Let's get real.

Robert Chung said:

Robert Chung said:

Robert Chung said:

Dan Connelly wrote:

>For OLH, compare with Cote Monte des Princes in stage 7, 6km @ 7%,
>or Col du Portet d'Aspet, 6.8% for 5.9km in stage 14, both cat 2,
>in the 2003 Tour. OLH averages 7.3% and gaining 393 meters
>in 5.4 km.
>
>You say Hwy 9 is 6.8% for 11km. In stage 15 this year,
>Col des Limouches is 10.7 km at 6.4%
>( http://www.cyclingnews.com/road/2004/tour04/?id=stages#summ ),
>and is cat 2.
>
> From 2003 (
>http://www.cyclingnews.com/road/2003/tour03/?id=stages#summ ), Col du
>Télégraphe is a 12.1 km climb @ 6.8 percent,
>also Cat. 2.
>
>So it seems Hwy 9 would also be cat 2.

http://anonymous.coward.free.fr/rbr/cols04.png
http://anonymous.coward.free.fr/rbr/cols04.csv

Ooops. Sorry. I've put up corrected data and plot.

Hmmm. I've just noticed that www.cyclingnews.com measures the climbs
differently than Velo Magazine. Jeff?

As usual, an excellent representation of the data from Robert. It makes sense to
add a few years to the data, but I should really do that.... I'll see if I have time.

A fundamental principle of rating hills should be that a subset of a climb have a lower
rating than the climb itself, or a superset. The problem with grade-distance is
that by taking a bigger chunk of road, one can convince oneself the climb is easier.
So the selection of endpoints becomes critical. Parameterization should really involve an
integration over a suitable discretization.

Dan

"Dan Connelly" <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote in message
news:[email hidden]...

Quoted message said:

Robert Chung said:

Robert Chung said:

Robert Chung wrote:

>Dan Connelly wrote:
>
>>For OLH, compare with Cote Monte des Princes in stage 7, 6km @ 7%,
>>or Col du Portet d'Aspet, 6.8% for 5.9km in stage 14, both cat 2,
>>in the 2003 Tour. OLH averages 7.3% and gaining 393 meters
>>in 5.4 km.
>>
>>You say Hwy 9 is 6.8% for 11km. In stage 15 this year,
>>Col des Limouches is 10.7 km at 6.4%
>>( http://www.cyclingnews.com/road/2004/tour04/?id=stages#summ ),
>>and is cat 2.
>>
>> From 2003 (
>>http://www.cyclingnews.com/road/2003/tour03/?id=stages#summ ), Col du
>>Télégraphe is a 12.1 km climb @ 6.8 percent,
>>also Cat. 2.
>>
>>So it seems Hwy 9 would also be cat 2.
>
>http://anonymous.coward.free.fr/rbr/cols04.png
>http://anonymous.coward.free.fr/rbr/cols04.csv

Ooops. Sorry. I've put up corrected data and plot.

Hmmm. I've just noticed that www.cyclingnews.com measures the climbs
differently than Velo Magazine. Jeff?

As usual, an excellent representation of the data from Robert. It makes

sense to

Quoted message said:

add a few years to the data, but I should really do that.... I'll see if

I have time.

Quoted message said:

A fundamental principle of rating hills should be that a subset of a climb

have a lower

Quoted message said:

rating than the climb itself, or a superset. The problem with

grade-distance is

Quoted message said:

that by taking a bigger chunk of road, one can convince oneself the climb

is easier.

Quoted message said:

So the selection of endpoints becomes critical. Parameterization should

really involve an

Quoted message said:

integration over a suitable discretization.

Dan

In order to categorize correctly, you need to factor in the preceding and
following hills on the stage. Categorization is not fixed for a given hill.

"Robert Chung" <[email hidden]> wrote in message news:<[email hidden]>...

Quoted message said:

Dan Connelly said:

For OLH, compare with Cote Monte des Princes in stage 7, 6km @ 7%,
or Col du Portet d'Aspet, 6.8% for 5.9km in stage 14, both cat 2,
in the 2003 Tour. OLH averages 7.3% and gaining 393 meters
in 5.4 km.

You say Hwy 9 is 6.8% for 11km. In stage 15 this year,
Col des Limouches is 10.7 km at 6.4%
( http://www.cyclingnews.com/road/2004/tour04/?id=stages#summ ),
and is cat 2.

From 2003 (
http://www.cyclingnews.com/road/2003/tour03/?id=stages#summ ), Col du
Télégraphe is a 12.1 km climb @ 6.8 percent,
also Cat. 2.

So it seems Hwy 9 would also be cat 2.

http://anonymous.coward.free.fr/rbr/cols04.png
http://anonymous.coward.free.fr/rbr/cols04.csv

I've read here in the past that there are few American roads that can
equal the climbs in the Alps. But looking at this listing of yours
suggests to me that the bay area's Mt. Diablo bears a striking
similarity to La Mongie or even L'Alpe d'Huez.

Tom Kunich said:

"Robert Chung" <[email hidden]> wrote in message news:<[email hidden]>...

Quoted message said:

Quoted message said:

http://anonymous.coward.free.fr/rbr/cols04.png
http://anonymous.coward.free.fr/rbr/cols04.csv

I've read here in the past that there are few American roads that can
equal the climbs in the Alps. But looking at this listing of yours
suggests to me that the bay area's Mt. Diablo bears a striking
similarity to La Mongie or even L'Alpe d'Huez.

La Mongie is cat 1, but L'Alpe is HC and is significantly steeper than
Mt Diablo.

To Carl's point that the position of the climb matters -- I completely
agree. I avoid comparisons to climbs which immediately follow other
climbs (like the cat 3 climb today, Col de Carri, which gains
152 meters!), or Villard de Lans, which gains 154 meters, either of
which would be cat 4 in different circumstances, I suspect.

Dan

"Tom Kunich" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Robert Chung" <[email hidden]> wrote in message

news:<[email hidden]>...

Quoted message said:

Quoted message said:

Dan Connelly said:

For OLH, compare with Cote Monte des Princes in stage 7, 6km @ 7%,
or Col du Portet d'Aspet, 6.8% for 5.9km in stage 14, both cat 2,
in the 2003 Tour. OLH averages 7.3% and gaining 393 meters
in 5.4 km.

You say Hwy 9 is 6.8% for 11km. In stage 15 this year,
Col des Limouches is 10.7 km at 6.4%
( http://www.cyclingnews.com/road/2004/tour04/?id=stages#summ ),
and is cat 2.

From 2003 (
http://www.cyclingnews.com/road/2003/tour03/?id=stages#summ ), Col du
Télégraphe is a 12.1 km climb @ 6.8 percent,
also Cat. 2.

So it seems Hwy 9 would also be cat 2.

http://anonymous.coward.free.fr/rbr/cols04.png
http://anonymous.coward.free.fr/rbr/cols04.csv

I've read here in the past that there are few American roads that can
equal the climbs in the Alps. But looking at this listing of yours
suggests to me that the bay area's Mt. Diablo bears a striking
similarity to La Mongie or even L'Alpe d'Huez.

Without crunching the numbers there are probably quite a few here in the
Sierras that would come to mind. I haven't rode them all, but the back side
of Sonora Pass comes to mind. Probably the climb from Lemon Cove or Three
Rivers to the top into Sequoia park. For a short thrill, Old Priest grade,
the climb from Bass lake to the where the main road turns off for Shuteye
peak, the climb from El Portal to the turnoff to Tioga pass, one of my
favorites, maybe the back side of Ebbets pass, the climb out of Death Valley
from Stove Pipes Wells, eh... that's enough. You might throw in the climb
from Lee Vining to the top of Tioga pass, a grinder.

Dan Connelly said:

To Carl's point that the position of the climb matters -- I completely
agree. I avoid comparisons to climbs which immediately follow other
climbs (like the cat 3 climb today, Col de Carri, which gains
152 meters!), or Villard de Lans, which gains 154 meters, either of
which would be cat 4 in different circumstances, I suspect.

Yes, the cat 2 rating given to the final climb out of Villard de Lans up
the Cote 2000 got hugely inflated because it ends a stage. Likewise, La
Mongie is only used as the end of a stage (else they'd continue up and
over the top of the Tourmalet); if you thought of it as a stand-alone
climb it'd probably be on the borderline.
Limouches and the Col d'Aspin are almost identical, but Aspin came toward
the end of the stage while Limouches was the first climb today.

Still, I was startled that the grouping seemed relatively clean.

I'm still distressed about the endpoint disparity between cyclingnews.com
and Velo Magazine (which is published by L'Equipe).

Sierraman said:

For a short thrill, Old Priest grade

Isn't there too much car traffic on this road for it to be enjoyable?

On Sun, 18 Jul 2004 23:14:12 GMT, "Callistus Valerius" <[email hidden]>

Quoted message said:

Let's get real, Voeckler would get dropped on the Sunday group rides where I
live. What the hell is this guy doing in yellow???

Then why isn't your club in the Tour?
--
Kelly Beard

"Dan Connelly" <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote in message
news:40FD4DDF.5070002@i_e_e_e.o_r_g...

Quoted message said:

Sierraman said:

For a short thrill, Old Priest grade

Isn't there too much car traffic on this road for it to be enjoyable?

Old Priest is a trap. Locals know to stay off, it's the tourists that get
stuck. There is an inn at the top and the guy who runs it has told me lots
of stories about this road. I used to climb it on a dare with friends. It
rises about 2000 feet in 2 miles. It nuts! With cars usually the brakes go
out and off the cliff they go. Ambulances used to hang out in that area.
Serious, that's just how many calls they used to get. So on the whole with
the New Priest grade being the new stretch of highway, I would say the
traffic is very light on the old grade. Everytime I went I rarely passed a
car or two who were wondering what the hell they had got themselves into.
Grade signs don't seem to make people think twice. Another time I climbed it
once on a dare usng a old very heavy schwinn to prove how strong I was. I'll
never do that again! I let one of the guys ride my very light roadbike and I
still beat him.

Sierraman said:

"Dan Connelly" <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote in message
news:40FD4DDF.5070002@i_e_e_e.o_r_g...

Quoted message said:

Sierraman said:

For a short thrill, Old Priest grade

Isn't there too much car traffic on this road for it to be enjoyable?

Old Priest is a trap. Locals know to stay off, it's the tourists that get
stuck. There is an inn at the top and the guy who runs it has told me lots
of stories about this road. I used to climb it on a dare with friends. It
rises about 2000 feet in 2 miles. It nuts! With cars usually the brakes go
out and off the cliff they go. Ambulances used to hang out in that area.
Serious, that's just how many calls they used to get. So on the whole with
the New Priest grade being the new stretch of highway, I would say the
traffic is very light on the old grade. Everytime I went I rarely passed a
car or two who were wondering what the hell they had got themselves into.
Grade signs don't seem to make people think twice. Another time I climbed it
once on a dare usng a old very heavy schwinn to prove how strong I was. I'll
never do that again! I let one of the guys ride my very light roadbike and I
still beat him.

For interested readers, the top of Old Priest Grade is here:
http://www.topozone.com/map.asp?z=10&n=4189461&e=738801&s=50&size=l&datum=nad83&layer=DRG25

Dan

"Sierraman" <[email hidden]> wrote in message

Quoted message said:

Old Priest is a trap. Locals know to stay off, it's the tourists that

get

Quoted message said:

stuck. There is an inn at the top and the guy who runs it has told me lots
of stories about this road. I used to climb it on a dare with friends. It
rises about 2000 feet in 2 miles. It nuts! With cars usually the brakes go
out and off the cliff they go. Ambulances used to hang out in that area.
Serious, that's just how many calls they used to get.

19% for average for 2 miles?

"Carl Sundquist" <[email hidden]> wrote in message news:<[email hidden]>...

Quoted message said:

"Dan Connelly" <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote in message

Quoted message said:

A fundamental principle of rating hills should be that a subset of a climb

have a lower

Quoted message said:

rating than the climb itself, or a superset. The problem with

grade-distance is

Quoted message said:

that by taking a bigger chunk of road, one can convince oneself the climb

is easier.

Quoted message said:

So the selection of endpoints becomes critical. Parameterization should

really involve an

Quoted message said:

integration over a suitable discretization.

Dan

In order to categorize correctly, you need to factor in the preceding and
following hills on the stage. Categorization is not fixed for a given hill.

That is correct. The TDF categorisation is based on the gear used by a car
to climb the hill, so it changed this year because Skodas have taller
gearing than Fiat.

-ilan

"Carl Sundquist" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Sierraman" <[email hidden]> wrote in message

Quoted message said:

Old Priest is a trap. Locals know to stay off, it's the tourists that

get

Quoted message said:

stuck. There is an inn at the top and the guy who runs it has told me

lots

Quoted message said:

Quoted message said:

of stories about this road. I used to climb it on a dare with friends.

It

Quoted message said:

Quoted message said:

rises about 2000 feet in 2 miles. It nuts! With cars usually the brakes

go

Quoted message said:

Quoted message said:

out and off the cliff they go. Ambulances used to hang out in that area.
Serious, that's just how many calls they used to get.

19% for average for 2 miles?

No, I exaggerated some. I think the grade sign says 18, not sure. I thought
actually it was more like 1800 feet in two miles. Dan might know exactly.

Dan Connelly d_j_c_o_n_n_e_l@i_e_e_e.o_r_g said:

Quoted message said:

Quoted message said:

Robert Chung wrote:
>http://anonymous.coward.free.fr/rbr/cols04.png
>http://anonymous.coward.free.fr/rbr/cols04.csv

Quoted message said:

A fundamental principle of rating hills should be that a subset of a climb have a lower
rating than the climb itself, or a superset. The problem with grade-distance is
that by taking a bigger chunk of road, one can convince oneself the climb is easier.
So the selection of endpoints becomes critical. Parameterization should really involve an
integration over a suitable discretization.

Not really a problem. Just measure the cumulative elevation gain, it's
robust. Doesn't matter as much where the climb "starts."
Using Robert's Velo mag data:

http://www.ucolick.org/~bjw/misc/bikes/cols04.l.grade.png

http://www.ucolick.org/~bjw/misc/bikes/cols04.climb.cat.png

Benjamin Weiner said:

Dan Connelly <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote:

Not really a problem. Just measure the cumulative elevation gain, it's
robust. Doesn't matter as much where the climb "starts."
Using Robert's Velo mag data:

http://www.ucolick.org/~bjw/misc/bikes/cols04.l.grade.png

http://www.ucolick.org/~bjw/misc/bikes/cols04.climb.cat.png

But cumulative fails to reward steepness. Instead of integrating
grade * dx, better to integrate grade^2 * dx. But then this
gives excessive reward for steepness at shallow grades. Still,
it's an improvement, and is essentially what is done
at:

http://www.altimetrias.com/htm/ComoCoefUK.htm

Dan

Benjamin Weiner said:

Dan Connelly d_j_c_o_n_n_e_l@i_e_e_e.o_r_g said:

A fundamental principle of rating hills should be that a subset of a
climb have a lower rating than the climb itself, or a superset. The
problem with grade-distance is
that by taking a bigger chunk of road, one can convince oneself the
climb is easier.
So the selection of endpoints becomes critical. Parameterization
should really involve an integration over a suitable discretization.

Not really a problem. Just measure the cumulative elevation gain, it's
robust. Doesn't matter as much where the climb "starts."
Using Robert's Velo mag data:

http://www.ucolick.org/~bjw/misc/bikes/cols04.l.grade.png
http://www.ucolick.org/~bjw/misc/bikes/cols04.climb.cat.png

I think it's still kind of a problem.

Based on data from Velo Magazine (and thus, L'Equipe)
http://anonymous.coward.free.fr/rbr/cols04.png

Based on data from cyclingnews.com
http://anonymous.coward.free.fr/rbr/cols04-cn.png

Robert Chung said:

Benjamin Weiner said:

Dan Connelly <d_j_c_o_n_n_e_l@i_e_e_e.o_r_g> wrote:

Quoted message said:

Quoted message said:

Quoted message said:

So the selection of endpoints becomes critical. Parameterization
should really involve an integration over a suitable discretization.

Quoted message said:

Quoted message said:

Not really a problem. Just measure the cumulative elevation gain, it's
robust. Doesn't matter as much where the climb "starts."
Using Robert's Velo mag data:

Quoted message said:

I think it's still kind of a problem.

I just meant that it bypasses Dan's problem about endpoints and integration.

Quoted message said:

Based on data from Velo Magazine (and thus, L'Equipe)
http://anonymous.coward.free.fr/rbr/cols04.png

Based on data from cyclingnews.com
http://anonymous.coward.free.fr/rbr/cols04-cn.png

Worrisome, yes. Anybody got a topo map? But for the purposes of
categorizing a climb, it doesn't matter what the actual elevation change
is. It only matters what ASO _thinks_ the elevation change is.

Benjamin Weiner said:

But for the purposes of
categorizing a climb, it doesn't matter what the actual elevation change
is. It only matters what ASO _thinks_ the elevation change is.

I had looked for that info in my Programme Officiel du Tour 2004 but
couldn't find it -- hence, the Velo Mag and cyclingnews.com data. However,
prompted by your words, I looked through www.letour.fr until I found the
data as ASO views them:

http://anonymous.coward.free.fr/rbr/cols04-tdf.png
http://anonymous.coward.free.fr/rbr/cols04-tdf.csv

Robert Chung said:

Benjamin Weiner wrote:

Quoted message said:

I had looked for that info in my Programme Officiel du Tour 2004 but
couldn't find it -- hence, the Velo Mag and cyclingnews.com data. However,
prompted by your words, I looked through www.letour.fr until I found the
data as ASO views them:

http://anonymous.coward.free.fr/rbr/cols04-tdf.png
http://anonymous.coward.free.fr/rbr/cols04-tdf.csv

Excellent, thank you.

http://www.ucolick.org/~bjw/misc/bikes/cols04.logl.grade.png

http://www.ucolick.org/~bjw/misc/bikes/cols04.logclimb.cat.png

The first rule of statistics: Every plot looks cleaner on a log scale.

A few of the obvious outliers are stage finishes (Villard-de-Lans)
or climbs toward the end of a long Alpine day (Forclaz). The Neronne
and Peyrol in the middle of stage 10 don't have that excuse for ranking
Cat. 2 and 1, though it was a beastly long stage. The RBR collective
will have to decide if Virenque's MP should be revoked.

Sierraman said:

"Carl Sundquist" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Sierraman" <[email hidden]> wrote in message

Quoted message said:

Old Priest is a trap. Locals know to stay off, it's the tourists that

get

Quoted message said:

stuck. There is an inn at the top and the guy who runs it has told me

lots

Quoted message said:

Quoted message said:

of stories about this road. I used to climb it on a dare with friends.

It

Quoted message said:

Quoted message said:

rises about 2000 feet in 2 miles. It nuts! With cars usually the brakes

go

Quoted message said:

Quoted message said:

out and off the cliff they go. Ambulances used to hang out in that area.
Serious, that's just how many calls they used to get.

19% for average for 2 miles?

No, I exaggerated some. I think the grade sign says 18, not sure. I thought
actually it was more like 1800 feet in two miles.

Holy Hell. I'll see if I can find my Topo disc.