Using two calibration courses to apply for certification

This is an old question. It seems to be more emotional than technical. 1000-ft = 1000-ft regardless of where the 1000-ft is. It seems that by doubting this, we are doubting our process. Is there a technical reason that different calibration courses shouldn't be used? What is the error introduced by the use of a second calibration course?

We have used 2 calibration courses on very rare occasions when measuring a race course far from home and far from the nearest calibration course.

This would be an advantage with a point to point race.
FYI - If I remember correctly we had several for the 1984 Olympic Marathon measurement

Using different calibration courses for the two measurements of a course would be a good check to make sure both cal courses are accurate. We should require that two different cal courses be used! Just kidding.

I have established a number of calibration courses in the Vancouver area and have used 2 calibration courses on occasion when the logistics saves me time and effort. I am with the Guido Brothers in that not allowing the use of 2 calibration courses suggests that we don't think they are the same and therefore not accurate.

I agree with Guido on that point. And along the lines of Mark's suggestion, it does seem that using two different cal courses gives an extra measure of certainty about the measurements, similar to the certainty from more measurements and measurements by different people.

I heard the same thing about the 1984 Olympic Marathon measurement, that there were several calibration courses-- as I understand it, they were along the way and to be recorded as encountered (I wasn't there though).

If I use my calibration course for pre-calibration, measure, then find my clicks/mile are more, based on someone else's course for my post-measurement calibration, do I then remeasure, or go back to my course for my post-calibration?

If I was smart enough to put down pieces of tape to get a reference measurement where I began the course measurement, and those pieces of tape tell me my tires are still inflated properly, then I have to question the post-cal course, right? Then, I would need to measure the cal course to see what I found.

I have no problem using a second cal course, just pointing out what could happen if you get a bad post-cal using that course.

Maybe this is why we should encourage measurers to use a variety of cal courses, when it is convenient enough. If there are cases like that, better to have them corrected than to just have each measurer in a separate world as determined by a possibly flawed cal course.

I think a more common reason for a "bad" post-cal (a high reading that you suspect might not be appropriate) is that a measurement is completed, then you travel a long way to recalibrate on a favorite close-to-home course, then find that a cold front has moved in and your counts go up.

Very constructive question, Gene.

I believe multiple courses are fine so long as we know we have a good basis to have confidence in the veracity of each of the calibration courses used.

What I'd like to encourage is the explicit recording of the cal course (or courses) used preserved in the historical record of each measure. I'd put this info on the cert itself and add a field to the database lookup to provide better data recall and technical transparency for each measure. This would provide a means to expediently deal with cal course problems as soon as they are discovered. We do not have that capability today.

We had a related question here in western NY a few years ago. State certifier Jim Gilmer wisely encouraged a process where we verify every Buffalo area cal course. In that initial effort we examined about twelve courses.

We found one cal course had a trivial variance and one popular cal course had a significant shortfall (about 18 inches in 330 meters if I remember correctly). The offending cal courses were carefully steel-taped and temp corrected and re-certified.

To this day we try to verify every new cal course that is produced in our region. I've done two verifications in the past two weeks. That gives me confidence that if I were to use 2 cal courses in a measure the results would be fine (verifiable) since the underlying cal courses have been checked in advance and we have good reason to believe that each cal course will produce the exact same working constant under similar temperature/tire pressure conditions.

My suggestion: Include a field on the cert form for recording the cal course #(s) utilized in a course measure. Then, if in the future a cal course is found to be defective we can quickly identify every course that may have been adversely affected and take immediate action as appropriate. Today, there is no expedient way to determine which certified courses may be faulty due to a bad cal course.

Any thoughts on this suggestion?

JJ

While it might be OK to recommend using the same cal course for pre- and post- I would not like to see it required. I have had occasions where I'll pre-cal on my home course, then use one nearer the race course for post if I know I won't be home before there's a significant temperature change (e.g. measure in the a.m., then go to work and not get home until after dark.)

Re: Jeff's comments: When I first started measuring David Reik was the CT certifier and required that a copy of the cal course accompany every application. Don't know if he recorded that data anywher; in my dim and distant memory I think there might have been a space on the application form asking for the cal course number - that could easily be reinstituted.

One other caveat: I would not be completely opposed to limiting different cal courses to ones that measurer had created; I've gotten into trouble using one of the Guido Brothers' 1,000- foot layouts thinking it was one of my 300m courses (thus disproving Pete's statement "1000-ft = 1000-ft regardless of where the 1000-ft is.")

I STRONGLY disagree with requiring use of just one cal course since I routinely use two.

For safety and efficiency, I measure at daybreak when I can see and traffic is minimal. My home cal course is a mile away, has good city lighting and so is easy to use before daylight. If there is a problem during calibration, there's a spare "whatever" a mile away. Then, ready to ride, I drive in the dark to the course to be measured. A distant cal course could be unusable for any number of reasons.

Recalibrating near the course allows you to determine the final S / F, Turnaround & Mile locations before installing permanent markers. With the numbers finalized, you can pound nails, take pictures or eat with peace of mind. Too many things can go wrong if you wait till you get home: falling temps late in the day causing higher Constant for the Day, unknown construction making distant course unusable, parked vehicles, slow leak, stolen bike, etc.

Notice that NE TN and SW VA have a BUNCH of certified cal courses.

Oscar Wagner

Here's a proposal for eliminating the post-cal in a large majority of measurements.

After doing a pre-cal at home drive to the course. Ride in a straight line for 300m (based on your pre-cal constant), marking the endpoints with a spot of paint. After measuring your course, ride the same straight line in the same direction.

1) If the counts for this post-measurement ride are more than 2 smaller than for the pre-measurement ride, then you are done. No post-cal needed.

2) If the counts are smaller by 2 or less, then go home and do your post-cal as usual.

3) If the counts are larger, then adjust the course based on the percent difference (and maybe add just a little to be sure) and then go home and post-cal as usual.

At least 90% of my measurements would fall into category #1.

I first saw this idea of doing a post-measurement calibration check at the course from Duane Russell (others have probably thought about it too). But I think we should take this a step further to eliminate the full post-cal.

I propose this with a grain of salt because I know it is unlikely to be accepted. But it should. We waste a lot of time doing post-cal rides when it is painfully obvious after the first ride that there is no way in h*ll the post-cal constant is going to be larger than the pre-cal.

Mark: AMEN.

A few years ago, I finished a course measurement moments before an intense summer afternoon thunderstorm blew through the area. I barely got my bike back on the car before the deluge and high winds arrived. I sat out the storm for about 40 minutes, as I recall, at the nearby cal course. As the skies began to clear, I got set to post-cal and I noticed that the temperature had dropped more than 20 degrees from the end time of the course measurement.

When I calculated the average constant, I could see that I would have to move the Start or the Finish a non-trivial amount to lengthen the course. I do not remember how many feet. I decided to look at the pre-and post-cal numbers carefully and decide how much the difference was likely caused by the temperature drop. Based on nothing more than experience and intuition, I concluded that the difference was 100% due to the change in the average pre-cal and course measurement temperature under a hot summer sun vs. a cool late afternoon post-storm post-cal. I therefore "violated" our guidelines by throwing out the post-cal because I saw no good reason to lengthen a 5 mile course whose two measurements varied by about 12 counts just because the air temperature dropped rapidly and significantly due to a passing storm.

Mark, it's an interesting idea and I'm not as opposed to it as I was when I first read it. I've been using this idea from Duane for a little while now and I think it's very helpful-- but I've thought of it as a predictor of what I would find out when I recalibrate.

I usually like to use a composite measurement (sum of shortest segments), and the average of my pre- and post-calibration constants. Your proposal wouldn't work for that (average should be based on two full calibrations), but maybe can work when you're using the higher of the constants.

Still, in my mind the notion of sandwiching any measurement between two full calibrations is a basic tenet of our work, and I'm a bit afraid of a "slippery slope" on this. Maybe we'd have to come up with further specs on how this is to be done, whether it has to be 300 m, etc, etc.

I understand what Lyman is saying and I've had a number of those cases. I agree that using Duane's procedure could provide a lot more confidence in making a decision about the final adjustment on a course. But I still feel the final calibration should be done. (Maybe not if weather isn't permitting!)

Bob I've just had so many times where I have done my first post-cal ride and it is 5 counts under my pre-cal average. And I think to myself, my next 3 rides would have to each be more than 6 counts over my pre-cal average (after the first ride was 5 counts under) for this to have any effect on the measurement. That's not going to happen, EVER.

I think one post-cal ride that CLEARLY indicates your wheel is bigger should be enough.

I vote with Bob on "slippery slope". Using judgement is fine for you guys with a 100+ course perspective but that is also about the sample size / experience level required to make those kinds of judgements.