Sonora Pass ride description & photos up

"Dan" wrote
"Dan Connelly" wrote
Dan wrote:

clip

I'd argue the natural convolution parameter for climbs is the amount one
can
climb in on a load of ATP, which lasts around 10 seconds. For example,
OLH
is around 390 meters, so if you can climb OLH in 21:40, this corresponds

clip

Okay, I had to lookup ATP
http://en.wikipedia.org/wiki/Adenosine_triphosphate
I have no idea what OLH is.

clip

Probably Old La Honda, a hill on the Peninsula often used as a test of
climbing ability.

I personally prefer "stairway to heaven" over "the golden stairs to
heaven" as in... http://tinyurl.com/8pzvv... it more catchy. Of course,
I replaced Led Zepellin's reference to "buying" with "climbing".

Steve

Dan Connelly wrote:
Dan wrote:

I have added a climb per 1/2 mile to the plots linked above. It shows
steeper bits steeper but we are headed for the much contested "how to
measure steepness" discussion. I am sure we could find a millimeter
along the way that is dead vertical if not overhanging.

...

Commonly, people suggest discritzation as the way to this end (ie measing
th eroad every 15 meters, or half-mile, or whatever). This is an inferior
choice, because the result depends on where you start. For example, if
you're measuring every 15 meters the the road climbs from 0 to 15, then
descends
from 15 to 30, the result is profoundly different if you measure at 0,
15, and 30
then if you measure at -7.5, +7.5, +22.5, +37.5. It's way better to
oversample
then filter down with convolution.

Dan


This sounds a lot like Mandelbrot's famous question--"How long is the
coastline of Britain?" The answer is from something like zero to
approaching infinity--depends on the length of you measuring stick.
Landscapes including profiles of roads have fractal properties over
quite a range of scales.

Mike

Mike McGuire wrote:

Commonly, people suggest discritzation as the way to this end (ie measing
th eroad every 15 meters, or half-mile, or whatever). This is an
inferior
choice, because the result depends on where you start. For example, if
you're measuring every 15 meters the the road climbs from 0 to 15,
then descends
from 15 to 30, the result is profoundly different if you measure at 0,
15, and 30
then if you measure at -7.5, +7.5, +22.5, +37.5. It's way better to
oversample
then filter down with convolution.

Dan


This sounds a lot like Mandelbrot's famous question--"How long is the
coastline of Britain?" The answer is from something like zero to
approaching infinity--depends on the length of you measuring stick.
Landscapes including profiles of roads have fractal properties over
quite a range of scales.

Mike

Precisely.

Dan

Dan wrote:
"Dan Connelly" wrote in message
et...


Okay, I am getting in over my head here so cut me some slack if I am off
base. I gather from your post that you are suggesting that a Gaussian filter
be applied to the data set. My understanding is that this is a low pass
filter in the frequency domain. The shape of this filter is non-linear and
based on the normal distribution. Filtering and smoothing certainly make a
data set look nicer but if the original data set is accurate and contains no
noise, filtering and/or smoothing will create a false data set with error
introduced into virtually every data point (no post-filter data point
represents reality). I have seen otherwise knowledgeable people filter out
the interesting part of a data set because they did not like it, understand
it or it did not fit their preconceptions. Seeing the world through a
Gaussian blur can be rosy but you can also miss important detail. There is
also the aliasing problem inherent in discrete sampling but as you suggest,
over-sampling will usually take care of this.

IMHO, the best way to judge the Sonora Pass climb is to go do it. Do it many
times and try changing the variables.



The goal is to establish a "peak grade" for the hill which correlates with how hard it is
to climb. My point is, even a 100% grade over 1 meter is barely relevent, as inertia, not
to mention anaerobic metabolism, allows one to "power" over it (if you don't skid). The
low-pass filter is to match the natural low-pass filter of inertia and anaerobic
power. My guess is 15 meters is close to the right distance of this "natural" low-pass filter.

Dan

So I'd convolve the profile with a Gaussian of sigma 15 meters. This
renders
as fully irrelevent your 1 mm example, yet a half mile discrization
clearly
is inadequate at reproducing the feel of the road.

My head is spinning. All I did was visually inspect the readouts from the
HAC-4 (altitude, distance & percent grade... the latter used only as an
indication of where a steep part might have begun or ended, nothing more)
and look for sections approximately .6 miles or greater that showed fairly
high numbers for percent grade. Then, using the altitude & distance numbers,
calculate percent grade over that section.

Obviously this won't tell you what the absolute steepest grade is, but it's
very good at determining where it's going to hurt the most.

--Mike-- Chain Reaction Bicycles
www.ChainReactionBicycles.com


"Dan Connelly" wrote in message
et...
Dan wrote:

I have added a climb per 1/2 mile to the plots linked above. It shows
steeper bits steeper but we are headed for the much contested "how to
measure steepness" discussion. I am sure we could find a millimeter along
the way that is dead vertical if not overhanging.

I'd argue the natural convolution parameter for climbs is the amount one
can
climb in on a load of ATP, which lasts around 10 seconds. For example,
OLH
is around 390 meters, so if you can climb OLH in 21:40, this corresponds
to
18 vertical meters per minute, or 3 meters per 10 seconds. Using a 20%
grade
as a reference, this corresponds to a lateral length scale of 15 meters.
(okay, actually sqrt(15^2 + 3^2) = 15.3 meters). This result "feels"
right
to me, based on my history of mental distress at encountering short steep
sections of road.

So I'd convolve the profile with a Gaussian of sigma 15 meters. This
renders
as fully irrelevent your 1 mm example, yet a half mile discrization
clearly
is inadequate at reproducing the feel of the road.

Commonly, people suggest discritzation as the way to this end (ie measing
th eroad every 15 meters, or half-mile, or whatever). This is an
inferior
choice, because the result depends on where you start. For example, if
you're measuring every 15 meters the the road climbs from 0 to 15, then
descends
from 15 to 30, the result is profoundly different if you measure at 0, 15,
and 30
then if you measure at -7.5, +7.5, +22.5, +37.5. It's way better to
oversample
then filter down with convolution.

Dan

That's a really nice photo report. Excellent captions for each photo...
makes for a good story and resource!

Since moving to California, it's puzzled me when I see the word
"Sierra" pluralized as you wrote "It really is this pretty in the
Sierras." Isn't it redundant to pluralize Sierra since it means
mountains or mountain range in Spanish? Small issue... I'm just
curious.

great report,
Steve
http://www.steephill.tv

"steephill.tv" wrote in message
ps.com...
That's a really nice photo report. Excellent captions for each photo...
makes for a good story and resource!

Since moving to California, it's puzzled me when I see the word
"Sierra" pluralized as you wrote "It really is this pretty in the
Sierras." Isn't it redundant to pluralize Sierra since it means
mountains or mountain range in Spanish? Small issue... I'm just
curious.

great report,
Steve
http://www.steephill.tv


You are right Steve, it should be Sierra not Sierras. I suspect that few of
use know as much Spanish (or is it Castilian?) as we should. Just as I have
gotten used to "PIN number" I have become accustomed to "Sierras."

In one way it makes sense because the range varies a great deal. The Whitney
region with high granite peaks, The loose rock of the Kaweahs, The
Palisades, The Minarets, the Cathedral Range, the volcanic peaks of the east
side, the Tahoe peaks, etc. They are all unique sub ranges.

"steephill.tv" wrote in message
ps.com...
That's a really nice photo report. Excellent captions for each photo...
makes for a good story and resource!

Since moving to California, it's puzzled me when I see the word
"Sierra" pluralized as you wrote "It really is this pretty in the
Sierras." Isn't it redundant to pluralize Sierra since it means
mountains or mountain range in Spanish? Small issue... I'm just
curious.

Improper Spanish or not it has become common usage in English for the
mountains in the Sierra Nevada mountain range. Just like "Rockies" is used
for those in the Rocky Mountain range.

Speaking of redundancies: Morro Rock, or one the ways to get the Cuesta
Grade.

Don
--
Ever had one of those days where you just felt like:
http://cosmoslair.com/BadDay.html ?

"Don Freeman" wrote in message
...

"steephill.tv" wrote in message
ps.com...
That's a really nice photo report. Excellent captions for each photo...
makes for a good story and resource!

Since moving to California, it's puzzled me when I see the word
"Sierra" pluralized as you wrote "It really is this pretty in the
Sierras." Isn't it redundant to pluralize Sierra since it means
mountains or mountain range in Spanish? Small issue... I'm just
curious.

Improper Spanish or not it has become common usage in English for the
mountains in the Sierra Nevada mountain range. Just like "Rockies" is
used for those in the Rocky Mountain range.

Speaking of redundancies: Morro Rock, or one the ways to get the Cuesta
Grade.

Don
--
Ever had one of those days where you just felt like:
http://cosmoslair.com/BadDay.html ?




I have always liked "The La Brea Tar Pits" or "The The Tar Tar Pits."