Why it’s built this way.

None of this is here to look rigorous. It is here because the dose I sell is smaller than the industry sells, and you should be able to check that rather than take my word for it.

Risk. Heterogeneity was high and the certainty was graded very low. Above about 140 minutes a week the picture is unclear rather than better.

Muscle. Twice a week isn’t a compromise between once and three times; for this outcome it’s the answer.

Strength. The strength arm is an open ended floor of three or more, not exactly three, and it’s private training rather than the program.

These are averages across thousands of people, not predictions about you.

It’s observational. Fitter people differ from less fit people in many ways, and the study cannot tell you that adding conditioning causes the gap.

Which is why conditioning is in every block rather than sold as an upgrade.

Breathing. A 2023 review found a small effect on stress and the authors warned about study quality. I teach it because it’s cheap, portable and it works on the night you can’t sleep, not because it’s a cure.

Mobility. Stretching improves the range you can hold. It does not prevent injury, and the evidence for that is weak.

Conditioning. Bring whatever you own. I use it to notice what’s changing week to week, not to score you.

Monitoring works, and works harder when you tell someone. 138 experiments, 19,951 people, 2016. The effect on reaching the goal was d=0.40. It was larger when progress was reported to another person, and larger again when it was physically written down.

Supervision alone changed the result. Eighteen trained men, twelve weeks, identical program. Sessions, sets and repetitions didn’t differ between the groups. Load did. Squat one rep max rose 33% supervised against 25% unsupervised, bench 22% against 15%. They didn’t do more work. They did heavier work, because they weren’t the ones choosing the load.

It only works if the person is worth answering to. A model of human support in health programs, 2011. Accountability functions through legitimacy, which has two halves: real expertise, and acting in the client’s interest. Where a coach seems illegitimate or controlling, accountability has been found to reduce adherence instead of raising it.

What none of them show. d=0.40 is a reliable nudge, not a transformation, and shouldn’t be sold as one. The supervision trial is eighteen trained men aged 18 to 35, supervised in person rather than over video, and says nothing about women, older adults or a first month of training; it’s the most useful of the three and the weakest. The third is a model, which is a well argued account of a mechanism rather than a measurement of one. And none of the three was conducted on you.