Activity Level & Eating Back Calories

If your Fitbit is accurately predicting your burn rate, the answer is that you should be eating back all of your caloric adjustment as shown AT MIDNIGHT, which is when the final accounting reconciliation takes place between Fitbit and MFP.

Please note that Fitbit and MFP substantially use the same baselines and formulas to estimate your burn. But it is an estimate and for some people the estimate will be close to accurate and for others less so.

Only your own logging and your own weight trends corresponding changes will allow you to figure out how closely the estimates match reality for you as an individual.

That said, erroneously high burns are more likely to be recorded by Fitbit when one's heart rate is high. Regardless of etiology Fitbit tends to interpret higher heart rates as higher burns.

And, with a larger burn, even a 5% or 10% error will be many more calories than for a smaller burn.

So I would more trust the 200 Cal on a sedentary day than I would trust the 2000 day on an extreme sports day.

This doesn't mean the 2000 Cal is not valid at all, or that it is only valid as 1000, but that I would more expect to find 150 Cal worth of error in the 2000 extra day than in the 200 extra day!

Again, of course, your own logging and results determine how close to estimates you're tracking. You can usually use 3500 Cal as equivalent to 1lb of weight trend change.

I can really relate to this question, since my calorie burn on exercise days is soooo much higher than on my rest days -- my primary form of exercise is cycling (road and mountain, but mostly on a stationary trainer this time of year). I don't wear a fitbit, but I do have a chest strap heart rate monitor that I wear only when intentionally working out, so my background daily activity is not logged as yours is. Over time, I have learned that the estimated calories burned based on my workout metrics are a little bit high, but not totally out of proportion.

In other words, over the long run, I do seem to drop about a pound for every estimated 3600 calories burned.***

On a side note for those that might be reading this and logging in a similar way, I have noticed that I need to log an effort tier lower than my actual perceived exertion to match the estimated calorie burn from my heart rate monitor. For example, last night I did a really tough 60 minute ride on my stationary trainer with an overall perceived exertion level of 8/10. My calorie burn (based on heart rate and power meter) was 592. Now, even though this workout felt like it should be logged as "Bicycling, 14-16 mph, vigorous (cycling, biking, bike riding)", to get the correct calorie burn I logged it as "Bicycling, 12-14 mph, moderate (cycling, biking, bike riding)". Anyway...this may differ for others, but maybe it will help someone out there. I don't think it is any secret that MFP estimated calorie burn for activities is notoriously overestimated.

For this reason (and this goes back to your original question, OP) I have had really good results eating about 40-70% of my calories burned from exercise back. I use that range as a guideline to account for any further errors in exercise calorie expenditure estimates, which are undoubtedly a mildly to wildly inaccurate more times than not. And if I'm really hungry and eat back closer to 100% sometimes, I don't get worried about it.

On a second side note, I can really see a performance drop off when I don't eat back at least SOME exercise calories. I have come around to realizing that it's not worth it to me to stress about this -- I would rather be improving my fitness with a slower rate of loss than be overly conservative on eating back these "extra calories" and have my energy levels take a hit. This helps me with athletic performance AND with energy levels and focus throughout the day at work and so on.

I love the idea of using your own results as feedback in an experiment (as PAV8888 hinted at above). Once I came around to that line of thinking, fitness and nutrition became more of a game and less emotionally-charged. The curiosity I have developed around how the data (calories in/calories out) were affecting (and in turn influenced by) my behaviors has been a valuable tool for me. Of course, there are many other variables at play that could affect your performance and results (mood, hormones, schedule, etc.), but it it very gratifying to get a handle on the quantifiable stuff.




***Interestingly for me, this works out to be roughly every 100 miles of cycling (at least on the road and/or trainer -- mountain biking is a whole 'nother thing). I tend to average around 15 or so mph (including recovery rides) and burn about 500-550 calories per hour. 6.66 hours of that effort (about 100 miles) equals 3,300-3,700 calories burned. Usually I am not restricting calories while doing this, so just eating around maintenance plus most of my exercise calories. I have only tracked this when I am doing super consistent training, like for an event or race -- and it's easier for me since I am so single-sport oriented -- but I find it really interesting!!

PAV8888 wrote: »

On an interesting note, issues exist with all sorts of our calculations and as individuals we do not, necessarily, track directly to the population averages which the formulas use.

One reason for the "eat half the calories" mantra that you often hear is a bit of a misunderstanding caused by the base levels of calories assigned to the time slot by MFP and by the meaning of NET calories in the abstract and in the context of the base level of activities selected on MFP.

MFP assigns calories to each minute of the day. To wit AF*BMR calories, with AF=1.25, or 1.4, or 1.6, or 1.8 depending on the activity level we select.

MFP and most devices use the compendium of physical activities to come up with a MET value for each minute of activity. (Fitbit, I believe, averages the detections over 5 minute increments and back-assigns the average of each 5 minutes to the 5 minute time slot. You can see how an activity that raised your heart rate and kept it high may "log" more calories than you were burning during activity lulls where your heart rate continued to be high even if your work output had lowered).

A simplified view of MET values is 1MET = 1x BMR per minute of activity.
A "normal" NET burn would be the value burned less 1x BMR.

As you can already see even a "true" NET value for calories burned would be skewed by the fact that MFP has already assigned at least 1.25* BMR calories for the time slot, not just 1.0

In any case... if you play around with the numbers and take a long duration moderate activity such as a long walk, you end up with gross calories in the 3.0 MET range. Which means that using a sedentary setting on mfp which assigns 1.25BMR to each minute of the walk, only 1.75* BMR of that burn was an extra burn, which means that only 58% of it was legitimately edible if you're tracking at the population mean. If you were setup at lightly active (1.4), then only 53% was a legitimately additional burn. Add a little bit of logging inaccuracy here and there and a desire for faster results and "eat back half" gets born.

This doesn't mean that there was any recording inaccuracy. Only a misunderstanding as to what was an actual additional burn!

That said, external devices such as Fitbit specifically calculate a TDEE value independently of MFP for the day and reconcile that TDEE value back with MFPs corresponding TDEE value.

As such the "eat back" calories given as an "exercise adjustment" are the difference between the TDEE values of Fitbit and MFP. And are legitimately presented as "detected net additional calories" that CAN be eaten (at midnight when the final reconciliation takes place).

The foible of Fitbit to assign a few extra calories to higher heart rates and movement is a different issue and may or may not be experienced by all other users.

For levels of more fun, the integration of Apple in particular with MFP (and last time I had read a thread with @heybales exploring things, the integration of Garmin watches too)... is not very correct in that the devices are NOT sending to MFP the whole day TDEE values it is expecting to receive.

Consequently the exercise adjustment shown ends up being very wrong making life more challenging for MFPers everywhere!

You left out one additional piece of the fun, @PAV8888: While it's true that the trackers (like Fitbit) make a more personalized and nuanced estimate of TDEE (compared with MFP or a TDEE calculator) . . . it's still producing an estimate of calories based on population averages.

A good tracker will produce a good TDEE estimate for most people, because most people are close to average (small standard deviation in these data, tall/narrow bell curve, loosely speaking). But not necessarily equally good for everyone.

It's all estimates, based on population averages . . . and we're each individuals.

But here on MFP, so many people think deviation from expectations is all about the exercise calories . . . when it can be any of the other estimates, or a combination of them all, leading an individual's results to differ from any of these sources (above and beyond the obvious flaw you mention with how MFP implements the METS-based estimates, which as you say can be quite meaningful for longer less intense activities, but possibly less materially so for short and intense ones).

PrimeSuspect87 wrote: »

Hi,

My current weight is 190lbs and target weight is 160lbs as I am 5.8ft.

I have a desk job and usually in the evening I just watch TV. I have set the activity level on MFP as Not Very Active which gives me base rate of 2194 and as I want to lose 1lbs per week MFP suggest I should eat 1694. I use fitbit to track my activity

Since the new year I have been active 5/7 days.

Yesterday fitbit gave me an adjust of 1247 calories. This came from 13000 steps plus resistance training. I would say 3000 steps was from just normal moving around the house based on desk job and the other 10000 was from playing sport for an hour and twenty minutes.

Now this gave me a total calorie burn of 3441 by fitbit.

Does this mean I should be eating back atleast 50% of those 1247 calories adjusted by fitbit if not all?

Also on the weekend I am inactive and end up with around 3000 to 4000 steps , the usual moving around the house. Now as I've set myself Not Very Active, when fitbit syncs those 3 to 4000 steps shows around 200 calories burned and MFP suggest I could eat additional 200 calories. However, I don't do this as I am inactive most day. Is this correct or should I eat those 200 calories back so my deficit stays at 500 to lose 1lbs per week?

Thanks in advance.

How long is the resistance training?
Looking at HR graph for that workout - how interval in nature was the workout?

That provides the worst inflated calorie burn using HR - which most Fitbit models use.

Some models use a method that when you select Weights as the workout, the rate of burn is correctly around 3.5 x BMR rate.
You can math that out to confirm.

Or go back to that day, notice the start and duration time of the Activity Record, and manually create a Workout Record of the same for Weights, and see what the given calorie burn is. That's more accurate.
You can leave that, it replaces the calorie burn on the prior Activity Record.
Fitbit is replace-only system, last added is used in figures.

Eat it all after that correction, per the reasons given by others.

[/quote]

Oh - you can select whatever entry you want, pick the real average speed you went - and can replace the calorie burn with your own better value.

[/quote]

@heybales holy smokes! I have been using MFP for **QUITE some time** and never knew you could/should do that. Thank you!

This is a fascinating discussion and giving me a lot of new perspective on how external devices interact with MFP. Thanks to OP for getting this conversation started and to everyone else for your input.