Principlev1
Run environmental experiments using crossover designs that
Run environmental experiments using crossover designs that alternate between conditions rather than testing sequentially, to control for temporal confounds like day-of-week effects or changing life circumstances.
Why This Is a Principle
This derives from tacit knowledge degrading without systematic improvement methods (Knowledge that exists only in tacit form degrades without), world model accuracy bounding performance (The performance of an agent is bounded by the accuracy of), and cognitive capacity varying in predictable cycles (Ultradian and Circadian Cognitive Rhythms). Crossover designs help separate experimental effects from temporal variation.