An interesting topic!
However I don't believe that the figures we see on packaging (or even those provided by reliable/professional resources) can ever be 100% accurate to the nearest carb or fraction thereof. They can only be taken as averages of the nutrients in different test batches, and these may well change depending on somtimes non-controllable conditions - e.g. for fruits and veggies: variables that might affect results could be the actual variety, ripeness, growing conditions, climate, etc. Nature isn't exactly a controlled chemistry lab.... so as far as I'm concerned we're essentially provided with and using a best "guesstimate" about the quantities of carbs, protein, fat, etc we eat... And if someone tells me they have eaten precisely 27g carbs today, I tend to think "Oh yeah - in your dreams maybe...!" But it's as close as we can get for valid data to work with... though I always try to allow for a bit of "slop" in my calculations.
Robboty