Climb estimates on Toaster / BikeHike

General cycling advice ( NOT technical ! )
bigfoot
Posts: 159
Joined: 27 Sep 2008, 8:06pm
Location: Frosham, Cheshire

Climb estimates on Toaster / BikeHike

Post by bigfoot »

It's always seemed obvious when riding courses created on Bikeroute Toaster, that the "Total Ascent" is less than my Edge 705 would report.

This was brought home printing off some back-up maps for my LEJOG next week on memory map. I happened to create a route card in Memory Map and noticed the total ascent was much higher than toaster reported.

So , Launceston - Taunton, Route in toaster shows 75.2 miles, 3290 feet of climb
The same tcx file loaded into Bikehike shows 79 miles 4140 feet of climb
and memory map with the same file loaded shows 73.8 miles, and 5454 feet of climb

I suspect the resolution of the elevation data is to blame. Any more experience of this.
snibgo
Posts: 4604
Joined: 29 Jun 2010, 4:45am

Re: Climb estimates on Toaster / BikeHike

Post by snibgo »

From playing with the numbers, I have concluded the Open Street Map elevations were collected in cars. At the slower speeds of a bike, there is greater resolution, so more detail, so greater ascents and descents. (The elevation curve of a route will be a fractal.)
User avatar
Mick F
Spambuster
Posts: 56359
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Climb estimates on Toaster / BikeHike

Post by Mick F »

Here's my ride from yesterday:
Pork and Peek.jpg
Up Pork Hill onto Dartmoor, then home down Peek Hill.

From my records in my computer taken from the actual Garmin figures:
28.69miles
3,407ft Total Ascent

From BikeHike tracing the route:
28.57miles
2,918ft Total Ascent

From BRT tracing the route:
28.54miles
2,546ft Total Ascent
Mick F. Cornwall
niggle
Posts: 3435
Joined: 11 Mar 2009, 10:29pm
Location: Cornwall, near England

Re: Climb estimates on Toaster / BikeHike

Post by niggle »

snibgo wrote:From playing with the numbers, I have concluded the Open Street Map elevations were collected in cars. At the slower speeds of a bike, there is greater resolution, so more detail, so greater ascents and descents. (The elevation curve of a route will be a fractal.)

It is the exact equivalent of the mathematical problem called the Coastline Paradox, the plotted graph of elevation of a route has a seemingly infinite degree of features depending on how closely you look at it, so the greater the resolution the bigger the result.

However there is a suggestion that the best solution is to measure features only down to the size of the unit used for the final result, so for a result in metres just take elevation measurements once every metre travelled horizontally. I do not know how often a Garmin does this or the resolution used in programmes like Bikehike and Bike Route Toaster, but they obviously must be different.

I think for cycling there needs to be an arbitrary agreement made, probably best decided by an organisation such as Audax UK.
Ayesha
Posts: 4192
Joined: 30 Jan 2010, 9:54am

Re: Climb estimates on Toaster / BikeHike

Post by Ayesha »

When I planned my E2E, I visually checked the Landranger 1:50,000 to see where the contours were less than 5mm apart. These, I counted as 'hills'. Everywhere else was 'flat'.

I put a red 'X' on the map at the base of 'hills' which I assessed as 'a challenge', and made sure I got some sugars down my neck twenty or so minutes before the 'X's.
Ayesha
Posts: 4192
Joined: 30 Jan 2010, 9:54am

Re: Climb estimates on Toaster / BikeHike

Post by Ayesha »

niggle wrote:
snibgo wrote:From playing with the numbers, I have concluded the Open Street Map elevations were collected in cars. At the slower speeds of a bike, there is greater resolution, so more detail, so greater ascents and descents. (The elevation curve of a route will be a fractal.)

It is the exact equivalent of the mathematical problem called the Coastline Paradox, the plotted graph of elevation of a route has a seemingly infinite degree of features depending on how closely you look at it, so the greater the resolution the bigger the result.

However there is a suggestion that the best solution is to measure features only down to the size of the unit used for the final result, so for a result in metres just take elevation measurements once every metre travelled horizontally. I do not know how often a Garmin does this or the resolution used in programmes like Bikehike and Bike Route Toaster, but they obviously must be different.

I think for cycling there needs to be an arbitrary agreement made, probably best decided by an organisation such as Audax UK.


The 'arbitary agreement' is made by yourself.

Digital mapping packages on the web give different results for 'total ascent'. I have found BikeRouteToaster to be the most suitable for my needs, as I 'THINK' it takes its elevation measurements ala OS contours.
On an OS map, the contours are 5m apart on Pathfinder. How much 4m up-and-down undulation occurs between contours is anyone's guess ( except the package which costs a blinding fortune to download ).

My own 'arbitary agreement' is that undulations of less than 5m elevation are not worth bothering about. They can be managed with a few pedal strokes of 'Honking' without a gearchange.

BTW, AUK's 'carved in stone' method of determining climbing for their Altitude Award is....
Contour counting.
Some AUK organisers are shelling out for Tracklogs digital mapping. ( The package that costs a blinding fortune ).

Estimating the climbing on a bike ride serves two purposes.
1/ Additional time to complete the ride, which is based on the individual's cycling ability and is quite complicated.
2/ Additional energy requirement over the 'flat ground' energy requirement. This is easy. Its kg x 9.81 x metres climbed, in kJ.
Last edited by Ayesha on 17 Jun 2011, 9:14am, edited 1 time in total.
User avatar
Mick F
Spambuster
Posts: 56359
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Climb estimates on Toaster / BikeHike

Post by Mick F »

My 705 is set to record every second, so fractals would be very small.

Online mappings show a smooth curve, so their fractals are small too. Try out a short route on one and look at the curve.

However, they work out their total ascents by only looking at the contour lines. Unless the route goes over a contour line, it's flat up to that line. The profile then becomes a staircase and the nuances of the actual profile are missed.

Therefore online mapping ascents need to be taken as LESS than reality.

Has that summed it up ok?
Mick F. Cornwall
Ayesha
Posts: 4192
Joined: 30 Jan 2010, 9:54am

Re: Climb estimates on Toaster / BikeHike

Post by Ayesha »

I had a chat with the organiser of the Meriden Memorial Audax.
He said he measured the climbing using Tracklogs because it has elevation data at 1m spacings !! A bit difficult to believe at first, but another rider confirmed the ascent figure with his Garmin that had a barometric altimeter ( Edge 705 ).

My Edge 605 with 3D positioning elevation calcuation showed neary 40% more. BRT gave a result of nearly 40% less.

The route was a bit bumpy, with three decent hills. All the little bitty bumps counted to Tracklogs and Edge 705. They didn't count into BRT, and my Edge 605 was so inacurate, it was bouncing up and down with interference from buildings and overhanging trees.

After a couple of years of using BRT as my 'climbing calculator', I simply double the result and multiply it by my kg and 9.81 to get the Joules for the climbing. 4186 Joules is 1 kCal.
Add this to 15 kCals / km and I do fine.

BTW Bigfoot, are you assessing climbing on your E2E for nutrition requirement purposes?
User avatar
meic
Posts: 19355
Joined: 1 Feb 2007, 9:37pm
Location: Caerfyrddin (Carmarthen)

Re: Climb estimates on Toaster / BikeHike

Post by meic »

Mick F wrote:My 705 is set to record every second, so fractals would be very small.

Online mappings show a smooth curve, so their fractals are small too. Try out a short route on one and look at the curve.

However, they work out their total ascents by only looking at the contour lines. Unless the route goes over a contour line, it's flat up to that line. The profile then becomes a staircase and the nuances of the actual profile are missed.

Therefore online mapping ascents need to be taken as LESS than reality.

Has that summed it up ok?


I dont know about the others but Anquet at least do not do that. They have used computer programs which have studied the adjacent contours and plotted a continuous plane over the surface of the virtual landscape. It isnt, of course, the same as the real landscape but it is a much more sophisticated approximation.
Yma o Hyd
User avatar
[XAP]Bob
Posts: 19793
Joined: 26 Sep 2008, 4:12pm

Re: Climb estimates on Toaster / BikeHike

Post by [XAP]Bob »

All the different methods used here show why it's a hard question.

I think the prime "difficulty" is in the difference between starting and ending elevation.
For every uphill grind there is a downhill to enjoy (assuming you go back home at some point...)

Having said that I'd prefer a short sharp climb followed by a long gentle descent to any of the other alternatives.
A shortcut has to be a challenge, otherwise it would just be the way. No situation is so dire that panic cannot make it worse.
There are two kinds of people in this world: those can extrapolate from incomplete data.
User avatar
Mick F
Spambuster
Posts: 56359
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Climb estimates on Toaster / BikeHike

Post by Mick F »

[XAP]Bob wrote:All the different methods used here show why it's a hard question.

I think the prime "difficulty" is in the difference between starting and ending elevation.
My example above, of my ride yesterday, started and ended at the same spot. Even accounting for Garmin elevation errors due the atmospheric pressure, there's a difference of over 800ft from Garmin records to BRT.

I feel an experiment coming on.
Read off the elevation from an OS map, cycle up a hill to another known OS elevation, and see how the various methods stack up.
Mick F. Cornwall
User avatar
Mick F
Spambuster
Posts: 56359
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Climb estimates on Toaster / BikeHike

Post by Mick F »

I've found a likely route:
http://www.bikeroutetoaster.com/Course. ... rse=264201
Start and end are at Spot Heights as per OS map.
Start is on a little bridge - 118mtrs
End is at the road junction down to Lydford - 267mtrs
Both places I know well.

I can get to the start and set my 705 to the correct elevation - 118mtrs = 383.5ft and start recording. I can stop at the end 5.75miles later 267mtrs = 867.75ft.

BRT shows total ascent of 672ft
BikeHike shows total ascent of 779ft

I must say that the top of the FIRST hill at Blackdown - western fringes of Dartmoor - is at about 1,000ft, so the total ascent just to there must be 617ft to start with!

I wonder what my 705 will say?
Not today, but maybe in the next day or two.

Perhaps someone else could have a test. Pouring down and windy out there here now!
Mick F. Cornwall
John-D
Posts: 241
Joined: 14 Mar 2010, 9:31am
Location: Haworth, West Yorkshire

Re: Climb estimates on Toaster / BikeHike

Post by John-D »

Mick F wrote:I've found a likely route:
http://www.bikeroutetoaster.com/Course. ... rse=264201
Start and end are at Spot Heights as per OS map.
Start is on a little bridge - 118mtrs
End is at the road junction down to Lydford - 267mtrs
Both places I know well.

5.75miles later 267mtrs = 867.75ft.

BRT shows total ascent of 672ft
BikeHike shows total ascent of 779ft



I've just downloaded that route on to Anquet and it came up with the following:

Flat length: - - 5.75m
Hidden length: - - 5.76m
Height asc: - - 829.58ft
Height desc: - - 327.53ft
Mx height: - - 1032.18ft
Min height: - - 379.26ft

John
User avatar
Mick F
Spambuster
Posts: 56359
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Climb estimates on Toaster / BikeHike

Post by Mick F »

Thanks John!
More info the better.

I'll see what the weather's like tomorrow ......
Saturday:
Another cool day with blustery showers across the region for much of the day, some heavy and thundery at times. Showers may ease in the evening. Brisk southwesterly winds. Maximum temperature 17 °C.


Outlook for Sunday to Tuesday:
Drier and brighter conditions developing on Sunday, making it feel warmer at times. Unsettled weather returning once again on Monday with outbreaks of heavy rain. Brisk southwesterly winds on Tuesday.


It looks like a Sunday ride!
Mick F. Cornwall
Ayesha
Posts: 4192
Joined: 30 Jan 2010, 9:54am

Re: Climb estimates on Toaster / BikeHike

Post by Ayesha »

That short route has a big climb, a descent and then another sharp climb.

Its a climb I would take note of. I've got it to 827 ft climbing by sub-dividing it on Garmin Topo GB 'Show profile'.

Will be back to see the result on Sunday evening.
Post Reply