Skip to main content

NON-FLAT CALIBRATION COURSES

A question arose when a certificate for a calibration course with a drop of 15 m/km was received. Should such a cal course be certified or used?

The ideal calibration course is straight, flat, at least 300 m in length, and close to the course to be measured. In many cases this ideal is hard to achieve. There’s no wiggle room on the length nor on straightness, but how much elevation change is OK? This is not rigidly specified in our procedures. Should it be?

In the group validation measurement of the 1996 Olympic marathon course, the 480 m calibration course had a drop of 4.2 m/km, and had a hill in the middle which was 9 m higher than the low end of the course.

In the group validation of the 2003 Olympic Trials marathon in Birmingham, Alabama, the pre-measurement 300 meter calibration course had a large rise (or drop) of 39 m/km. The postcal was flat.

The above courses were chosen for their proximity to the start of the measurement. It would have been time-consuming and difficult to locate an ideal precal course that did not require mass transport of the bicycles.

In spite of the hills, calibration rides did not show much difference. In Birmingham the typical ride took about 3300 counts. Uphill counts were generally less than downhill, but not in all cases. Moreover, the difference between uphill and downhill was generally on the order of 2 to 4 counts.

My feeling on this is that this is not a problem requiring a solution. Could be that any cure would be worse than the disease.

As long as the cal course is ridden in alternating directions, the errors should cancel out. Perfectly? No. Good enough? I think so.

Also, any windy day will produce count differences depending on whether the measurer is riding upwind wind or downwind.

Commentary?
Original Post

Replies sorted oldest to newest

I agree that it should average out fairly well, although not perfectly.

I had an interesting experience recently when I was making an adjustment to the Vancouver Half Marathon course where a new road alignment had shortenend the route between mile 8 and mile 9 by 111 m. About half of the distance between mile 8 and mile 9 was uphill with a total elevation gain of about 30 m. I ended up measuring between mile 8 and mile 9 four times, because I got such a big difference between uphill and downhill. The differences between my uphill rides and my downhill rides were both 2.9m or 0.2%. My 2 uphill rides and my 2 downhill rides were both within 3 counts when comparing uphill to uphill and downhill to downhill. Of course I used the shorter of the two differences to adjust the course, but I wonder if it wouldn't have been more accurate to use the shorter of 2 averages; ie average the first uphill and first downhill and then average the second uphill and second downhill.
While it seems intuitive that an uphill cal ride should show more counts, and a downhill ride fewer counts, than a flat cal ride, there is virtually no actual data (that I'm aware of) that supports that assumption.
It also seems intuitive, although somewhat less so, that averaging uphill and downhill rides should cancel out the error caused by each, but again, there is almost no data to support that.

2-4 counts in 3300 is similar in size to the SCPF.
I can see that the uphill and downhill rides would balance each other. How about a Cal course containing some up and down sections? Would you be able to tape it properly?

I feel that the Calibration course can be accepted if it's flat or has a constant increase/decrease in elevation. A rolling Calibration Course is a No!
Jim, when I read this thread initially, that's the only thing that occurred to me that would make it a bad course. If the elevation change is such that the tape is pulled off the ground when tension is applied, you're going to get an artificially long calibration course.

So that would be my litmus test--if the tape lifts off the ground, it's not usable.

Add Reply

Post
×
×
×
×
Link copied to your clipboard.
×