Why might sensor spacing patterns create different types of extrapolation errors than uniform dimensional scaling?

This is a thoughtful question about measurement geometry, but the provided preprints don't directly address the general principle you're asking about.

However, I can sketch the underlying tension using what's in the archive:

Uniform scaling vs. discrete spacing patterns

When you scale a measurement apparatus uniformly (e.g., double all sensor distances), errors typically scale predictably with your new resolution. But when sensors are *spaced non-uniformly*, you introduce a fundamentally different problem: aliasing and interpolation bias.

The clearest analogy in these papers appears in cosmology. When estimating correlation functions from a discrete, randomly distributed sample of sky observations, the spacing pattern of your sample—not just its density—matters [AG-2025.02-311]. The "sparsity covariance" effect shows that uneven sampling introduces uncertainties that depend on *where* your measurements fall, independent of experimental noise. This is a pure geometry-of-sampling problem.

Similarly, in causal set theory, measuring distances from a discrete substrate [AG-2024.01-482] requires choosing which causal relationships to use as "rulers." A non-uniform spacing pattern in your discrete events creates different projection artifacts than uniform downsampling would.

Why the difference?

Uniform scaling preserves symmetry: errors propagate in predictable directions. Discrete patterns break symmetry: your estimator now depends on *which* discrete points you happened to sample, creating systematic biases that don't vanish smoothly as you refine resolution.

The provided preprints don't develop this general principle for sensor arrays or extrapolation theory specifically, so you may need sources focused on interpolation theory or sampling geometry.