[Closed] Accuracy of myfitnesspal

I find MFP to be extremely 'optimistic' in terms of calories burned during most exercises. I've linked strava to it now as that seems a little more realistic for runs and rides.

It is a reasonable guide. In the past it has tied in pretty close to what my Garmin gave me but I always think it is overly generous in its estimations.

Its never going to be the most accurate thing in the world...

It's not an impossible figure. Assuming that the app knows accurate how much energy you've used (which on a road ride it should be able to make a reasonable guess), then it's down to what percentage efficiency myfitnesspal thinks you convert calories to mechanical energy. Typically the range is 20% to 25%, which depending on what value they use gives a potential overall error of about 20%

Are any of the GPS etc apps accurate? Garmin and endomonmdo seem way over the top to me.

Assuming you're on the road they should be reasonably accurate , well within the error above.

This is from the last 90 min road ride with a power meter

.............................avg power.....Work Done......Kcal
Garmin................214................1210kj.............1277
TrainingPeaks...214................1207kj...............1207
Strava..................215................1212kj..............1352

They all pretty much agree on avg power and work done (which you'd expect)
There's a bit more variability on calories but all roughly in the expected ball park.

Going back to the OP 1500cals over 100minutes would need an avg power of ~ 258 Watts, which is ~ 22mph on the hoods on a flat road. So if the OP did 35 miles on flattish roads, then the figure would be about right.

You may not know, and it might be of vague interest.

Training Peaks converts kj's recorded by the power meter to kcals 1:1.
The conversion ratio from kjs to Kcals is almost exactly balanced by the efficiency of Kcals of food to kjs of output for the average human.

I find mine can be a mad at times, for example it might give me 700 cals for a ride, then add 1000 more for extra steps at the same time, it does self-correct after a while though.

Strava isn't a great feed for cals though as it doesn't allow for terrain -doesn't matter if your climbing on a road on a road bike, or hopping up rocks on an mtb the count is the same.

I use mfp and sometimes think 'I just swam off 711 calories?' It lies!

Truth is - I don't know whether mfp calculations of exercise cals accounts for body weight and current daily intakes etc

Reading elsewhere - an hrs swimming can range between 413 calories burned up to 931calories. So 711 seems about right for my weight and speed at least based on the below:

[url] http://www.active.com/triathlon/articles/how-many-calories-does-swimming-burn [/url]

A 130-pound person swimming freestyle for one hour will burn 590 calories swimming fast, and 413 calories swimming slower.
A 155-pound person swimming freestyle for one hour will burn 704 calories swimming fast, and 493 calories swimming slower.
A 180-pound person swimming freestyle for one hour will burn 817 calories swimming fast, and 572 calories swimming slower.
A 205-pound person swimming freestyle for one hour will burn 931 calories swimming fast, and 651 calories swimming slower.

Calorie estimates are generally imprecise unless you have something like a heart rate monitor or power meter to help guide the estimation. Even those can be way off. I find Strava the most consistent and conservative in its estimates, whereas Endomondo is all over the place.

Polar are meant to hold the algorithm for the best guess. IMO the Garmin one is laughably over optimistic. No way do I believe you'd burn 1200 calories in 90 minutes as per Ian's stats.