I vote for ignoring the percentages altogether. The recommended macro percentages are originally, I believe, derived from a 2500-calorie daily intake. The quantities work out to be reasonable if figured on that basis, at any rate.
(1) It is pretty clear that people have a definite carbohydrate threshold, above which their insulin level is high enough to inhibit the liver from producing ketone bodies, and to inhibit the metabolism of fatty acids by the muscles. That threshold appears to be different for different people, and only experimentation can reveal what your personal threshold is. But the point is that the threshold is not a percentage of caloric intake but a specific quantity of carbohydrate. So if your threshold is 42.38 grams/day, those 168-odd calories will be a greater or lesser percentage of your caloric intake, as the latter rises and falls.
On these forums, we recommend a carbohydrate limit of 20 g/day, as being low enough to get everyone but the most metabolically deranged into ketosis. (I read somewhere that Richard and Carl really wanted to make it 0g, but were afraid of frightening everyone off, lol!) Dr. Phinney recommends staying below 50 g, and he used to talk about staying under 100-125 g in his older lectures.
(2) Protein recommendations are tied to lean body mass, and are intended to be a guide to avoiding muscle loss. Again, the quantity will vary as a percentage of caloric intake as total caloric intake varies. Richard and Carl recommend 1.0-1.5 g of protein / kg of lean body mass / day, but I’ve seen recommendations as low as 0.6 g (Dr. Ron Rosedale) and as high as 2.0 g (Dr. Ted Naiman, Prof. Benjamin Bikman).
(3) The remaining recommendation is to eat “fat to satiety,” because fat is a low-insulinogenic substitute for the carbohydrate no longer being consumed. Note that it is fat to satiety, not “as much fat as one can cram down one’s gullet.” Or even “fat to some arbitrary percentage of calories consumed.” The reasoning is that once one’s insulin has receded to a low enough level, the satiety hormones leptin and peptide YY can start registering in the hypothalamus again, thus making appetite a reliable indicator of caloric need. The human body is pretty adept at matching caloric intake to caloric output by making adjustments on both sides of the equation. Ate a bit too much? The body can step up the metabolism. Worked a little too hard? The body can increase hunger until we eat more. In the absence of carbohydrate, there is no need to calculate how many calories we need to eat. or what percentage each macronutrient should be—let the body do the calculations and tell us. It’s much easier to relax, and not try to out-think 2,000,000 years of evolution.