Crr and CdA when riding a route created from GPS

[Steve Edmonds]

I have created some workouts from activities recorded with GPS. CRS files.

How do I set Crr and CdA for use when speed from power is set in Options>Training>Preferences.

Are they set in Edit>Estimate power values. If I make changes here and save them, when I go to Edit>Estimate power values again the changes are not shown, is this normal and I just have to assume the previously entered values are active?

Are the saved values global or only for the selected workout in the train tab, are the settings saved between exits from GC.

Thanks, steve

[Steve Edmonds]

I notice settings as below in configglobal-general.ini, can I put values specific to me in here to be used in "Estimate power values"

fixderivepower\bikewheight=9.5
fixderivepower\cda=0
fixderivepower\crr=0.0031

[Ale Martinez]

El sábado, 14 de agosto de 2021 a la(s) 19:32:36 UTC-7, Steve Edmonds escribió:

I notice settings as below in configglobal-general.ini, can I put values specific to me in here to be used in "Estimate power values"

fixderivepower\bikewheight=9.5
fixderivepower\cda=0
fixderivepower\crr=0.0031

The user interface to update Edit menu Data Processors parameters is documented in: https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_Preferences_Data-Fields#processing

 

On Sunday, 15 August 2021 at 13:20:08 UTC+12 Steve Edmonds wrote:

I have created some workouts from activities recorded with GPS. CRS files.

How do I set Crr and CdA for use when speed from power is set in Options>Training>Preferences.

Are they set in Edit>Estimate power values. If I make changes here and save them, when I go to Edit>Estimate power values again the changes are not shown, is this normal and I just have to assume the previously entered values are active?

Are the saved values global or only for the selected workout in the train tab, are the settings saved between exits from GC.

Virtual bike for simulated rides in Train mode is currently unrelated to Edit > Estimate Power 

[Steve Edmonds]

I think I have answered my own question by changing the values in configglobal-general.ini

GC was drastically overestimating the downhill speed but with fixderivepower\bikewheight=28, fixderivepower\cda=0.55, fixderivepower\crr=0.011 it is pretty close to my observed downhill speeds.

I'm training with a 13kg pack hence the bikewheight=28

I am not too sure the uphill estimation is so close (GC seems to be overestimating speed) so will review this info.

BTW, I assume that fixderivepower\bikewheight is bike weight and not a height factor, see it also in src/Core/Settings.h

[Steve Edmonds]

I see that  Edit > Estimate Power draws it's default values from configglobal-general.ini as does Virtual bike for simulated rides in Train mode which led to a little confusion for me thinking Edit > Estimate Power wopuld save back to the .ini

[Steve Edmonds]

In reference to the uphill speed, with the settings I entered in configglobal-general.ini GC has me riding up a 10% slope at 4.8km/hr with 100 watts input.

From https://www.gribble.org/cycling/power_v_speed.html climbing at this speed would require ~145 watts and "feels" a little closer to riding the actual hill. If I can find it I will take a look at the code and see if I can see what data GC is actually taking for the calculation for Virtual bike for simulated rides.

[Eric Christoffersen]

Hi Steve.

All the bicycle sim parameters are at tools/options then go to the training tab, then the 'virtual bike specifications' sub-tab. There's tool tips for each. Everything that is used in sim is there except rider weight which is an athlete parameter.

To test make a crs with a 10% slope, use the robo trainer which by default emit an average of 100w.

I just did that test and found the steady state speeds exactly match that web site.

Beware because of sample noise a raw gpx needs quite a bit of altitude smoothing to feel realistic. Smooth your workout, save as json, then load json as workout and ride it.

Eric

[Steve Edmonds]

Thanks Eric.

On 16/08/2021 02:17, Eric Christoffersen wrote:

Hi Steve.

All the bicycle sim parameters are at tools/options then go to the training tab, then the 'virtual bike specifications' sub-tab. There's tool tips for each. Everything that is used in sim is there except rider weight which is an athlete parameter.

From the derived front wheel mass I see the value entered for front wheel mass is effectively the rim, spokes, and hub combined to which the other components are added.

To test make a crs with a 10% slope, use the robo trainer which by default emit an average of 100w.

I just did that test and found the steady state speeds exactly match that web site.

I had previously made a big dipper (2.5%, 5%, 10%) slope down to check freewheeling terminal velocity and back up again.
Now aligns close enough thanks.

Beware because of sample noise a raw gpx needs quite a bit of altitude smoothing to feel realistic. Smooth your workout, save as json, then load json as workout and ride it.

The routes I am training for are multi day rides I have not yet ridden so I am getting the GPX files of other riders off RWGPS who already do a reasonable amount of elevation smoothing. I am then importing them with around a 20m segment length. This is not unrealistic and possibly a little too much smoothing has resulted. I noticed one place in your code you are doing a 50m look-ahead, not sure if this applied to the virtual ride because I haven't yet got my head around where the various bits of code are used.

Eric

On Saturday, August 14, 2021 at 10:14:47 PM UTC-7 Steve Edmonds wrote:

In reference to the uphill speed, with the settings I entered in configglobal-general.ini GC has me riding up a 10% slope at 4.8km/hr with 100 watts input.

From https://www.gribble.org/cycling/power_v_speed.html climbing at this speed would require ~145 watts and "feels" a little closer to riding the actual hill. If I can find it I will take a look at the code and see if I can see what data GC is actually taking for the calculation for Virtual bike for simulated rides.

On Sunday, 15 August 2021 at 16:58:51 UTC+12 Steve Edmonds wrote:

On Sunday, 15 August 2021 at 16:49:25 UTC+12 Ale Martinez wrote:

El sábado, 14 de agosto de 2021 a la(s) 19:32:36 UTC-7, Steve Edmonds escribió:

I notice settings as below in configglobal-general.ini, can I put values specific to me in here to be used in "Estimate power values"

fixderivepower\bikewheight=9.5
fixderivepower\cda=0
fixderivepower\crr=0.0031

The user interface to update Edit menu Data Processors parameters is documented in: https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_Preferences_Data-Fields#processing

 

On Sunday, 15 August 2021 at 13:20:08 UTC+12 Steve Edmonds wrote:

I have created some workouts from activities recorded with GPS. CRS files.

How do I set Crr and CdA for use when speed from power is set in Options>Training>Preferences.

Are they set in Edit>Estimate power values. If I make changes here and save them, when I go to Edit>Estimate power values again the changes are not shown, is this normal and I just have to assume the previously entered values are active?

Are the saved values global or only for the selected workout in the train tab, are the settings saved between exits from GC.

Virtual bike for simulated rides in Train mode is currently unrelated to Edit > Estimate Power 

I see that  Edit > Estimate Power draws it's default values from configglobal-general.ini as does Virtual bike for simulated rides in Train mode which led to a little confusion for me thinking Edit > Estimate Power wopuld save back to the .ini

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/0b0ea36a-13ab-4618-8634-8ec0c05951c0n%40googlegroups.com.

[Eric Christoffersen]

> The routes I am training for are multi day rides I have not yet ridden so I am getting the GPX files of other riders off RWGPS who already do a reasonable amount of elevation smoothing. I am then importing them with around a 20m segment length. This is not unrealistic and possibly a little too much smoothing has resulted. I noticed one place in your code you are doing a 50m look-ahead, not sure if this applied to the virtual ride because I haven't yet got my head around where the various bits of code are used.

Awesome.Lets see if we can get you a first class experience. First, I don't know about the 50m, maybe share a line to the line in the source?

Second... you used the keyword "segment length"... that implies you are importing using the workout creator. YMMV but I think the workout creator is a broken vestige that should be deleted from the project, it takes your wonderful workout and flattens it into a crs, you lose location info so won't have a live map and it discretizes your slopes so your ride slopes won't be continuous. There is much better tech available now.

I think the right way to go is:
1) Import your file into gc as an activity: Activity->ImportFromFile. That causes your activity to be imported into gc's internal json format. Advantage is that it doesn't have weird limitations and can hold all fields.
2) You can import multiple files as activities, then append them to make a single large one.

3) Smooth the new activity using Edit->FixGPSErrors. You might want to get fancy and smooth route itself to look good on a map, but for ridable workout you just need to smooth the altitude data.

FixGPS: Add ability to prune outliers · Issue #3320 · GoldenCheetah/GoldenCheetah (github.com)

4) Once you have the activity looking like you'd expect (smooth continuous non-insane gradients), export the activity as JSON file. Give it whatever name you want, this is your new workout file to use with trainier. Put it in a directory with your other workouts.
5) I train mode, do a 'scan for workouts' and it will find and automatically load your new json workout.
6) Be sure and add live map so you can see where you are on map while doing your workout. Much more fun than just gradient. You can tell yourself "sprint to next switchback".
7) Iterate on 3,4,5 until you get the ride the way you want. You can edit specific data (edit valus, delete rows/columns) in the activity details tab.
8) Very important: under tools/options, training tab, preferences sub-tab: enable "simulate speed from power" which enables the simulation.
9) You might want to enable "simulate relative hypoxia" which will steal your watts based on the delta of your actual training altitude and the virtual ride's altitude. (Uses "actual training altitude" from virtual bicycle specifications page.)

Simulated Hypoxia for Train Mode by ericchristoffersen · Pull Request #3699 · GoldenCheetah/GoldenCheetah (github.com)

Let me know what you think!
Eric

[Ale Martinez]

El lunes, 16 de agosto de 2021 a la(s) 07:49:17 UTC-6, zak...@gmail.com escribió:

> The routes I am training for are multi day rides I have not yet ridden so I am getting the GPX files of other riders off RWGPS who already do a reasonable amount of elevation smoothing. I am then importing them with around a 20m segment length. This is not unrealistic and possibly a little too much smoothing has resulted. I noticed one place in your code you are doing a 50m look-ahead, not sure if this applied to the virtual ride because I haven't yet got my head around where the various bits of code are used.

Awesome.Lets see if we can get you a first class experience. First, I don't know about the 50m, maybe share a line to the line in the source?

Second... you used the keyword "segment length"... that implies you are importing using the workout creator. YMMV but I think the workout creator is a broken vestige that should be deleted from the project, it takes your wonderful workout and flattens it into a crs, you lose location info so won't have a live map and it discretizes your slopes so your ride slopes won't be continuous. There is much better tech available now.

I think the right way to go is:
1) Import your file into gc as an activity: Activity->ImportFromFile. That causes your activity to be imported into gc's internal json format. Advantage is that it doesn't have weird limitations and can hold all fields.
2) You can import multiple files as activities, then append them to make a single large one.

3) Smooth the new activity using Edit->FixGPSErrors. You might want to get fancy and smooth route itself to look good on a map, but for ridable workout you just need to smooth the altitude data.

FixGPS: Add ability to prune outliers · Issue #3320 · GoldenCheetah/GoldenCheetah (github.com)

4) Once you have the activity looking like you'd expect (smooth continuous non-insane gradients), export the activity as JSON file. Give it whatever name you want, this is your new workout file to use with trainier. Put it in a directory with your other workouts.
5) I train mode, do a 'scan for workouts' and it will find and automatically load your new json workout.

I think we could add an option to Workout Wizard to simplify this workflow: similar to create workout from current activity, but opening an instance of Fix GPS for the smoothing and import the result afterwards.

[Steve Edmonds]

On 17/08/2021 13:06, Ale Martinez wrote:

El lunes, 16 de agosto de 2021 a la(s) 07:49:17 UTC-6, zak...@gmail.com escribió:

> The routes I am training for are multi day rides I have not yet ridden so I am getting the GPX files of other riders off RWGPS who already do a reasonable amount of elevation smoothing. I am then importing them with around a 20m segment length. This is not unrealistic and possibly a little too much smoothing has resulted. I noticed one place in your code you are doing a 50m look-ahead, not sure if this applied to the virtual ride because I haven't yet got my head around where the various bits of code are used.

Awesome.Lets see if we can get you a first class experience. First, I don't know about the 50m, maybe share a line to the line in the source?

Second... you used the keyword "segment length"... that implies you are importing using the workout creator. YMMV but I think the workout creator is a broken vestige that should be deleted from the project, it takes your wonderful workout and flattens it into a crs, you lose location info so won't have a live map and it discretizes your slopes so your ride slopes won't be continuous. There is much better tech available now.

I think the right way to go is:
1) Import your file into gc as an activity: Activity->ImportFromFile. That causes your activity to be imported into gc's internal json format. Advantage is that it doesn't have weird limitations and can hold all fields.
2) You can import multiple files as activities, then append them to make a single large one.

3) Smooth the new activity using Edit->FixGPSErrors. You might want to get fancy and smooth route itself to look good on a map, but for ridable workout you just need to smooth the altitude data.

FixGPS: Add ability to prune outliers · Issue #3320 · GoldenCheetah/GoldenCheetah (github.com)

4) Once you have the activity looking like you'd expect (smooth continuous non-insane gradients), export the activity as JSON file. Give it whatever name you want, this is your new workout file to use with trainier. Put it in a directory with your other workouts.
5) I train mode, do a 'scan for workouts' and it will find and automatically load your new json workout.

I think we could add an option to Workout Wizard to simplify this workflow: similar to create workout from current activity, but opening an instance of Fix GPS for the smoothing and import the result afterwards.

This is proving interesting. The results seem somewhat dependant on the file logging parameters (rate of data point logging). For the GPX file I initially used the CRS I produced is closer to expected riding. I can see the B-spline smoothing with Edit->FixGPSErrors would be fine for a road ride with slower slope changes but it may not be most appropriate for a trail ride with more up and down. I am just analysing some GPX logs at the moment to see what is typical variability. I suspect a GPX recorded with barometric elevation may need to be handled differently from GPS or ground radar elevation.

To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/7c5ec22e-bca1-4be7-a986-d967ad9bb295n%40googlegroups.com.

[Eric Christoffersen]

>  This is proving interesting. The results seem somewhat dependant on the file logging parameters (rate of data point logging). For the GPX file I initially used the CRS I produced is closer to expected riding. I can see the B-spline smoothing with Edit->FixGPSErrors would be fine for a road ride with slower slope changes but it may not be most appropriate for a trail ride with more up and down. I am just analysing some GPX logs at the moment to see what is typical variability. I suspect a GPX recorded with barometric elevation may need to be handled differently from GPS or ground radar elevation.

Yeah, I was just speaking in general about importing gps tracks. If you're downloading someone else's route then maybe its already been manicured and doesn't need any smoothing. In that case just import as activity and export as json. As for over-smoothing... did you play with the smoothing parameters? Its pretty flexible and you can set it so it doesn't smooth at all.

The fundamental reason to avoid crs for virtual ride is that when used for virtual ride the gradient interpolation is turned off. With crs if you have:

.1 10%

.1 -10%

.1 10%

As crs this is taken literally and you'll get a steep sided sawtooth. But same ride exported as json the virtual ride sees 3 location points on route and interpolates a spline between them. You'll ride a smooth sine-wave type thing. At least with an actual route the interpolated spline is much more pleasant and realistic and the inter-point variance is in the tenths of a %.

The general route smoothing is a terrible problem because data has so many ways to be corrupted. I bet you're right that barometric would be naturally smooth, downside is that your route will rise and fall with the weather so a pass' altitude gain may be wrong. Gps is really bad because it gets distorted in different ways depending on the reception, especially up against cliffs, in canyons or in big forests. The very worst data I've found is the 'correct using internet map' because it is sometimes derived from a topomap with quantized altitude. You can tell that's what you have if the route is a set of flats with 5m steps... :) The slope gets nasty when the internet can't decide if a road is + or - and rapidly toggles between. Was thinking of writing a special case filter to detect this and de-quantize.

[Steve Edmonds]

>All the bicycle sim parameters are at tools/options then go to the training tab, then the 'virtual bike specifications' sub-tab. There's tool tips for each. Everything that is used in sim is there except rider weight which is an athlete parameter.

What is the assumption in respect of tyre weight and diameter. I assume diameter (radius) is required for rotational inertia. My front tire is over 1/2 my front wheel mass. ISO/E.T.R.T.O. = 559 and I have set outer rim diameter to 570/2. Outer tyre (wheel) diameter is 690 and most of the mass would be in the outer tread/casing as the side walls are relatively thin and flexible.

Looking at BicycleSim.cpp, should I enter outside wheel radius for Front/Rear Rim Outer Radius in the 'virtual bike specifications' sub-tab.

Thanks, steve

[Eric Christoffersen]

Ok that’s lunatic fringe, I like it! There’s inner rim radius and outer tire radius. Inertia math assumes density is uniform from inner to outer. It’s just the I for cylinder of uniform density.

https://github.com/GoldenCheetah/GoldenCheetah/blob/f9fa6ca401898fde0b6a79decf519c71bd172a67/src/Train/BicycleSim.cpp#L176

The ui displays computed I for the wheels, you could measure your wheels I then increase inner radius to drive center outward so I matches reality.

Or not. The rotational inertia makes a pretty small difference to ride time, especially since trainer has no brakes. Next you’ll complain it doesn’t model crankset inertia. :)

How is the virtual ride working?

[Steve Edmonds]

I was not sure if you were modelling rotational inertia, I saw a J in the equations somewhere which is what I used once for rotational inertia. The tyre weights would not be atypical for a MTB trail riding.

Where the UI asks for Front Rim Outer Radius I will use the wheel outer radius of my tyre and assume rim+tyre mass is distributed between inner rim and outer rim values.

Last night we entered the highest lockdown level with the first delta outbreak so there will not be any real riding for a few weeks.I am considering 2 sources of rides for virtual rides, my own logs at 1 or 2 second intervals and those of others I can download. My rides have plenty of points to be able to get a reasonably "smooth" and realistic elevation profile. Other peoples rides I download from Ride with GPS seem to be resampled at 10 second intervals where as my own rides download at original accuracy the ones that are resampled can have some unrealistic grade changes. This could potentially be from riding a switchback where the 2 points in the log end up almost above each other with a short distance between and large elevation change.

I am analysing files at the moment in a spreadsheet and then comparing to see what the various settings in GC Fix GPS errors does to the actual ride.

I have a trail I mowed in the paddock that I can log during lockdown, a bit steep in places but not uneven, with high accuracy and compare it against the file RWGPS would give me were it someone elses, i.e. resampled.

To add to the effort, RWGPS downloads other peoples rides in GPX format without times for TRKPTs but does download TCX files with time but these don't import into GC. The coordinates of the points in the files correspond, so I can put times into the GPX from the TCX.

[Eric Christoffersen]

COVID lockdown sucks. I tried to mtb pre vaccine and there were just too many people on nearby trails, had to fall back on trainer. I found mtb reality was just too varied to get reasonable simulation times, mud and sand and curves made the times too far off. The actual trails are too varied and I was constantly sneak attacked by steep slopes. Wasn’t fun without video. Big giant logging road climbs worked well.

That sucks that your tcx doesn’t load. It’s a nice format. Gpx can be pretty sparse, it doesn’t contain time and distance is optional. Maybe you can debug and fix it? Or share file here and someone can debug?

I’ll be really surprised if you ride the downloaded files without additional smoothing. The fix gps approach is powerful but seems like lots of people have trouble using it. Understand there’s two splines (smooth math curve) the first built with all points, the second is made with points remaining after outliers from first spline are discarded. Each splines basis is how many points are used for each spline control point, so bigger is smoother spline. When you finish your route is rewritten with values interpolated from the second spline. There’s a separate altitude pass because altitude tends to need a heavy hand, location spline is for location aesthetics.

For example, if you set outlier of 1000m then (probably) no points are discarded and the second basis is your smoothing spline. Set basis to 3 and the smoothing will be almost nonexistent. 

Way i use it is set first pass to largish number, outlier distance to something small like 0.5m, second spline to smaller basis. The first smooth curve exists to toss outliers, the second is the smoothing after outliers are removed. So you can set second basis to three and it’s your original with just outliers removed. Yes? The actual values depend on the noise and point variance, you can’t smooth as aggressively if your sample spacing is noisy.

 I’d love feedback if it doesn’t do what you want. There are other route smoothers on the Internet, I’d like to hear what you end up using.

Have you tried riding with live map? Can it keep your attention? I really dislike the non video ui so end up gravitating to virtual rides with movies. I really like alucards umbrail pass rlv.

[Steve Edmonds]

On Thursday, 19 August 2021 at 01:06:10 UTC+12 Eric Christoffersen wrote:

COVID lockdown sucks. I tried to mtb pre vaccine and there were just too many people on nearby trails, had to fall back on trainer. I found mtb reality was just too varied to get reasonable simulation times, mud and sand and curves made the times too far off. The actual trails are too varied and I was constantly sneak attacked by steep slopes. Wasn’t fun without video. Big giant logging road climbs worked well.

 I can get about 4km in my paddocks on the MTB with a lot of zig-zagging, as you say logging roads or even well graded trails work OK. I am not so much looking a times in the sim. but wanting to make sure I am prepared for the climbs. The sneak attack of the slopes or lack of video doesn't bother me much, ROUVY has a nice feature of a little preview of about 30 seconds of zoomed in slopes in advance, also on the map overlay marks the course with red hotspots for climbs, but that gets a bit depressing.

That sucks that your tcx doesn’t load. It’s a nice format. Gpx can be pretty sparse, it doesn’t contain time and distance is optional. Maybe you can debug and fix it? Or share file here and someone can debug?

 I was looking at the import code last night for the TCX, I hadn't realised the file structure could be so comprehensive. It has been a long time since I coded anything in C, let alone C++ and I was using an IDE for easy debugging. I downloaded a TCX from the test folder and I might be able to spot the difference. Otherwise I could easily script a TCX->json file conversion for the files I have.

I’ll be really surprised if you ride the downloaded files without additional smoothing. The fix gps approach is powerful but seems like lots of people have trouble using it. Understand there’s two splines (smooth math curve) the first built with all points, the second is made with points remaining after outliers from first spline are discarded. Each splines basis is how many points are used for each spline control point, so bigger is smoother spline. When you finish your route is rewritten with values interpolated from the second spline. There’s a separate altitude pass because altitude tends to need a heavy hand, location spline is for location aesthetics.

For example, if you set outlier of 1000m then (probably) no points are discarded and the second basis is your smoothing spline. Set basis to 3 and the smoothing will be almost nonexistent. 

Way i use it is set first pass to largish number, outlier distance to something small like 0.5m, second spline to smaller basis. The first smooth curve exists to toss outliers, the second is the smoothing after outliers are removed. So you can set second basis to three and it’s your original with just outliers removed. Yes? The actual values depend on the noise and point variance, you can’t smooth as aggressively if your sample spacing is noisy.

 I’d love feedback if it doesn’t do what you want. There are other route smoothers on the Internet, I’d like to hear what you end up using.

 It is definitely doing what I want, I can identify areas where I need improvement and design workouts. I have only had a direct drive trainer with power for 2 weeks and it's a whole new world of learning. Luckily I have my GPS logs of my 3 hour and 4 hour rides for virtual rides.

Have you tried riding with live map? Can it keep your attention? I really dislike the non video ui so end up gravitating to virtual rides with movies. I really like alucards umbrail pass rlv.

 Not yet,  I will have to have a play with this. I have no internet in the building with the trainer so hopefully the map is cached.

[Eric Christoffersen]

I think the video overlays make for a much better experience. But it means you need a video to overlay. You could just put any video on and it will overlay it.

Here's my current state, umbrail pass. Can see the short altitude view at the top center and the full course along the bottom.

If you just want climbing the videos work pretty great. They don't need internet to use, you must download them.

[Steve Edmonds]

Finally got around to a virtual ride, with map, without video. I will work on the video aspect. Haven't got around to the GPS smoothing yet as I am having a good look at the irregularities in GPX logs.

The smoothing doesn't affect me so much as I don't have a smart trainer that suddenly changes loading with grade, I just go faster or slower and alter cadence/gear to suit and match my normal speed up the hills. The good news is that I matched my actual ride time within a minute or so with a similar weariness of the legs.

[Eric Christoffersen]

Oh yeah. If no smart trainer then theres no real issue. On smart trainer the  jaggy gradients play hell with knees. Out of saddle to compensate and then resistance falls away to nothing. Torture chamber.

But you can just spin. A consequence free environment.

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---

You received this message because you are subscribed to a topic in the Google Groups "golden-cheetah-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/golden-cheetah-users/vAhgiHnY5LI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to golden-cheetah-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/c1a4c402-a443-4287-b82a-ba4efb571c3en%40googlegroups.com.

[Steve Edmonds]

Yes, if you know what your speed and cadence should be for any given slope you just fiddle with the gears to get the same in the virtual ride.
Now I'm happy riding will get back to the issue of the quality of the GPX files.