Just what do the exercise modes do?

[chris2121]

I've noticed some really strange inconsistencies between having an exercise mode turned on or having the Versa non in an any exercise mode (when it should auto detect).  

Firstly I sometimes forget to cancel when I've set the Versa in exercise mode (usually the Bike mode because I cycle most days).  So it might still be in Bike mode for an hour or two after I've stopped cycling.  This almost always results in a much higher than reasonable calorie burn and AZM tally for the non-cycling time than could reasonably be the case.  For example it logged 2000 cal burn over a 5 hour period when only 1 hour or so was actually cycling and much of the rest of the time I was fairly inactive.

Conversely if I forget to set the exercise mode and rely on auto detection (which correctly works out that I'm cycling) I get a much lower calorie burn and AZM tally than I would normally expect for the identical ride.  I did this yesterday and on a route I often take and was working particularly hard (I managed Strava PBs on several sections) the calorie burn was about 50% 'normal'.  

Looking at the data it would appear it's down to the HR tracking and it seems that by setting the Versa into the Bike mode I change the way this works so I log higher heart rates than if I don't.

Has anyone else noticed this?  

[JohnnyRow]

To summarize, in auto-detect biking calorie burn is too low.  Using the exercise app, biking calories seem okay, but higher than expected after the ride if forget to end it.  Have I got it?

I'm going to take the other view, that they are treated the same, unless you can give more data to convince me otherwise.

Auto-detect: gives you 50% of 'normal', but how are you determining 'normal'.  By normal do you mean in the exercise mode, which you also don't trust?  Or do you trust the biking part of exercise mode, but just not the period after the ride?  Make sure it is including the full time of ride.  If you just look at the workout, not noticing start and end time, it could be not detecting the full ride, giving fewer calories.

Exercise app:  Wound need breakdown of those numbers; how many of this 2000 calories were during the one-hour ride, and how many in the 4 hours afterward?  If you did the same one-hour ride in auto-detect, how many calories do you burn in the 4 hours afterward.

[chris2121]

@JohnnyRow You are quite correct in that there's not enough data there to draw any hard conclusions.  And in the great scheme of things it isn't that important, I'm just curious to understand what the exercise modes actually do.

When I look at the data where I've inadvertently left it in bike mode what I can see in the 4 hours after the ride is wild fluctuations in HR with some very high peaks.  So a did a couple of experiments while sitting at my desk working yesterday and tried a couple of different exercise modes.  What I saw was much the same with a tendency to record much higher than expected HR values.  I also noticed that it was dropping HR detection quite frequently and the high readings were most commonly after it reacquired a reading.  The HR would then drop down to a more reasonable value.  

The device wasn't intended to be used like this so perhaps I shouldn't be surprised if odd things happen.  Maybe the lesson is just remember to switch exercise mode on when you need it and off when you've finished.  As you quite rightly point out there may well be a delay incurred if we rely on auto-detect.  

The reality is of course that data like calorie burn from a Fitbit is and can only every be approximate because of all the assumptions that are made in it's calculation.  What I'd like is for it to be consistent so it's possible to compare one day with another.