Is obesity caused by too much insulin, or just by too many calories?

The three are not testing the same proposition. Hall's trials restrict carbohydrate and measure ad libitum intake and energy expenditure in healthy or overweight volunteers over two to four weeks. Fung's proposition is about insulin rather than carbohydrate specifically, his primary lever is eating frequency rather than macronutrient composition, and his population is people with established type 2 diabetes on insulin, which is the group whose insulin is furthest from normal and the group Hall's trials did not study. Norton is answering a body-composition question in people trying to change it.

The one place all three genuinely meet is the claim that lowering insulin by itself produces fat loss independent of intake, and on that specific claim Hall ran the experiment and Fung disputes the interpretation rather than the data.

All three accept that reducing refined carbohydrate and sugar is reasonable advice. All three accept that adherence dominates outcomes over the long run: Fung's own "the best fast is the one that fits into your lifestyle" is Norton's argument in his own words.

Fung concedes Hall's 2016 data explicitly ("the data are excellent") and declines to attack its sample size or its calorie-controlled design. Nobody in this set recommends grazing across the whole day.

An inpatient trial that manipulates eating FREQUENCY rather than macronutrient composition, with calories equated, insulin measured, and metabolic endpoints in people with established type 2 diabetes. That experiment tests Fung's actual claim, and neither Hall nor Fung has run it: Hall's trials vary what is eaten, Fung's evidence is uncontrolled clinical observation.

Short of that, a randomised comparison of his published 24-hour-fast protocol against equivalent continuous caloric restriction, with medication changes as an endpoint, would settle whether the fasting adds anything beyond the intake reduction.