To play devil's advocate, I wonder what kind of calorie-testing the reviewers used. Did they measure total calorie count (including calories from fiber and sugar alcohols)?
Even some of the "experts" do not realize that humans cannot digest fiber calories AT ALL (as a result manufacturer's are allowed to subtract these calories, but are not required, and with sugar alcohols they're able to subtract the "average" useable calories - though newer research suggests that sugar alcohols can be more caloric for some people than for others).
Diet products are most likely to do the allowed subtracting (because it makes their product look better... and really it's usually the more accurate calorie count... at least if sugar alcohols aren't in the product).
With the sugar alcohols it gets iffier because not only do some sugar alcohols contain more calories than others, some people can digest some of them better than others. There's no hard and fast rule, but the fewer unpleasant symptoms a sugar alcohol causes a person, the more likely the person is to be absorbing more of the calories.
Thus for example, xylitol might be more caloric for me (since I don't seem to have a problem with it, unless I eat ridiculous amounts of it), but sorbitol and mannitol give me MAJOR digestive problems. Even small amounts can make me feel like my lower body is turning itself inside-out.
I found this chart that lists the calorie count per gram for sugar alchols (but these are averages. Some people absorb more calories, and some fewer, and the amount of intestinal distress appears to be inversely correlated with the calorie absorption... so just because xylitol is listed as having 2.5 calories per gram, doesn't mean it has 2.5 calories per gram for everyone. Some people may be able to pull only 2 calories per gram and others might be able to pull 3....)
http://lowcarbdiets.about.com/od/wha...aralcohols.htm
Calorie counting isn't as precise a science as we'd like to believe. There are many factors that affect how well we are able to pull calories from certain foods. Most calorie counting resources (in the USA) act as if our bodies are furnaces, and will use every bit of burnable energy in the food, but this isn't true.
Hay and wood are quite high in calorie, for animals that can digest cellulose (plant fiber). It wasn't unusual in the past for "diet" breads and other baked goods to include powedered cellulose filler that was essentially byproducts of the lumber and papermill industry (wood and paper "dust").
Cellulose has calories (because it can burn... calories are esentially a measure of "burnability"), but the cellulose calories aren't available to humans, and yet in the USA (many countries have better calorie-information laws) however in the USA those calories ARE usually still listed in the calorie counting resources (and worse the manufacturers don't have to tell you how they did their math, so you don't know if the fiber calories have already been subtracted... the only way is to check their math).
I also suspect that manufacturer's are allowed to use "stock" or "default" calorie counts for some products (for example, I noticed in many brands of instant oatmeal, virtually all instant oatmeal boxes will list 160 calories as the calorie count for a 40 to 42 gram serving size... even if the math of their nutrition label doesn't add up to 160 calories).
This confused me until a person working in the food industry (on the line of a plant that made packaged food products) told me that a company doesn't have to do nutrition testing if their product recipe is sufficiently similar to the industry standard... then they can just use the accepted calorie count for that class of product).
I don't know this is true for a fact (because I don't know if this is just a rumor this guy heard at work, or whether it's true, but it did make sense to me, in regards to what I was seeing on these oatmeal packets).
Because of all the ways in which calorie counts can be mis-estimated, I prefer counting my calories by an exchange plan. If I eat mostly whole foods, and weigh my food, I can get a pretty decent estimate of calories - but I know I can easily be off by 10 to 30 calories or in some cases more. Usually though the discrepancy is only about 20 calories max though.
When I count calories to the nearest calorie (or even the nearest 10) this inaccuracy bugs the hades out of me, and I get all OCD and crazy about whether that peach was a 55 calorie peach or a 65 calorie peach...
I can also start using weird calorie logic (that 50 calorie oreo is a better choice than that 70 calorie apple, because it's 20 calories fewer).
So exchange plans help me control my calories (an 1800 calorie exchange plan is going to average 1800 calories) and do it without going all nutty with the math. Yes, I know that my small piece of fruit might contain 60 calories or 75 (the plan assumes 70 calories), but it really doesn't matter in the long run. Eating by "food group" or exchange categories helps me eat a more balanced, more whole food diet (because whole foods are easiest to count), and most of the counting has been done for me.
If I do come upon a food that I don't know and can't find the calorie count for, I decide what food group it's from, look up the calorie count and figure out how much of it I can have to satisfy that exchange (I know that sounds hard, but it's not). For example I didn't have the exchange count for longan (a fruit that sort of looks like a skinned grape covered in a leathery suit) so I looked up the calorie count and determined how many pieces I could have for 70 calories (about 11 fruit). This was pretty close to my own estimate because I knew the fruit was very much like lychee in flavor and size and I knew that I could have 10 lychee fruit (longan are a tiny smidge smaller than lychee so this makes sense). If I'd not been able to find the longan data I would have used the lychee data and been happy (and I only would have been off by 4 calories).
Most fruits I can guestimate very well just be tasting it and comparing it to the most familiar fruit I eat regularly (and have memorized the exchange values for). Most berries I get about a cup (even though with raspberries I get 1 cup, blackberries I get 3/4 of a cup, and strawberries I get 1.25 cups... which is why I usually choose strawberries).
This comes in handy when I'm away from home (as I don't have a smart phone), and I'm rarely wrong (I don't get home and find out that the fruit I ate was twice the calories I thought it was. I'm rarely off by more than 10 to 15 calories... and not just for fruit, but for meats and breads and such as well... though if I had a portable scale I could be even more accurate when eating away from home).
As usual, I've written a novel to say something relatively simple: I never trust the manufacturer to have done the math correctly. I always double check based on what I know about the class of food (for example, all cereals run about 80 to 100 calories per ounce, so if a cereal contains way more or way fewer calories according to the label, I become suspicious and I look in the ingredient label to explain the discrepancy).
Cereals that contain more than 100 calories per ounce, usually have added fat (or fatty ingredients such as nuts).
If a label seems "too good to be true," I assume the worst (and look to the ingredients to see if I can confirm or debunk the label).
And yes I do eat fewer processed foods because it's harder to decypher the label (how many calories are in xanthan gum, ascorbic acid, modified food starch? I don't have a clue... though I know I could look it up, but even I'm not willing to do THAT much math).