earth-a-verified-answer-key-does-not-change-the-question

# Earth: a verified answer key does not change what the study is Community assessments in this field ship as an archive with two halves: the raw submissions that went in, and the combined product that came out. Within the first hour you will notice the second half, recompute the source's headline numbers from it, and find that all of them land. What happens next decides the run. ## The gift, and the trap The gift is real. A verified reference means every descriptive statement you go on to make can be checked before you publish it, so you can afford to state *more* of them than a run working without a reference, not fewer, and each one arrives with a residual attached. An archive that contains its own answer is the cheapest quality control you will ever be handed. The trap is that a confirmed answer feels like a closed question. The run reasons that quoting the shipped result would answer nothing, goes looking for a question still open, and finds one immediately: the source's method has choices it never tested — a weighting, an exclusion, an independence assumption, a coverage gap. Those are genuine, they are cheap to attack, and each one is refutable, which is exactly what a hypothesis stage is hunting for. From that moment the study is indexed by the critique. Figure slots go to the arms that adjudicate it, results subsections are named after its verdicts, and the numbers verified in hour one sit in a validation file with nowhere to be written down, because a descriptive...