4 miles in 31:03. I felt a little tired today, so I just took things easy.
Regarding the watch, I was able to fix the step count problem with a restart of the watch. I'm still not sure why it bugged out on me to begin with. HR again was off for the first mile of my run, recording values below 100 bpm, and above 200 bpm, however it gave me reasonable values thereafter, except for a few instances when there was some obvious noise in the signal and there were brief dips or spikes in the recorded HR. I wish that I could get my hands on the raw data to see what is causing problems as I wonder if I could devise a filtering routine that could fix things.
The distance run is consistently measuring short, as over the past 4 days the watch has given me a total distance run of 4.00, 3.94, 3.98, and 3.96 miles (measured on the course tool at 4.05 miles), while each of the individual runs is showing as 3.9 miles when uploaded online by the watch. Part of the variation I am seeing in distance is due to some of the inherent limitations of consumer GPS devices, and I can't complain too much about the amount the recorded distance is off, however there are still some issues that I would change if I were in charge. I would allow for more user control by allowing the user to force the sample rate to increase to 1 Hz. Currently, data is only collected when movement is noticed, which means that sometimes there is a 1 second gap between data points, and sometimes the gap is 2 or 3 seconds. That is not adequate when you are shooting for peak accuracy. I would also try to improve the TTFF (time to first fix) so that it would not be necessary to wait for so long before collecting data, and improve the app start time so that the app started recording immediately when data collection was initialized. I would also smooth out some of the pace readings. Currently, pace is recorded on a more absolute level so that any inaccuracies in the GPS show up as large variations in the pace being run and make the pace readout look silly. I would smooth out these curves so that the absolute pace being run is actually displayed as a running average over multiple data points. This would tend to smooth out some legitimate brief surges in pace, but would also eliminate the crazy changes of pace where it looks like the running pace has changed by 1:00/mile from one second to the next, and would make this type of data more usable. Finally, I would allow for data to be collected even without a GPS signal. Currently, the watch will not save a run if there is no GPS output from the watch. Sometimes I might want to use a treadmill, etc, and it would be nice if the exercise time could still be saved in these instances.
| 