I don't think counting calories is useless. But it's not as easy or simple as put out. Bodybuilders seem to be on to this--insistence on CICO being "it" there comes with all sorts of warnings about difficulties like inaccuracies on labels, the tendency for NEAT (I think that's non-exercise activity thermogenesis) to go down when intentional exercise goes up, etc. I don't think you can with exquisite accuracy control calories in versus calories out.

What I often see--first, determine maintenance calories by trial and error. If you're like most people, a creature of habit--you'll probably be eating a limited number of foods that you favour. Say you're eating 2500 calories--that might actually be 2300, or it might be 2700. Cut by 20 percent of calories, while eating more or less the same foods--you might be cutting 460 calories, you might be cutting 540. But as long as it's pretty much scaling down what you'd been eating in maintenance--it's probably going to put you in a deficit. Or what would be a deficit, if you don't compensate in activity or something.

Calorie restriction has never failed for me with weight loss--when I manage to restrict, I lose weight. I make no claims for anybody but me from this. Where it fails for me--when this is my main strategy, later on I end up bingeing. This actually happens not while I'm actively losing weight, but when I get to goal and am thinking about maintenance. I think this actually fits into the insulin hypothesis. If I'm eating 1800 to lose weight, but need to eat 2400 for maintenance--I've just added 600 calories of insulin secretion stimulating food to my diet. Going to more ketogenic food--more fat, even lower carbs, restrained protein--allows me to eat at maintenance without binges. Basically, switching to the least insulinogenic maintenance diet available.