Fun with elevation readings

145 views
Skip to first unread message

jk

unread,
Sep 25, 2023, 5:45:16 PM9/25/23
to randonn...@googlegroups.com
I'll admit I found the elevation reading variations on our just-completed 8k600 interesting.

Bottom line: if you're building an 8k600 or SR600, which use OpenRunner as the source of record, keep an eye on it and ensure that it's in the same ballpark as RideWithGPS and Strava.

When I was building the route, OpenRunner would swing from over 8,000m (the target) to 6,000m (quite low).  While I was confident we'd met the criteria for an 8k600, I was a bit curious on what riders' GPS units would read.

Hoping this table copy/pastes correctly:

TypeSourceElev (ft)Elev (m)
RouteRideWithGPS25,7907,861
RouteOpen Runner26,7988,168
RouteOpen Runner (weird)20,7256,317




Rider


1Garmin Edge 53027,0218,236
1Strava adjusted27,0718,251
1Garmin Edge 530 (rwgps)27,1658,280
2Wahoo ELEMNT ROAM27,0288,238
3Wahoo ELEMNT BOLT32,7959,996
4Garmin Edge 1030+28,9738,831
5Garmin Edge 103027,1008,260


From the data, it looks like the non-funky OpenRunner route was pretty close.  27k feet of climbing is probably closest to reality.  It sure did feel like a lot of climbing!

Jonathan K
Cleveland, Ohio

Noah Swartz

unread,
Sep 25, 2023, 10:33:44 PM9/25/23
to jk, randonn...@googlegroups.com
Best advertisement for the ELEMNT BOLT I've seen yet - impress all of your friends with your amazing climbing ability!

--
You received this message because you are subscribed to the Google Groups "Randonneurs USA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to randonneurs-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/randonneurs-usa/1692404454.5651585.1695678188856%40mail.yahoo.com.

Rob Hawks

unread,
Sep 25, 2023, 10:44:18 PM9/25/23
to Noah Swartz, jk, randonn...@googlegroups.com
In the 'each device of the same model is different' category, my Bolt v.2 always records lower than rwgps estimates, and lower than what Strava will correct to using their 'correct elevation' feature. I wouldn't take one reading from one device and assume all others will replicate that.

rob

Joshua Haley

unread,
Sep 26, 2023, 9:36:36 AM9/26/23
to Rob Hawks, Noah Swartz, jk, randonn...@googlegroups.com
Obviously, the rider with the BOLT was riding epically, ascending and descending through clouds at a rate the GPS couldn't keep up with.  Just by coincidence, I happen to ride with a BOLT.  Or the data is just "meh" 😂.  My "KneePS" system clocked the ride at "way too friggin hard," and they are rather achy today still. 

Josh

Eric Peterson

unread,
Sep 26, 2023, 1:20:18 PM9/26/23
to Randonneurs USA
Could you elaborate on the following?

  • if you're building an 8k600 or SR600, which use OpenRunner as the source of record, keep an eye on it and ensure that it's in the same ballpark as RideWithGPS and Strava.
In particular what is involved in "keep an eye on it" and "ensure that it's in the same ballpark"?

Eric Peterson

David Weigel

unread,
Sep 26, 2023, 2:57:59 PM9/26/23
to Randonneurs USA

I’ve noticed that different base maps lead to different ascent totals, even within the same app. Here is an example of the same route, mapped out in different base maps:

I haven’t been able to find any definitive answers yet as to why this happens. My best guess is that different base maps use different geoid models and/or interpolation algorithms.

As far as differences in recordings: It is my understanding that there are too many variables to compare devices with much accuracy as it is dependent on calibration, incoming storms, satellite positioning, interference with other devices, etc.

 - David

Bill Gobie

unread,
Sep 26, 2023, 3:54:58 PM9/26/23
to Randonneurs USA
According to this discussion of mapping apps,* RwGPS switches between two different routing engines depending on which map you choose. This has helped me understand when to change maps. Climbing is not discussed but you might expect to get only two different climbing totals, not three. The climbing calculation must be independent of the routing engines, at least sometimes. If a person really wants to go down the rabbit hole you can get route data by appending ".json" to a RwGPS route link, then massage the elevation data to your liking.

Bill

*Originally posted to the FB group Northwest Gravel Riders

--
You received this message because you are subscribed to the Google Groups "Randonneurs USA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to randonneurs-u...@googlegroups.com.

Jake Kassen

unread,
Sep 26, 2023, 6:05:55 PM9/26/23
to Randonneurs USA
Elevation aside, there's notable differences in the coordinates of roads and other structures between the basemaps. It's to the point where if you drew the same route in two separate basemaps, the distance might not even be identical.

On funny but inconsequential thing about RWGPS: The website stores route distances in meters and round to the nearest 1/10 of a mile or km when displaying the total. RUSA's policy is to truncate to whole kilometers. There are a few perm routes that are <500 meters short of the credited distance. In some cases, the route shouldn't have been approved if you want to be extremely strict.

So next time you feel like it was an easy ride, tell yourself it's only because you got a free 400 meters. :)

Jake

> -------Original Message-------
> From: Bill Gobie <gobie...@gmail.com>
> To:
> Cc: Randonneurs USA <randonn...@googlegroups.com>
> Subject: Re: [RUSA] Re: Fun with elevation readings
> Sent: 26 Sep '23 15:55
>
> According to this discussion of mapping apps,* RwGPS switches between
> two different routing engines depending on which map you choose. This
> has helped me understand when to change maps. Climbing is not
> discussed but you might expect to get only two different climbing
> totals, not three. The climbing calculation must be independent of the
> routing engines, at least sometimes. If a person really wants to go
> down the rabbit hole you can get route data by appending ".json" to a
> RwGPS route link, then massage the elevation data to your liking.
>
> Bill
>
> *Originally posted to the FB group Northwest Gravel Riders
>
> On Tue, Sep 26, 2023 at 11:57?AM David Weigel <weig...@gmail.com>
> wrote:
>
> > I’ve noticed that different base maps lead to different ascent
> > totals, even within the same app. Here is an example of the same
> > route, mapped out in different base maps:
> >
> > * 2,605ft Openrunner (Topo)
> > * 2,516ft Openrunner (Openrunner)
> > * 2,613ft RwGPS (RWGPS Cycle)
> >
> > * 2,593ft RwGPS (Google Map)
> >
> > * 2,604ft RwGPS (OSM Cycle)
> >
> > I haven’t been able to find any definitive answers yet as to why
> > this happens. My best guess is that different base maps use
> > different geoid models and/or interpolation algorithms.
> >
> > As far as differences in recordings: It is my understanding that
> > there are too many variables to compare devices with much accuracy
> > as it is dependent on calibration, incoming storms, satellite
> > positioning, interference with other devices, etc.
> >
> > - David
> >
> > On Tuesday, September 26, 2023 at 11:20:18?AM UTC-6 Eric Peterson
> > wrote:
> > Could you elaborate on the following?
> >
> > * if you're building an 8k600 or SR600, which use OpenRunner as the
> > source of record, keep an eye on it and ensure that it's in the same
> > ballpark as RideWithGPS and Strava.
> >
> > In particular what is involved in "keep an eye on it" and "ensure
> > that it's in the same ballpark"?
> >
> > Eric Peterson
> >
> > On Monday, September 25, 2023 at 4:45:16?PM UTC-5 jk wrote:
> >
> > I'll admit I found the elevation reading variations on our
> > just-completed 8k600 interesting.
> >
> > Bottom line: if you're building an 8k600 or SR600, which use
> > OpenRunner as the source of record, keep an eye on it and ensure
> > that it's in the same ballpark as RideWithGPS and Strava.
> >
> > When I was building the route, OpenRunner would swing from over
> > 8,000m (the target) to 6,000m (quite low). While I was confident
> > we'd met the criteria for an 8k600, I was a bit curious on what
> > riders' GPS units would read.
> >
> > Hoping this table copy/pastes correctly:
> >
> > TYPE
> > SOURCE
> > ELEV (FT)
> > ELEV (M)
> >
> > Route
> > RideWithGPS
> > 25,790
> > 7,861
> >
> > Route
> > Open Runner
> > 26,798
> > 8,168
> >
> > Route
> > Open Runner (weird)
> > 20,725
> > 6,317
> >
> > RIDER
> https://groups.google.com/d/msgid/randonneurs-usa/CA%2Bsg%3DPqy_0s45EQPa31_Kw%3DGCm%3DM1wZ8-a6VWmOk3LAV9uMzZw%40mail.gmail.com.
>

Jonathan Karpick

unread,
Sep 27, 2023, 8:16:25 AM9/27/23
to Randonneurs USA
Eric - I had two things happen while developing this route.  I'd wager the issues are compounded when the elevation gains result from relentless smaller climbs.  The challenge is real - that's a much tougher task than measuring elevation gain from a few (more a bit more than a few) large climbs:

1) I was building in RwGPS and exporting GPS to import into OpenRunner, obviously keeping an eye on the 8,000m target.  At the beginning, all was good.  OpenRunner was close to RwGPS.  It looked like we could get a feasible route in.  Then, something happened.  OpenRunner started showing 25% less climbing.

2) I decided to build the route directly in OpenRunner.  As I got over ~400k in distance, the reported elevation gain started DECREASING as I ADDED hills and distance to the route.  That shouldn't happen.

I ended up keeping an eye on that elevation and beating the OpenRunner route into submission by tweaking route points so that elevation wouldn't dip as I added climbs.

My guess is that after a certain distance, OpenRunner drops the 'sample rate' across the route, and possibly in a way that clipped off the top of many of our 'little' climbs.  It's just a guess.  It's also possible that by being careful with my route points, I was somehow forcing it to recognize climbs.

The other lesson learned:  OpenRunner isn't a great tool.  Avoid it unless forced by ACP.

-Jonathan

Raphaël

unread,
Sep 27, 2023, 11:34:41 AM9/27/23
to Jonathan Karpick, Randonneurs USA
+1 on OpenRunner not being a user friendly tool. RideWithGPS or Komoot are much better. 

Raphaël 

On Sep 27, 2023, at 05:16, 'Jonathan Karpick' via Randonneurs USA <randonn...@googlegroups.com> wrote:

Eric - I had two things happen while developing this route.  I'd wager the issues are compounded when the elevation gains result from relentless smaller climbs.  The challenge is real - that's a much tougher task than measuring elevation gain from a few (more a bit more than a few) large climbs:
--
You received this message because you are subscribed to the Google Groups "Randonneurs USA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to randonneurs-u...@googlegroups.com.

Martin Shipp

unread,
Sep 27, 2023, 8:49:10 PM9/27/23
to Randonneurs USA, Jonathan Karpick
"the reported elevation gain started DECREASING as I ADDED hills and distance
I've experienced the same in RWGPS.  
 
For as long as I've been developing routes on-line, I've thought that the elevation on a given route, or to compare routes, was fun and maybe interesting, but not reliable or accurate or precise and sometimes not even re-producable. 
 
The distances have always been good (I'm referring as far back as 2011); the auto-generated cue sheets have been mostly reasonable in recent years (however, circa-2012, they were TERRIBLE); but as for the elevations reported -- see the first paragraph. 
   
...Martin  


On Wednesday, September 27, 2023 at 08:16:28 AM EDT, 'Jonathan Karpick' via Randonneurs USA <randonn...@googlegroups.com> wrote:

... 

2) I decided to build the route directly in OpenRunner.  As I got over ~400k in distance, the reported elevation gain started DECREASING as I ADDED hills and distance to the route.  That shouldn't happen.

... 

-Jonathan 

Eric Nichols

unread,
Sep 28, 2023, 11:06:34 AM9/28/23
to Randonneurs USA
Various maps and apps use different digital elevation models. Some follow the road, others follow the land surface, all with varying levels of accuracy.  

A good example is when a road crosses a high bridge.  Some maps and apps will assume the route plunges down the canyon to the water surface, then climbs up the other side (!).  

Another example is when the road is cut into a steep canyon wall -- small errors in the position of the road can result in false climbs and descents. 

Every map has inaccuracies, some small, some large. I've been helping to improve the OSM Cycle maps.  These have gotten much better over time. 

Eric Nichols 

Eric Peterson

unread,
Sep 28, 2023, 7:50:07 PM9/28/23
to Randonneurs USA
Jonathan - thanks for explaining the process for working with the  different tools to get a route that met the ACP climbing requirement. 
I've not used OpenRunner but there is certainly a consensus that it is not a easy tool to use. 
I can only imagine your frustration in seeing the amount of climbing decrease as you added climbs. Your explanation seems reasonable.
I'm glad to hear that you worked your way through those difficulties and came up with what from all reports was a great route. If you like climbing that is. 

All these tools are imperfect to some degree I think, and the ability to use different base maps only compounds the inconsistency in reported elevation gain. 
The GPS devices themselves don't seem to produce consistent numbers either.
I found this variation pretty annoying when I first started using GPS devices. The truth is out there, but the only way to know for sure would be to traverse a route with a level, a ruler, and a notepad and even then there would be measurement errors. 

Eric Peterson

Josh Brown

unread,
Sep 28, 2023, 8:11:41 PM9/28/23
to Eric Peterson, Randonneurs USA
I use a Garmin 830 and a Garmin Enduro watch and they consistently come up with different ride metrics both in miles and elevation, I of course keep the one that is smaller.

Josh in NYC 

--
You received this message because you are subscribed to the Google Groups "Randonneurs USA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to randonneurs-u...@googlegroups.com.

Pete Dusel

unread,
Oct 2, 2023, 8:41:50 AM10/2/23
to Randonneurs USA
  One thing to remember, is everybody has their own algorithm for computing climb, and they don't often share it.
How much up does it need to be to be called climb? Do we ignore everything below 1"?, 10 feet? 20 feet? Speed bumps?

 For example, my old Polar, which documented this, had to see 30 feet of up, before it added that climb to the total climb. A sting of 20 foot rollers added no climb. No clue what my Garmin uses.

  Also, altitude is the least accurate GPS measurement. Some GPS units add a barometric altimeter to help.

  In the end, I don't think there is an absolute correct answer for climb on road routes.

Pete

Joshua Haley

unread,
Oct 2, 2023, 9:02:15 AM10/2/23
to Pete Dusel, Randonneurs USA
I think my take away here is nothing is perfect, but only OpenRunner is "correct" for the ACP. It's worth making the route there, rather than expecting a RWGPS route to tick that 8k600k box.

Josh

--
You received this message because you are subscribed to the Google Groups "Randonneurs USA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to randonneurs-u...@googlegroups.com.

Pete Dusel

unread,
Oct 2, 2023, 9:06:14 AM10/2/23
to Randonneurs USA
I agree, just do it in Openrunner.

As we used to say, there's the Right answer, the Wrong answer, and the ACP answer.
Guess which one matters here!

Pete

Mike Sturgill

unread,
Oct 2, 2023, 10:16:57 AM10/2/23
to Pete Dusel, Randonneurs USA
Agree re/Openrunner, to a degree. I tried using OR for my 8k600 route so that I could ensure it met the criteria. But, it started giving me messages that I needed to pay and subscribe in order to route anything over about 330k. I then went thru the RUSA route committee to get them to upload my RWGPS route and give me the climbing/distance figures. It would be nice if I had access to do it myself, but didn't want another subscription for a one-off route creation.

-Mike

Rob Hawks

unread,
Oct 2, 2023, 10:40:21 AM10/2/23
to Mike Sturgill, Pete Dusel, Randonneurs USA
Hi all,

I'm not so quick to point fingers at either the ACP or OpenRunner for this marriage. There is sense in picking one platform and sticking to it rather than accepting submissions from many. Openrunner is likely far more established in France than Ridewithgps is and therefore it is what they know best. Further, can they really be faulted for picking a platform from their own region? Many of us favor Ridewithgps, and a fair bit of that favor could be because it is what we know already, and Ridewithgps is local to us.

As an aside, I've not used Openrunner myself and can only base my comparison on comments I've heard from others that have used it. I would not be surprised at all to read that an honest comparison would show Ridewithgps to be the better product.

There is very likely an advantage in leaving the 'testing' of elevation gain to the RUSA routes team. The more 8K600 routes they process, the more adept they become in using Openrunner and the more consistency they can offer in performing the task. I think we may be close if not there yet in having enough data to say that a very careful creation of a route on Ridewithgps will produce a predictable Openrunner elevation gain. Plus, why do more work yourself when the Routes team will be doing it anyway when they approve your (Ridewithgps) route submission. I might be wrong about this part, but it is possible that the Routes team is able to extract the Ridewithgps route as a file and upload that to Openrunner. That would provide a consistent process, no?

rob

Joshua Haley

unread,
Oct 2, 2023, 10:57:50 AM10/2/23
to Rob Hawks, Mike Sturgill, Pete Dusel, Randonneurs USA
The problem we saw in Ohio with our short climbs is the route built in Open runner checked the box. Exporting the route from RWGPS and pushing to open runner did not and was short.

Actual GPS readings and Strava elevation had us all over the climbing minimum. Might be a unique problem for our region with tons of short punchy climbs. 🤷‍♂️ Maybe this is a "10%" of the time issue we address as needed.


Josh


Reply all
Reply to author
Forward
0 new messages