We often hear, and even I've said, do not count calories on a low-carb diet. The advice, on its face, is sound because when someone starts a low-carb diet, overweight or obese, counting only grams of carbohydrate, they reduce their calorie intake even when eating ad libidum; so having to count calories over-complicates a dietary approach that already lowers calorie intake anyway.
But saying "don't count calories" does not mean that calories do not matter. I wrote about how it is a myth that calories don't count back in 2005 (here).
They do, they always have and always will - but they're not always perfectly matched to calories-in/calories-out (CICO) since nutrient-density of the food you eat also matters, along with the fact that the body compensates daily with homeostsis and how much we move, so it's not always burning exactly the same calories each day. But over a long-term average, calculated active metabolic rate (AMR) is a fairly accurate target for calories needed on average each day to maintain weight when you're meeting requirements for essential nutrients.
So, should you count calories?
Based on years of experience, my suggestion remains - when you start a low-carb diet, do not count calories.
If you stall, look at calories - you may be eating too many, or (gasp) not enough.
If you gain, look at calories - I can promise you that if you're keeping your carbohydrate restricted, gaining almost always means you're eating surplus calories from fat and protein in excess of your active metabolic rate (AMR); small fluctuations are normal, but gain on gain on gain while restricting carbohydrate is excess calories.
It is a MYTH that you cannot gain weight if you just keep your carbohydrate restricted, the fact is calories do matter. In addition,hormones, glycogen stores, and metabolic rate matter too, but if you're gaining weight while keeping carbohydrate restricted, you're not gaining in a calorie deficit, you're gaining in a calorie surplus.