-60f from published run after pushing the 2nd boulder onto the switch. Nothing else changed. May not be optimal. Initially this trick was found when the first published run got submitted.