What determines the current verdict for an indicator when multiple sources provide different reliability scores and verdicts?