Your pick — B's confident critic + C's three reads. The grade leads; the three reads give the at-a-glance split; the real quotes are the “why.” No internal terms anywhere. On the right: what happens when a set barely has reviews — the answer to the whole “trust” question.
“The most beautiful thing LEGO has ever made.” The layered build and the density of detail win almost everyone over.
At ~$500 it's a real commitment — some call that steep for a set you display rather than play with.
Only 2 reviews so far — “gorgeous, but I want more time with it.” This grade will sharpen as more land.
This right-hand card is the answer to what you were stuck on. We're not grading the AI in a lab, and we're not stamping “unproven” on everything. We just show what real people said, count them out loud, and blank out what we honestly can't call yet — with a plain “Early read” that promises to firm up. Trust isn't a badge or a project. It's the receipts, in plain language.
Does this feel like the Drop Score? If yes, I'll write it up as the design spec — the card anatomy, the three reads, and the plain-language “early read” rule for thin data. If something's off — the order, the two-quote “why,” the read labels — tell me and I'll adjust the mock first.