Best way to count calories burned during exercise?

I think a FitBit is very accurate. It is all based on your heart rate.

enigmaticphrase wrote: »

Does anyone know what the most reliable way to count calories burned during exercise is?

I find that when I use an elliptical or cycle in the gym, I am given a wildly different burn estimate than when I enter that exercise into this app. At first I thought that perhaps it was because the app has my weight and stats but the discrepancy remained even when I entered my stats into the machine. Sometimes it is as much as 200 calories different. The machine almost always estimates fewer calories burned than the app does.

So what’s more reliable — the estimate in my fitness pal, or the estimate on the screen of the machine?

Go with the most conservative estimate. For cycling, if the bike has a power meter that measures watts, that is going to be the most accurate for cycling. Most everything else is pretty up in the air. A HRM can provide a reasonable estimate for calorie expenditure, however, algorithms used assume a steady state cardio event and an average heart rate...the further you get away from those things, the less accurate a HRM will be.

For example, someone who is pretty unfit is going to have a higher HR performing a particular exercise than someone with a reasonable fitness level or high level of fitness, so the HR would inflate the calorie expenditure. Also keep in mind that when using a HRM, in most cases you're getting a gross expenditure, not net...meaning the calories you would have expended had you done nothing are also included in the total.

I echo the power meter answer. For the Concept 2 rowing machine, use the calorie calculator on their website (which requires knowing the stats from the workout - find them in the memory area). While I almost never use their BikeErg, I really don't trust the calorie estimate for that.

It's a cr*pshoot, mostly. It's all estimates, though some have better odds of accuracy than others in certain scenarios. Fortunately, close enough works fine, in the context of the weight management process.

OP, in your example, I'd use the lower estimate, and go on with life. If your weight loss ends up about where you expect after a month of logging everything, you're golden. Otherwise, adjust intake.

More abstractly, here's what I think, for cardio(-ish) things:

- Machine estimate if well power-metered (cycle power meter, Concept 2 rowing machine + web page calorie adjustment): Closest accuracy.
- Heart rate monitor or fitness tracker for steady state moderate intensity cardio, if you know your actual max heart rate and the device will use it if you enter it. Possibly slightly better odds of being accurate if the device knows what specific activity you're doing.
- For walking, compare the flat-terrain rule of thumb estimates for net calories (I always forget the multipliers: Is it 0.3 times distance in miles times body weight in pounds for walking, 0.6 (0.65?) multiplier for running?). I don't much walk/run, so am not sure.)

If intervals (especially high intensity intervals), it's a little dicier. I'd compare sources and use the lowest, personally.
If a fitness tracker doesn't know your true max heart rate or can't use it, also dicier for any intensity. Use with skepticism, and compare other sources. Use RPE (rate of perceived exertion) scales to guess whether your age-estimated HR max is reasonably close, as you get fitter and can work harder; get tested (athletic capability test, not medical stress test) if you're able, when fit enough; or find a suitable sub-maximal test for estimating it, when you're fit enough for that.

Be sensitive to whether the source is giving you gross calories (includes at least resting metabolic rate on top of the exercise-specific increment) or net (without RMR), generally . . . but especially for long-duration and easy-effort activity (like casual walking, where RMR is a bigger factor). You can back out an estimate of RMR, to improve gross estimates.

By "other sources", I'm suggesting finding online calculators specific to the activity (from rational organizations, not "Joe's calorie calculator site"), looking at the Compendium of Physical Activities** for comparative METS, looking at the MFP estimate (mostly METS-based), and that sort of thing.

When considering relative accuracy, I don't get excited about what other people seem to consider big differences: If my TDEE is around 2000, for example, and two sources are say 50-75 or so calories apart for an hour, I'll probably use the low one, but I don't consider them meaningfully different even if that's a fair fraction of the exercise calories. That's down in the noise of error rates on estimating food, TDEE, etc., IMO. Close enough.

For strength training:

- I'd use a METS-based estimate like MFP's myself. It's probably still wrong, but it's about as good as it's going to get. HRM estimate is very likely to be wrong, possibly very wrong.

Personally, I'll use the weight-adjusted Concept 2 or fitness tracker estimate (with chest belt) for machine rowing (I've learned that they aren't usually dramatically far apart for routine workouts, for me; my tracker knows I'm machine rowing, knows my tested max HR).

I use the tracker estimate (with chest belt) for on-water rowing because I don't have anything better, and my machine rowing experience suggests it might be in the ballpark. (I always use the chest belt because the wrist HR doesn't work reliably with either kind of rowing - too much arm movement). MFP database estimates for on-water rowing are extra dumb: They suggest that rowing or canoeing at the same speed burns the same number of calories, which seems ridiculous to me, as someone experienced with both. (Even different canoes burn different numbers of calories at the same speed, if you ask me.)

For spin class, I use the tracker estimate (no power meter, and what the bike says when it picks up my HR is clearly absurd) - I did compare online calculator estimates when I first started logging.

For strength training or yoga/stretching, I use the MFP METS-based estimates, which are low and therefore plausible. (I rarely count yoga/stretching at all, unless I do much, much more than a normal short session).

** https://sites.google.com/site/compendiumofphysicalactivities

Jthanmyfitnesspal wrote: »

@jhanleybrown : The formula you list comes from this:

kCals "burned" = (Watts applied) * (3600 seconds/hour) / (4184 kcal/joule) / (human metabolic efficiency factor)

The first two factors are just unit conversions. The final one is squishy, anywhere from .2 to .25, I've heard, depending on a number of factors, including level of fitness and pedaling style. This is a 20% overall error.

So, even though a cycling power meter is an absolute measurement of Watts applied, you are still in a bit of a pickle for getting your calories burned to better than ~20%. Oh well, that's probably about as good as any estimate!

But, given that the factor won't change very quickly for an individual, a power meter gives you a very good relative measurement of calories burned! I would get one as a fun toy, but they're still rather expensive and my numbers won't be very good, anyway.

@AnnPT77 : Wow, what a comprehensive answer!

@Jthanmyfitnesspal
Sorry your maths and therefore your conclusion isn't correct.

Re the bolded - efficiency is normally stated in percentages so 20% to 25% and the difference isn't therefore 20% it's 5% at most. It's not the size of the difference between 20% and 25% it's the range of estimates versus reality.

However, most cyclists aren't at 20, 21 or 25% - they would be outliers. So assuming 24% (which the simplified x3.6 formula does) means the effective inaccuracy is very small. The high efficiency outlier would be over estimating by 1% and low outliers by underestimating by 3% or 4%. Which in the exercise estimating world is extraordinarilly small.