During conversations where people discuss their running routines, I noticed two constants; they invariably use a smart watch to record their data and they are faster than me.
This trend could be because of observer bias, that people present their best times as their average, people tend to inflate their numbers during comparison, or that they are in fact faster than me. While these points are certainly likely, especially the latter, I've always thought that the inherent 3m to 5m inaccuracy in mobile GPS devices causes an over estimation in distance travelled which a treadmill (my more common means of running) doesn't have.
While there is a fair amount of data on the Internet that suggests this idea is correct, and that the software doesn't account for the variance in all cases, at some point in the future I'll have to test the theory out myself. But in the meantime, I thought I'd calibrate the treadmill I use. Which is what a normal rational person would do. After all, I need to be sure my data is accurate.
Running a length of masking tape around the treadmill belt and subsequently measuring the tape with...well a measuring tape, shows the belt length was 3.135m (for my own future reference). Then marking a point on the belt (again with the masking tape) allowed me to time how long it took to complete a certain number of rotations.
The data showed that on average, the treadmill speed is nine percent more than expected. Frankly this is great news, as it means I'm faster than I thought.
I ran the self calibration function on the treadmill and repeated the test. The treadmill reduced the speed at all settings but still measured an average of six percent faster than expected. This makes me think that the treadmill is compensating for the weight of the individual using it and as such it runs a little fast when not under load.
The next step is to either rig up an Arduino to count the number of rotations while I use the treadmill, enlist a fellow runners help, or make a new video.