Any ideas what's going on here? I was under the impression Blend used a weighted sum, as it approached the end of a dataset.
Reversing the order, just ignores the second input since the first is much larger, and thus should be weighted heavily.
