For a condition, interventions ranked by strength of evidence — not by a confidence score. We tested our old confidence number and it was no better than a coin flip, so we don't show one. Human-trial evidence is ranked above hypotheses, and contested findings are marked and demoted.