I began to notice the number of "steps" I took each day, and I actually liked when it told me to "move" after an hour of inactivity. There are still some kinks to work out, like when it told me to "move" while I was riding my bike at 20 miles per hour, with the GPS engaged.
I started to wonder if the 920 could give me a more accurate count of the number of calories I burned each day than my usual method of entering all my exercises into MyFitnessPal.
So for the last couple of weeks, I decided to compare the calories tracked by my Garmin and MyFitnessPal.
This very complicated graph gives you the results:
The light blue and the yellow lines give you the raw data from MyFitnessPal (MFP) and Garmin. As you can see, there are some wide discrepancies. On March 15 MFP counted 1680 calories versus 3726 for Garmin. There are also a few days when the Garmin count was significantly lower than MFP.
I noticed that the days when Garmin was way lower than MFP were days when I did activities that the Garmin doesn't capture, such as the elliptical trainer and strength training. So I made an "Adjusted Garmin" count (in brown) that took care of most of that -- but the two numbers were still off by quite a bit.
Then I noticed that the days when Garmin was way higher corresponded to long rides. Clearly Garmin is giving me way too much credit for my calorie burns on rides because I was basing my eating off the MFP numbers those days and I didn't notice feeling exceptionally hungry and I certainly wasn't losing tons of weight. So I created a "double-adjusted" Garmin count (in red on the graph) that ignores about 40 percent of the calories burned cycling. As you can see, the red count is much closer to the light blue MFP count.
But Garmin was still giving me credit for about 30% more calories than MFP, and MFP wasn't giving me any credit for days when I took a lot of steps (like when I was timing a race). So I created an adjusted MFP count to account for that. This one is just 18% different from the Adjusted Garmin count, and there's less day-to-day variance as well.
So it seems I could get a decent count using the activity monitor on my Garmin if I could get it to track the few additional activities like the elliptical trainer. Since Garmin consistently overestimates calories, I would have to factor that in, but in the end I could have a totally-automated system for tracking calorie burns.
Another option would be to get a separate activity monitor that also tracks heart rate. Since heart rate tracking can give more accurate calorie counts, this would solve two problems at once. Anyone want to get me one of these babies as a gift?