Hacker News Reader

Comment by jackvalentine | Hacker News Reader

Having tried implementing calorie counting... how exactly do you do it?

You either cook everything yourself and weigh every ingredient. Or you buy totally packaged meals (Lite 'n Easy, etc), or you buy fast food (And even then, the calorie counts in fast food are... inaccurate at best).

If I go to the corner shop and get a sandwich I have no idea what's in it and there is no accurate way to estimate.

While it's good to be aware just how energy dense foods are (like muffins... aka breakfast cakes) there is no method that I'd consider both achievable for the average person to consistently do AND accurate enough to be considered 'counting' as opposed to washy estimates.

View on HN

I calorie counted for years.

You take things apart and weigh them, or make good estimates based on total weight. You don't need to be very accurate, anything within ~25% error is fine. You just need to be very consistent.

The USDA has a great database that provides you all the raw materials (and many of the packaged composite items) to use, and once you've eaten a thing and recorded the values, you can simply reuse them infinitely. One ham sandwich, if you're consistent about it, is a given calorie intake.

The key is what that lets you do: It lets you analyze your intake versus body composition and weight, determine whether you're eating too much, too little, or a bad ratio, and then make slow changes over time to fix that problem.

If you're like me right now, you're overweight. Okay, look at your calorie count, reduce it by say 5% or 10%, wait a month, see how your body composition changes, and reanalyze.

With this method you don't need to be dead-on-balls-accurate with your calorie counting, only consistent and reasonable.

As bizarre as it may sound, I have found that buying only pre-packaged food for a few weeks really helps people get a ballpark feel for what 1900 calories (or whatever) looks/feels like. Because the calories on packaged food exist by law (even if they aren't always completely accurate), it's a really easy way learn about it. After that people can develop some intuition (like, "I only ate a few mouthfuls of cheese. How many calories could that be?" or "It's just trail mix. It's healthy. Peanuts and raisins. Like 0 calories!" or "I just have a single bran muffin and a coffee for breakfast. It's not even a meal" or... well you get the idea).

Some years ago I spent a lot of time calorie counting. I got pretty good at estimating but more importantly became keenly aware of where the calories are hidden. You start to develop simple rules of thumb: beware of the breads and sugars, fruits are full of vitamins but can pack a calorie punch (bananas and grapes on the higher side, strawberries and melons being in the low side), most raw vegetables, negligible. Lean meats, surprisingly low relative to their help in making you feel well fed. Be cautious with cheese; sauces and dressings can be killer.

You don't have to be perfect (or honestly even that precise) but it's immensely helpful to do the exercise for a few weeks to dial your brain in. When you calorie reduce at the same time, you grow razor-sharp appreciation for foods that you can eat a massive amount of that provide few calories. When you're hungry, a choice between a muffin and a tray overloaded with vegetables and turkey slices becomes a lot more clear.

I've often wondered if an app that has me take a picture of everything I eat in a day and then sends me a collage of all yesterday's consumption by email in the next morning would be a suitable replacement for calorie counting. I wonder if the process of taking my phone out to photograph things before I eat them would cause me to more carefully consider my choices, and the private shaming the morning after on days I don't do well would serve as a behavior modifier for the day.

> If I go to the corner shop and get a sandwich I have no idea what's in it and there is no accurate way to estimate.

uh, why not? i'll do it right now for you off the top of my head, which anyone can do after they look them up for a while.

100g of white bread = ~ 250 calories / 2 tbsp of mayo = ~ 200 calories / 4 oz of deli turkey = ~ 100 calories / 1 tbsp mustard = ~ 10 calories / 2 slices (2 oz.) of cheese = ~ 240 calories / 2 slices of tomato = ~ 20 calories / 1 slice of lettuce = ~ 5 calories / 1 bag of chips = ~ 280 calories (probably on the bag)

there. 1105 calories total. is that perfect? no, but is that useful? yes, a whole lot more useful than "it's too hard, so i might as well not do it."

as an aside: this is why people are fat. a sandwich and a bag of chips is insulin-spiking can easily be over 1k calories, and most people would probably guess it's 300 calories, and 'not fattening', whatever that means.

throw in a soda and you've got a recipe (literally) for disaster.

I was about to start writing a similar comment, but you saved me the time. I just wanted to emphasize that this is why people consistently under/over estimate how many calories they eat. Every single person who has asked me for help either gaining or losing weight has thought they knew how much they were eating until they actually started counting.

You're literally guessing the quantities of the ingredients of the sandwich. (and mixing measurements... grams AND ounces?)

How is that counting? That's my point.

You're getting a vague idea of what you're eating but it is in no way calorie counting.

mixing g and oz - that's a novel and interesting excuse to avoid the work of weight loss. usually people just blame their thyroid but this bullshit excuse at least has some kind of passable scientific pedantry behind it. i'll have to file that one for later.

https://en.wikipedia.org/wiki/Analysis_paralysis

Who said anything about avoiding weight loss?

I've had far more success just keeping a simple food diary and going "wow I eat a lot of cakes" and cutting them out.

I have a theory that 'calorie counters' are shockingly inaccurate (one other poster mentioned a 25% margin is 'fine'!) and their success comes from the very act of attempting to quantify what they're eating being a defacto food diary.

it doesn't matter if it's accurate, it matters if it's precise, and consistent. kind of like the scale you stand on, or weigh your food with.

I disagree that you can possibly be precise and consistent across foods without preparing them yourself.

I actually agree with you - to a large degree, counting the calories is enough to change behavior in itself, without even attempting to modify your diet at all.

The point remains, though, that it really doesn't matter how accurate or precise you are, simply that you're gathering data and using it to make measurable changes in your diet.

Your way is perfectly fine too! The nice things about calories (just like money) is that they're a universal medium of comparison, so you can compare your cakes to steaks. But if you simply want to look at a category and say "I am eating N of these, I need to eat N-1 to lose the weight", that's perfectly functional!

The thing I hate is when people make totally non-empirical diet changes, and then lament that they aren't losing weight. You just have to measure and adjust.

My point though is what you're "measuring" is quite difficult to do in a clinical setting, let alone every day living your life.

The above poster talking about the deli sandwich mentally breaking down all the ingredients... I'm happy to be proven wrong but I simply don't believe you can do that with a reasonable degree of accuracy.

I suppose this is turning in to quite a pedantic argument about the definition of "count". I'd call what everyone has said they're doing more accurately a "calorie estimating food diary".

My initial post was genuinely curious about how people can be so accurate when eating foods from a variety of non pre-measured sources. It turns out they're not being accurate.

>The above poster talking about the deli sandwich mentally breaking down all the ingredients... I'm happy to be proven wrong but I simply don't believe you can do that with a reasonable degree of accuracy.

You are just overestimating both the difficulty and the degree of accuracy required.

While at the same time, still relying on even more vague terms, to determine if you ate too many cakes etc (as per your other example).

I'm not claiming any degree of accuracy in the estimates, just that without tricking yourself in to think you're being rigorous you can achieve pretty much the same results. I outlined the mechanism that I think these 'counters' are actually achieving results by in spite of their atrocious data collection.

I can't believe people are defending such poor data collection practices. You'd be all over it if someone else was selling their results based on such inaccurate data but in this case it's fine?

If you're going to call it counting you need to be accurate.

>I outlined the mechanism that I think these 'counters' are actually achieving results by in spite of their atrocious data collection.

The mechanism is simple: they reduce their caloric intake, because they can track how much they eat. More or less: it doesn't have to be perfect, nor is it "atrocious" if it isn't. And you can easily just round the numbers up ("I calculated 500 for this thing, but let's say it's 600 just to be safe").

You seem to believe that any kind of "back of the envelope" / "ballpark" calculation is useless. Or that people only eat complex multi-part meals with no nutricion information, and have to gauge everything from zero all the time.

So I went looking for any studies about peoples ability to estimate calories, since any under/over estimation compounds either way in terms of results.

It turns out like everything, that's a tricky thing to study. When people are aware that their meals are going to be scrutinised they change what they eat for the duration of the study.

I wasn't able to find a recent study that investigates people's ability to estimate calories and the variance in their estimations.

> You seem to believe that any kind of "back of the envelope" / "ballpark" calculation is useless.

I do think that anything above a "good, bad, not sure" estimate is probably going to be so inaccurate as to not be worth the effort. However the act of trying to count calories itself promotes a mindfulness of what we're eating and that can induce change.

>I have a theory that 'calorie counters' are shockingly inaccurate (one other poster mentioned a 25% margin is 'fine'!)

Accurate or inaccurate is only meaningful related to the task and its requirements. 25% can be totally acceptable margin of error for the task. We use even bigger margins in lots of ventures (determining which startup will have a succesful exit to fund, for one).

And yes, the mere act of quantifying helps. But quantifying with even 25% error is still better than just writing down "ate 5 cakes", especially if one doesn't eat too many repeats of the same food.

Not sure in what reasoning one can complain for a method with 25% margin of error (say), but be OK with a method like "5 cakes" which still applies quantification, just in an even more vague and hazy sense.

5 cakes is much worse than 5 carrots, for example, but with merely writing down how many you ate, you have to rely on a far more relative guesstimation of their relative harm than you would be if you were counting their calories and being off by 25%).

> Not sure in what reasoning one can complain for a method with 25% margin of error (say), but be OK with a method like "5 cakes" which still applies quantification, just in an even more vague and hazy sense.

One method implies rigour and the other one is honest about what it is setting out to do

But you seem to be the only one assuming and/or bringing up rigor in this discussion.

Everybody said it's a quick ballpark figure / back of the envelope style calculation.

The mere fact "cake bad, carrots good" everybody knows. It's not much information concerning "Did I ate too much today?". One can have a caloric budget and stay within it (more or less) without having to be perfect in measurements (or sticking to carrot because it's easier to know its light).

>You're getting a vague idea of what you're eating but it is in no way calorie counting.

No, that's exactly calorie counting in the real world.

This is not advanced math or chemistry lab, and you don't need to get to even 80% accuracy for everything. Just to be consistent and get a good feel.

It's close enough to get results. Nobody is going to put a sandwich from the local sandwich shop into a bomb calorimeter to figure out how many sandwiches they can eat and still lose weight.

If you're going to get hung up on the term calories let's call them approximate calories (or ACs) instead. If I eat mostly the same foods all the time I can use reasonable guesses to get the ACs in each food. Foods I eat infrequently will probably have a greater deviation between my ACs and the actual calories but that doesn't matter much because I don't eat them very often. Foods I eat frequently will probably have less error and the error doesn't matter much as long as I'm consistent with the numbers.

Let's say I decide I want to lose weight. I can start off with the general rule that for a male of my age, build and activity levels I need about 2500 calories a day to maintain weight. Since I want to lose some weight I'll set my daily AC allowance at 2000. After a couple weeks I check how the weight loss is going. Am I losing weight too slowly? Drop the ACs a bit. Am I losing weight too quickly? Increase the ACs a bit. Keep checking and adjusting every couple weeks until I'm happy with the weight loss.

This approach works in the real world. I do weigh and measure some of my food but if I had to weigh everything down to the last gram to make sure I ate exactly 1947 calories per day for ideal weight loss I'd quit pretty quickly.

A can of soda is only about 140 calories. I.e. half the calories in the bread. I think it gets a way worse reputation than it deserves.

What about a medium sized soda at a fast-food place? Those places use consumer surplus as an incentive to get a larger drink, so seeing people with huge cups, presumably filled with soda is very common.

Besides, 140 calories is a lot, considering it has about the same calorie density as milk and the number of great drinks available to you at 0 calories, like black tea, green tea, coffee and plain water.

Soda is sugar only, always, so it's much easier to throw your daily balance of carb/protein/fat off with soda than any other drink (e.g. milk).

One time, as a continually hungry teen, I did a one day calorie count for a scout project. I was suitably horrified to find out at the end of the day that I had consumed around 10,000 calories.

I had success with a smartphone app (loseit) and estimating. You really don't need to be more precise. Lost 3-4 kilos in a month. Doing a rough calculation will often do the trick, that one sandwich could be 400-500 calories, and you had two beers, which is already 400-500. A weight losing calculator suggested you can eat about 2000 calories that day. So that's already half o your daily intake. But if you go run or cycle for an hour you can eat maybe 300-400 more. Keeping track breaks the bad eating habits.

I have a scale in my kitchen and I weigh everything. For most restaurant food you can look up the calorie information really easily.

So for me it is a very small minority of food that I can't get an exact number on and I've been weighing and entering things enough that for those edge cases I can usually do a decent job at eyeballing serving sizes.

If you go to the corner store and get a sandwich, just put a rough estimate of a sandwich based on the sandwich options in your app and what you think seems right. You learn to make smart estimates when you start counting the calories on everything. Even then, washy estimates are just fine, just put them on the conservative side (ie, overestimate when you have to estimate).

Calories are never an exact numerical science anyways, even if you did weigh everything precisely. You really don't care if you have a margin of error of 50-200 calories if you are lower than your goal most days. You still loose weight.

I've been using MyFitnessPal on and off throughout the years and have always been able to consistently gain or loose weight (depending what I am aiming for) even when relying highly on estimates. This is just my experience though.

The Fitbit app is great for counting calories. Very easy to use and it has a great database.

Yeah it's interesting, I find myself choosing a potentially less healthy food option simply because it's nutrition information is more readily available and easily quantifiable.

Spend a couple of weeks/months preparing your own food and weighing everything. Use an app like MyFitnessPal to record and count your calories. It doesn't take very long to develop an intuition that's "good enough" about how many calories a deli sandwich contains.

2% extra calories per day ~= gaining 6 pounds a year. I doubt you can get within 2% making this far less useful.

Basically, you need a feedback mechanism like your actual weight. Counting calories is not about actual calories it's a way to cut back if your not getting the results you want.

Yes, sorry I should have mentioned that. You need to track your weight along with counting calories and make adjustments to your caloric intake based on the weight you are gaining/losing. You don't need to be within 2% accuracy (frankly that's not feasible anyways) just consistent.

I do strict calorie counting. I cook most of what I eat and do volumetric estimation on the rare days I eat out.

Reply on news.ycombinator.com