Vasaloppet adventure

Try here, Peter:
http://www.fis-ski.com/pdf/2006/CC/2...06CC2845RL.pdf

"Peter H." wrote:

That 1380m of climbing just kind of bothered me, so I took
a look at the course profile and tried to remember from about
20 years ago when I skied it. I do realize the profile won't show
climbs of 10 or 15 meters.
The horizontal lines are 100, 200, etc. meters.
It sure looks a lot closer to 380 m. total than 1380.

The Eurosport website is the only place I've seen the figure,
and I suppose web sites rank below newspapers
(and those somewhat below some
sports scientists) for credibility.
So it would be interesting to know whether that 1380
figure is given somewhere more believable.
If it's anywhere close to that much climb,
then I was a lot better skier in those days than I recall.

As my friend John B. has said:
"The older I get, the better I used to be!"

Best, Peter

Thanks very much, Gene. I cannot really argue
with the official FIS number for elevation gain.
Still, it's hard to believe that I
shuffled down that (not quite) 90km. in 8 hours
with a bad cold, very poor technique then, no idea
about proper training and nutrition, but still didn't
feel all that wasted, if it really was that much total
climbing (20 years ago in the Oppet Spar). We did have
a small tailwind, and the snow was plenty fast enough,
so maybe that's why, but I'd done the Finlandia 70km
just 4 days before.

The FIS gives 90,000 meters as the distance, so apparently
they tolerate some error, since the distance was listed as 89km
for several years some time ago.
But maybe the course has changed just enough
to be the full 90, which sounds more impressive. And they put
in a height difference, which usually implies that larger is harder,
but here it's the opposite with the finish lower than the start.

However, my (now lessened) skepticism about the elevation
gain had to do with a much bigger possible error than referred to
above. Looking carefully at all the little and bigger bumps on
a course profile, then adding 20 or 30%, seems to give a pretty
good estimate for most courses, in my experience, in terms of
what is stated as the official height gain. But here (Vasa), that
method certainly gives under 1000m. elevation gain.

With his experience in orienteering,
Terje M. probably knows at least as much as anyone here about
the accuracy of GPS's and altimeters, so here's a question for him
(and anybody else). How do they typically get these numbers for
elevation gain? Whatever else is true, it would be a lot less work
for a typical WC, even 50km., than for the Vasaloppet; i.e.
surveying a 16.7km loop, as opposed to a 90km point-to-point.

I notice that the expensive GPS's often have a separate altimeter,
though in principle, the GPS gives your position in the universe,
not just 2-dimensional position on the face of the earth. Not-so-fancy
GPS's like the Timex system will accumulate erroneous total distance
pretty quickly when you turn it on and just stay in the same place,
presumably 1 meter at a time due to slight observational errors
accumulating. So the GPS positioning system itself perhaps wouldn't
do the job very well. Even if a really high-end one
got to be accurate to a few centimeters
(but it's about 100 times less accurate, I understand),
just jiggling around on the skier's body would give all
sorts of little errors which could accumulate significantly.
For example, you could get 4500ft. of spurious
elevation gain over 90km. by just adding 1 foot per
20 meters of the skier's upper arm moving up and down,
if that's where the unit was worn.

This parameter (elevation gain) really does make a huge
difference in course difficulty, so it's very interesting to me
to learn how it's measured. Joseph might be interested to hear
that we figured about 3300ft. for the Birken a few years ago. It's
an easier estimate, with very few small undulations, just really
three very long climbs. Still very hard to believe that Vasaloppet
is more, despite the extra distance.

Best, Peter

Peter H. wrote:
Thanks very much, Gene. I cannot really argue
with the official FIS number for elevation gain.
Still, it's hard to believe that I
shuffled down that (not quite) 90km. in 8 hours
with a bad cold, very poor technique then, no idea
about proper training and nutrition, but still didn't
feel all that wasted, if it really was that much total
climbing (20 years ago in the Oppet Spar). We did have
a small tailwind, and the snow was plenty fast enough,
so maybe that's why, but I'd done the Finlandia 70km
just 4 days before.

The FIS gives 90,000 meters as the distance, so apparently
they tolerate some error, since the distance was listed as 89km
for several years some time ago.
But maybe the course has changed just enough
to be the full 90, which sounds more impressive. And they put
in a height difference, which usually implies that larger is harder,
but here it's the opposite with the finish lower than the start.

However, my (now lessened) skepticism about the elevation
gain had to do with a much bigger possible error than referred to
above. Looking carefully at all the little and bigger bumps on
a course profile, then adding 20 or 30%, seems to give a pretty
good estimate for most courses, in my experience, in terms of
what is stated as the official height gain. But here (Vasa), that
method certainly gives under 1000m. elevation gain.

With his experience in orienteering,
Terje M. probably knows at least as much as anyone here about
the accuracy of GPS's and altimeters, so here's a question for him
(and anybody else). How do they typically get these numbers for
elevation gain? Whatever else is true, it would be a lot less work
for a typical WC, even 50km., than for the Vasaloppet; i.e.
surveying a 16.7km loop, as opposed to a 90km point-to-point.

OK, here goes!

GPS is getting incredibly accurate these days, as long as you are in an
area with good visibility and WAAS/EGNOS/ESTB coverage.

Last summer I did a lot of testing over a 2-week period, using custom sw
I wrote to download binary data from my Garmin GPSmap 76S:

Combining all the thousands of position fixes, and assuming that the
average location is/was correct, I got 67% (i.e. approximately 1 sigma)
within 1.1 m, 90% within 2.0 m, 95% within 2.65 m.

Over the entire test period, I found a single bad period, lasting less
than 30 seconds, in which the position moved out about 30 m, stayed
there for a few seconds, then quickly returned. Excepting this one
anomalous measurement, all the remaining samples were within 3.0 m!

This takes care of X/Y (lat/long) coordinates, elevation is a very
different problem!

First is sat geometry: All the satellites are above you! This means that
average vertical precision is twice as bad as the horizontal one, which
means that over an hour or so you can get a _lot_ of elevation ripples,
giving false total elevation numbers.

For measuring the length of a ski trail, in open terrain, simply
carrying a GPS in the top pocket of a backpack will give you a result
that's probably accurate at the 10-30 m level over 10+ K. The less
twists and turns, the more accurate.

A trail that moves into tree cover, particularly wet/snowy trees,
reduces accuracy a _lot_, I have tried many times to carry a GPS in a
night-O battery carrier with little luck: The resulting track log is
never very close to my actual route choice.

Elevation profiles

By far the best way to get a good 3D track log is by combining two
devices: A GPS for horizontal position and a barometer for vertical changes.

With the default setup of my Garmin, the barometer is continuously
delivering altitude measurements, while at the same time the GPS is used
with a long (3 hour) time constant to calibrate the barometer. The
result is dependable altitude measurements, with particularly good
altitude changes which is the most important part.

Terje

--
-
"almost all programming can be viewed as an exercise in caching"

Terje Mathisen wrote:

OK, here goes!

GPS is getting incredibly accurate these days, as long as you are in
an area with good visibility and WAAS/EGNOS/ESTB coverage.

Last summer I did a lot of testing over a 2-week period, using custom
sw I wrote to download binary data from my Garmin GPSmap 76S:

Combining all the thousands of position fixes, and assuming that the
average location is/was correct, I got 67% (i.e. approximately 1
sigma) within 1.1 m, 90% within 2.0 m, 95% within 2.65 m.

Over the entire test period, I found a single bad period, lasting less
than 30 seconds, in which the position moved out about 30 m, stayed
there for a few seconds, then quickly returned. Excepting this one
anomalous measurement, all the remaining samples were within 3.0 m!

This takes care of X/Y (lat/long) coordinates, elevation is a very
different problem!

First is sat geometry: All the satellites are above you! This means
that average vertical precision is twice as bad as the horizontal
one, which means that over an hour or so you can get a _lot_ of
elevation ripples, giving false total elevation numbers.

For measuring the length of a ski trail, in open terrain, simply
carrying a GPS in the top pocket of a backpack will give you a result
that's probably accurate at the 10-30 m level over 10+ K. The less
twists and turns, the more accurate.

A trail that moves into tree cover, particularly wet/snowy trees,
reduces accuracy a _lot_, I have tried many times to carry a GPS in a
night-O battery carrier with little luck: The resulting track log is
never very close to my actual route choice.

How about power lines and their influence on distance and height? I use a
Forerunner 301, and it seems that where it is power lines, the measurements
are not accurate..

--
Terje Henriksen
Kirkenes