “this strategy returns X% per year” — averages across all regimes and conceals the dependence. The honest question isn’t “what’s the average?” but “where did the returns actually come from, and what happens when that condition disappears?”

Regime dependence is why a strategy can be both genuinely profitable in the past and genuinely dangerous in the future. It worked — but only under conditions that won’t always hold.

Key takeaways
  • 1A regime is the prevailing market character: trending, flat, volatile, calm.
  • 2Strategies tested in one regime can fail badly in another with the same rules.
  • 3A strong average return can hide that the gains came from a single favorable period.
Where you'll see this

regime dependencepart of how we stress-test every strategy.