Just musing and looking for your thoughts on the questions.
I'm not sure exactly how to do it but the best approach - which I assume is what the current method is attempting - is one that allows the the instructor to determine where/when/why/how the STUDENT will FAIL. It's cliche, but we learn by our mistakes, not our successes. In order to understand where/when/why/how a student will fail, the instructor needs to seem him/her fail. If they don't, you can't be sure whether they "did it right" by dumb luck.
Lynn will relate: it's why clinical trials are done the way they are. To determine whether a drug/procedure/intervention is effective you need a sufficiently large number of observations - and statistical power - to be certain that the results of the trial are not simply to due to chance (p<.005).
Training is the same way. You do a lights out, no-mask exit once; big deal. Twice? Lucky. Three times? Still lucky. Four times? Now I don't know if you've learned it, or are simply aping it. So I need to introduce another variable, because I need to see you fail, in order to know if you've learned. And unless I want to spend weeks, maybe months, waiting for you to fail on your own, we're going to need some artificial
charlie foxtrot's and/or some long days.
Back to the clinical trial example, for Lynn anyway, think about the HOPE study (or any other landmark CV trial) where the study population was +55yrs, documented CV risk, plus one or more
additional risk factors such as smoking, overweight, diabetes, previous MI, family history, etc, etc. Why did we do it that way? Because we needed enough primary FAILURES (MI/Stroke/CV deaths) and secondary FAILURES (hospitalizations/revascularizations/etc) over a reasonable period of time in order to determine efficacy. If the study was done in otherwise healthy individuals we'd still be waiting for enough bad outcomes to detect the study effect.
My cavern instructor was pretty clear about his intentions: "You will fail at some point, I have no doubt. I'm not looking to see if the cave, or your gear, or you - or I - can cause you to fail. I already know they can. I need to understand
where, when, why, and how you will fail. Because that's where I can intervene and teach you."
It's hard to detect the "efficacy" of the training, unless there's enough observations of failure to ensure that the findings are not due to chance.