Last weekend another measurer and I measured a long course in the local area. Because of its length, we elected to set a calibration course approximately midway. We measured approximately the first half of the race course, re-calibrated, and proceeded to finish the course measurement.
When we sat down to calculate the course length, it became apparent that his splits over the first half of the course were consistently shorter than mine. Over the second half, though, we were within a meter.
We compared calibration data. On the mid-calibration and the post-calibration, our numbers were consistent; his counts over the 300 meter course averaged 1.75 to 2 counts less than mine. On the pre-cal, though, they were almost the same. My pre-calibration counts were within 1, but his varied as much as 5 from highest to lowest. His counts on the other two calibrations were much more consistent with each other.
I remember that there was once a standard regarding how close calibration rides needed to be to each other, so I went back to some of my old manuals. There it was, in the original 1985 manual, it stated that calibration rides should be within 0.07% of each other. That verbiage was missing from later manuals.
I've pretty much always held to that standard. I generally only calibrate over 300 meters now, but I always make sure I have four rides, including two in each direction, that are within 2 counts of each other. I adjust that on the rare occasion that I use a different length course. If you can't ride four rides within two counts, something is wrong and you need to attend to it.
We were eventually able to make things work using average constants, but I learned a lesson. Always make sure your calibration numbers agree with each other. If they don't agree, keep riding the calibration course until they do, whether the manual says you have to or not. There are a lot of things that can make two measurement rides not agree with each other. Don't make inconsistent calibrating be one of them.
Original Post
