measuring xc courses?

So why is it that I can get the widest disparity of measurements with my Jones Counter in comparison to my bike computer on the same bike? My GPS is also much more consistant than the Jones Counter in all of my tests.

So you really think you can ride a bike at various speeds while trying to keep the front wheel on a lane line and occassionally glance at the click count with no deviation from that lane line? In reality it is much easier to do that with a wheel.

If a country mile is 16 FEET you are correct about my liberal allowances of how others have measured. I really don't believe that I'm ever off that much with any method I use.

Also, riding a bike on CC Courses is not possible in my area on many of the courses.

dingle wrote:

Bobby1 wrote:

It's an interesting study but fails to consider the possibility of randomization of errors and that most would cancel out over a longer course.

This is highly unlikely. If you have an error ridden process, as you repeat it the average squared deviation from the real value will probably increase, even if the expected value of the measurement remains the true value.

Are you sure you meant what you wrote? Are you saying that all of the errors will be of exactly the same direction and that it is highly unlikely the correlation is less than one? How many courses have you measured and have you obtained the accuracy of the people discussing this above. How many statistics and probability classes have you done very well in. Explain how the average squared deviation is going to increase; is each subsequent error going to be not only in the same direction [the highly unlikely part of your post] but sequentially larger; your tape measuring device is flawed and the flaw is getting bigger?

When laying out a track the tolerances have to be very small because longer distances repeat the same errors while on the road, the mistakes are not all in the same direction unless it is in your basic measuring device (e.g., you tape measure is 99.9m not 100.0m). However, even here the average squared deviation is not getting larger, it is staying the same. If you make errors not related to the measuring instrument bias, then the average error gets smaller.

I'm quite sure of what I meant and I won't engage in the ad hominem/ argument from authority arguments that you are.

The person I was replying to suggested that if you add a bunch of measurements up where each one is an unbiased but error prone measurement, that the errors would cancel out. The process described can be viewed as a Gaussian random walk the mean squared error mean (Distance measure - distance traveled)^2= sigma^2 n where sigma^2 is the squared error in an individual measurement and n is the number of measurements you make.

I think that you probably misunderstood the terms (how many stats classes have you taken?) I used. I didn't say that the error of each individual measurement changes, but I did say that the average measurement error at a given distance increases with the distance measured. (i.e. if you measure a 8k 100 times with a 100m tape, the average absolute error of that total measurement will be greater than if you measured a 5k with the same tape).

"We expect that random errors should cancel on average, that is, many measurements of the same quantity should produce a more reliable estimate. Statistical errors can be controlled by performing a sufficiently large number of measurements."

http://www.physics.smu.edu/~scalise/mechmanual/error/

Bobby1 wrote:

"We expect that random errors should cancel on average, that is, many measurements of the same quantity should produce a more reliable estimate. Statistical errors can be controlled by performing a sufficiently large number of measurements."

http://www.physics.smu.edu/~scalise/mechmanual/error/

That is the difference between measuring the same thing 100 times versus measuring one thing with 100 measurements. i.e. measuring a 100m course 100 times, versus measuring a 10k with 100 100m tape measurements.

So if you are arguing that people should take a crappy measuring device and measure their entire course 100 times, I agree. If you are suggesting that a crappy measuring device magically becomes a good one if you add all its measurements over the course, then you are wrong.

You seem to like to make assumptions, the measuring device is not crappy, there are random errors in any measurement and they can cancel each other out.

Bobby1 wrote:

You seem to like to make assumptions, the measuring device is not crappy, there are random errors in any measurement and they can cancel each other out.

Yes by crappy, I mean they have error (small or large). Regardless, they do not "cancel out" if you are adding them, they do "cancel out" so to speak if you are averaging them.

Do you still think that if you are measuring a 100m course the error is less than the error in a 10k course that you measure with the same 100m tape?

dukerdog wrote:

http://ocrs.wlu.edu/Images/whyOCRS.pdf

thats funny because w&l's course is 650 meters short

Look at the animation of a random walk here:

http://animation.yihui.name/prob:brownian_motion

note how on average, the balls move away from the center where they start. This is the sum of a bunch of random errors.

Admiral, The wrote:

dukerdog wrote:http://ocrs.wlu.edu/Images/whyOCRS.pdf

thats funny because w&l's course is 650 meters short

it gets even worse, their track is short too

This term "average squared deviation" is what causes the problem. It is easy to read this as the assertion that, essentially the standard deviation divided by the total distance does not decrease. You are much 'cleaner' (and hence it is clear more precisely what you meant) in your reply that indicates that your original post was and is correct; it was a little hard to parse (different backgrounds use words differently, mine comes from applied econometrics). Sorry for the 'umbrage' in my remarks.

I know Jones!!!!!

Because you're obviously doing it wrong.

And you're doing it wrong, because you think wrong is right and right is wrong, therefore wrong must be better than right.

I just ride the bike the same way that I'd run the course, taking the shortest route possible. No, I don't stare at the ground and I don't look at the count, which might explain why you're wobbling around like you do with the wheel. I take care that the speed of the bicycle is reasonable to maintain the integrity of the wheel with the line that I'm traveling and the ground. That means that I won't be keeping the same speed all the time, unless the course is exactly the same all the time, but that's usually not the case, even on the roads.

If you ride the course twice with a Jones Counter on a bike, and you ride the course the shortest distance each time, then you should get the exact same measurement, which should not be off more than a few inches each time. For these to be accurate, the bicycle needs to be calibrated both before, and after, the measurements. This is to verify that no significant change has occurred to the pressure of the tires, or anything else that would affect the calibration of the bike.

Walking or jogging with a wheel is not accurate because of the wobbling of the wheel, and there there is no weight on the wheel. When you are walking or jogging, your feet are not on the ground all the time, and that affects the motion of the wheel.

J.R. wrote:

Walking or jogging with a wheel is not accurate because of the wobbling of the wheel, and there there is no weight on the wheel. When you are walking or jogging, your feet are not on the ground all the time, and that affects the motion of the wheel.

So how come I can get 100m/400m/1600m/etc. when I use my wheel on the track when walking or running to check calibration?? Also, it is not difficult to apply light steady pressure on the handle of the wheel while walking or running with the wheel.

I really don't see how your measuring unit and techniques are so superior to other products, and your theories seem a little out there (your quote above?!?!).

Interesting.

Question-

Which hand are you holding the handle of the measuring wheel when you measure the track that you claim you get 100m/400m/1600m????

Do you switch hands at all?

In my early days of measuring (back in the 70's)I did a great deal of measuring with the wheel....even full maratons!

26mi235 wrote:

This term "average squared deviation" is what causes the problem. It is easy to read this as the assertion that, essentially the standard deviation divided by the total distance does not decrease. You are much 'cleaner' (and hence it is clear more precisely what you meant) in your reply that indicates that your original post was and is correct; it was a little hard to parse (different backgrounds use words differently, mine comes from applied econometrics). Sorry for the 'umbrage' in my remarks.

Damn, you economists screw everything up ;)

David -

The wheel I use most often has a small kick stand that I leave down and use as guide when following curves. I undestand the difference of a curbed and non curbed track, 12" vs 8" in lane one. I'm right handed and always use my right hand and when measuring/calibrating I use the kickstand following the lane line between lane one and lane two. This keeps the wheel 8" from the inner portion of the lane line. The wheel is always within a meter/1600m no matter if I walk or run.

How do you know that you actualy measured 1600 meters if you are in lane 2?

Your accurate within 25 cm per 400m regardless of running or walking!...Wow, that's good. If I could do that I would be considered one of the best measurers in the world...but I don't think I could. :)

David Katz wrote:

How do you know that you actualy measured 1600 meters if you are in lane 2?

Your accurate within 25 cm per 400m regardless of running or walking!...Wow, that's good. If I could do that I would be considered one of the best measurers in the world...but I don't think I could. :)

1600m Relay - work off that stagger... pretty simple