I do like that they started putting confidence intervals on estimates of replicability. That said, "Thirteen methods for evaluating replication success provided estimates ranging from 28.6% to 74.8% (median of 49.3%)"
Wide range y'all. And what's a median of metrics mean?