Garmin Vivosmart HR (8) ~ Statistics for the offset

Statistics for offset between the chest based sensor and the wrist based sensor

Statistics for the offset

   Today, I show a statistics for the offset between the chest based heart rate sensor and the wrist based heart rate sensor.  This is based on the past half month data.  Since the number of samples for each data is different, using the percentage of the total samples for each data to compare them.   As you can see, more than 90
% samples are within less than 4.

   Then, I can look into the worst a few data.  The worst one is a data on September 14.

The worst data on September 14

  The orange is a data for Forerunner 910XT with the chest based heart rate sensor.  In the middle section, the chest bases sensor reading shows much higher value.   However, those reading is probably incorrect reading.   The heart rate should not be around 180 bpm in that section.  So I would assume the data read by Vivosmart HR could be close to the real value.   This means, the offset is due to wrong reading by Forerunner 910XT in this case.

   The second worst data is September 23.

The second worst data on September 23

   The blue is a data for Vivosmart HR with the wrist based heart rate sensor.  In the beginning, there are a several spikes on the reading by the Vivosmart HR.    For this data, I do not believe the reading value by Vivosmart is correct reading and the offset is due to this wrong reading.

The both sensors might have incorrect reading

   This data shows the both sensors have a chance to get incorrect reading.   However, I have no idea why it happens and where it happens.   For the both examples, they show a part of period gets larger offset and the other parts are matched well.   The offset is not due to a problem of the wrist based sensor.   As long as the both sensors read reasonable value, the trend would be close enough.   I am not sure if this offset due to how I wear the sensors or not.  But the bottom line is the both sensors might have incorrect reading.  So as long as, people recognize this, it should not be a big problem for general usage.  The data shows majority of samples are reasonably close each other.
   Please note, this data is collected for running / walking and cardio exercise.   For those type case, the wrist based sensor could be all right to use.   Actually, it is a benefit not wear the chest based sensor.

   I will continue to collect the both heart rate reading data during my training for a while.  To see this result is consistent or not.   Overall, the incorrect reading happens in different type of exercise.  It also happens in random.