Calories way off when I switched from zip to alta

jayemes

Got an alta for Christmas and just started using it yesterday. With my zip even though the calorie burn was high, it was *reasonably* high. I would get about 500 calories burned for 10k steps (and I would assume it was more likely half that)
Just started syncing my new Alta yesterday and for 8k steps it's telling me I burned 1000+ calories. Any idea what's up? Is the Alta just different at calorie estimation than the zip or is there a setting in the setup I'm missing?

malibu927

Regular Alta or HR?

jayemes

It's the regular alta

heybales

Steps results in a distance.
Distance and time gives a pace.
Pace and weight moved give a good estimate of calories.

You cannot merely compare steps and the calorie adjustment.

Find a prior day on old device with same steps about, and is distance the same?
If yes, is daily burned the same (not adjustment, Fitbit daily burned)?


Perhaps the steps is off because of bogus being seen on wrist now, and some adjustments are in order to tweak accuracy.

jayemes

heybales wrote: »

some adjustments are in order to tweak accuracy.

That's my question - is there anything that can be done to increase the accuracy?

heybales

Depends on what you found for accuracy of steps.
There are settings for non-dominant wrist if too many bogus steps being seen.
Are there?

Beyond that, the following are other improvements.

First and easiest is the the stride length setting.
Since for many 23 hrs of day is not exercise, but daily activity, and of that 5-7 is sleeping and BMR level burn, perhaps another 8-11 sitting at work/home. That leaves 5-10 hrs where steps has biggest influence on calorie burn.
Mainly for walking, unless you do a fair bit of running weekly also.

Since Fitbit is attempting to adjust each step to it's dynamic stride length right then, based on the static stride length figure, it's best if the static setting is right in the middle ground of you potential lengths.
So not grocery store shuffle end of range, not exercise level walk end of range.
Setting the static figure at either end of range means the other ends gets inaccurate when done.

And in the scheme of exercise level pace being done during the day, if that's for only 1 hr, a slight inaccuracy there is to be favored over the much greater time of daily level pace being slightly inaccurate.

So finding a known distance track (US high school tracks have many times been converted to metric on main markings, have to know the 1/4 mile ones to use).
Attempting to walk say 1/2 mile at avg daily pace - it'll seem very slow out on track meant for going fast.
You log the steps at start and stop.
Now with known distance and steps the math can be done.

Same with running if it's done frequently enough and usually at same pace.
I do Run or Jog/Walk intervals during training to save the joints best I can - and I've had the Fitbit distance be surprising right on to good GPS measurement. Which probably means the walking is underestimated, running is over, but net effect is correct.


After that, next potential improvement is logging workouts - depends on how long and frequent they are for effect.
If device is giving HR-based calorie burns for workouts - some workouts are just going to be inflated using that method - manual is better.
Anything not steady-state aerobic - so interval type workouts and lifting. Now if workout is 15 min few times weekly - then inaccuracy isn't that great.
If device is giving step-based calorie burns - then lifting again, and whole host of non-step-based cardio like biking, swimming, elliptical, stairs, ect.


After that, getting Fitbit to start foundation using a BMR that is more accurate than gender, age, height, weight.
If you have a good bodyfat % measurement (handhelds, scales, skin calipers, measurements can all be decent and terrible depending on how done), and that BMR calc leads to something 5% different.

Then height can be tweaked to improve it. This also requires the manually set stride length.
If a huge change like 10% BMR difference, that height correction could interfere with HR-based calorie burn accuracy (which uses BMI) - but again that depends on how much HR-based calorie burn you are getting anyway weekly, compared to BMR level burn.

jayemes

Thank you @heybales !
I'm going to read through that incredibly detailed response and see what I can do to make the alta work a bit better for me