We have more rate data in this issue than any before it, and the main thing it proved is how little of it can be compared yet. Stanford posts 4.49% on 60-month used auto. We post 5.49%. That looks like a hundred basis points, and it isn't. The tier thresholds and fee structures underneath those two numbers don't match, so the subtraction doesn't work. HELOC margins over Prime gave us one clean comparison, because the index is public and several peers disclose their spread. The rest is observation without position. This issue maps every cell: what we collected, what's blocked, and what evidence resolves it.