Filtering a signal by a state variable from the same bars
Regime conditioning works when the state is exogenous and lagged; measured on the same bars it becomes arithmetic.
- Verdict
- Settled by the data
- Decided by
- Circular conditioning
- The number
- Effect survives on shuffled data
The hypothesis, as pre-registered
Conditioning a strategy on a market state — volatility, trendiness, participation — isolates the regime where it works.
What the test found
When the state variable is computed from the same bars the strategy trades, the bucketing is circular. High-volatility buckets contain large moves by construction, so any strategy whose payoff scales with range will look better there whether or not it has an edge.
The tell is that the effect survives on shuffled data, where no real regime structure exists.
Regime conditioning is a genuinely useful technique — but the state has to be exogenous and lagged: computed from prior sessions, or from a different instrument, and known before the trade is taken. Done that way it holds up. Done the easy way it is arithmetic.
A related trap: a 'high volatility owns the drawdown' headline is usually a unit conversion, not a finding. Drawdown scales with volatility mechanically.
What would make this work
The state variable would have to be computed from information available strictly before the signal bar, and the conditional result would have to disappear on shuffled data.
Every insight ends with this line: the exact condition that would turn the finding into a tradeable edge. That is what makes each entry a next experiment rather than a full stop.
Screened September 1, 2026
Testing something similar?
The screener puts your version of this idea through the same gate stack, on your market, and answers in seconds. Your first verdict is free.