You are assuming our athlete has a lifetime to grow stronger.
1. If our trainee is 16 and wants to not get smashed on the football field, then he better get strong, quick and efficiency is not irrelevant.
2. If our trainee is 32 and has a family and a career and a mortgage and two car payments and can maybe sneak away to the gym may 3 hours per week, if lucky, then he better get strong, quick and efficiency is not irrelevant.
3. If our trainee is 53 going on 73, overweight, diabetic and heading down the road to polypharmacy, she better get strong, quick and efficiency is not irrelevant.
4. If our trainee is actually 73 and sarcopenia and osteopenia are crashing down around her on all sides then she better get strong, quick, and efficiency is not irrelevant.
I cannot think of a meaningful situation where we ignore time's scarcity. To do otherwise, but then claim to be purveyors of fitness, may make sense only on a P&L Table at Planet Fitness.
The 401k analogy is useful, but only to a point. Whatever the shape of the decay, a higher starting point is ALWAYS better. And, it has been my experience, that once a person has experienced some measurable increase in strength, and he sees that he can make that change, it is easier to fight that decay going forward. It is certainly easier than trying to convince a 70 year old to squat for the first time.
Optimality is not in the eye of the beholder, that's the point of my entire article. It is objectively verifiable.
Strength is the most useful adaptation as it influences other expressions of fitness. Strength can be measured. Since it can be measured and is most useful, progression should be maximized. Progression is maximized, I think, when a program takes advantage of the Novice Effect allowing for just enough recovery, but not too much down time. We test if the results are numerically superior to all other known programs for some measure of time that allows for meaningful adaptation to occur. If we can't find a better way to get people stronger as efficiently, then that's optimal (for now) and that's what any trainer should be coaching.
Now if there is something in the program's nature that causes it to attract people with a low rate of compliance, i.e. a high rate of idiosyncratic volatility, then certainly questions should be raised. But inserting compliance as measure of optimality is not a good thing, I don't think, since an optimal program, by the definition in the article, is going to be hard. We cannot discount a program's optimality simply because it's hard and some people will decide not to finish it.
Either way, this model has a mechanism to test rates of non-compliance (when collecting the data, the coach/client designates if the failure to achieve progress is a result of not eating, not training, etc). Hell, once the data are collected, a survival model could test if certain programs have a higher "hazard" rate (dropping out of the program = non-survival).
Here's the best way to convince them: get their squat to increase in three weeks by 30-45 lbs. Training Efficiency and Low Programmatic Volatility are really the same ingredient in the cake. And the cake is not a lie.