The performance failure is the one to watch, because it feels like following the instructions correctly.
Intent has a smell, as Voss puts it. A label delivered as a checkbox is received as a checkbox, and now they know a technique is being run on them. The tools work only as extensions of genuine curiosity, the one input you cannot fake on a deadline.
The other thing to respect is the ceiling. Voss puts hostage negotiation at 93 percent and is blunt that the last seven are unreachable however long you spend. That is a claim about process, not outcomes, and it sits next to Annie Duke's thinking in bets: losing the hand with the best available process is not evidence the process was wrong.
Even inside Voss's own firm the sequencing is contested. In this session a Black Swan instructor says nine times out of ten he would reach for a label where Voss would reach for a calibrated question, and that both are right. The order is not fixed. The confirmation you drive toward is.