Moving and Elasped Time For Strava Segment

Page may contain affiliate links. Please see terms for details.

BSRU

A Human Being
Location
Swindon
Just noticed the last two tacx virtual rides I have uploaded, have an elasped time much longer than the moving time for strava segments.
For the previous year or so the elasped/moving time have been almost identical, even though on longer segments the ride has been auto paused for interruptions.
But now the elasped time seems to be including paused time aswell.
This is causing a problem now as I cannot compare segments from new rides to segments from old rides.
Strangley the overall elasped time seems to still be using the overall moving time, i.e. not including pauses.
 

Legs

usually riding on Zwift...
Location
Staffordshire
Just noticed the last two tacx virtual rides I have uploaded, have an elasped time much longer than the moving time for strava segments.
For the previous year or so the elasped/moving time have been almost identical, even though on longer segments the ride has been auto paused for interruptions.
But now the elasped time seems to be including paused time aswell.
This is causing a problem now as I cannot compare segments from new rides to segments from old rides.
Strangley the overall elasped time seems to still be using the overall moving time, i.e. not including pauses.

Simple answer is not to stop. If you're doing virtual rides, and you want to have comparability of segment times, why would you need to stop?
If you do really need to stop (for a pee, or whatever), why not do it on a descent?
 

CXRAndy

Guru
Location
Lincs
Can't you just deduct the pauses from the total ride time. Quite easy to highlight the stops. It then tells you the duration

Or don't stop riding
 

Alex321

Guru
Location
South Wales
It has always had different values for elapsed and moving time for real rides, I'm surprised it didn't for virtual rides.

TBH, it should include paused time, because every time you pause does give to a bit of rest.
 

Dogtrousers

Kilometre nibbler
Segments and virtual rides really don't mix.

Here's why. (This is for TACX trainers. Others may differ).

It's a bit long and complicated (and boring) I'm afraid.

A real ride outside records your time and location every second (say). It also records stuff like HR, power, speed etc if you have sensors for these. This is great because your location changes every second. If you move your bike a little bit, the recorded location changes a little bit. Over a minute you will have 60 (say) records, each with a different location. This all makes perfect sense. We end up with a series of points like this. I'm going to assume that you are recording Heart Rate, for the sake of illustration:

13:00:00 Location=L1, HR=H1
13:00:01 Location=L2, HR=H2
13:00:02 Location=L3, HR=H3 and so on where L1, L2 and L3 are all different locations as measured by your GPS and H1, H2, H3 your heart rate as measured at that time.

Things are VERY different for a virtual ride. A virtual ride consists of a series of waypoints, like a route. They are close enough for the route to have a recognisable shape, but there are significant gaps between them. Let's say for the sake of argument that they are 30m apart and you are going at a virtual speed of 27km/h. That means you will pass one route point every four seconds.

Now, your TACX trainer still records your data every second. It records your HR, power, virtual speed etc. as you pass through the various route points. But it does not interpolate between route points. So lets say that for four seconds your nearest simulated route point is R1. It will stamp four records with the location of R1. Then after that your nearest route point becomes R2, it will stamp the next four records with R2 which is 30m away. So it will look like this.

13:00:00 Location=R1, HR=H1
13:00:01 Location=R1, HR=H2
13:00:02 Location=R1, HR=H3
13:00:03 Location=R1, HR=H4
13:00:04 Location=R2, HR=H5
13:00:05 Location=R2, HR=H6
13:00:06 Location=R2, HR=H7
13:00:07 Location=R2, HR=H8 where R1, R2 are route points.

Do you see what's happening? Instead of a smoothly changing location, your location is jumping from route point to route point.

If this was a real ride your location would be smoothly changing showing all the intermediate locations between R1 and R2. But in the trainer records it jumps from one route point to the next.

This means that - apart from for plotting a map of where you have "been" (actually where you haven't really been) the locations recorded by a TACX trainer as you do a virtual ride are almost meaningless. Or at least they require very sympathetic handling.

Now along comes Strava and tries to match a segment to this rather weird series of locations, leaping from point to point.

If the segment is quite long and the route points quite close together this won't make all that much difference. But if this isn't the case things start to get crazy. Quite how crazy really depends on looking into the mind of the Strava matching algorithm, and that's not easy to do.

First of all it will only match route points that are inside the segment. So lets say your first route point R1 is 25m from the start point of the segment, and your last route R2 point is 25m from the end. It measures the time you took between those route points. You've just trimmed 50m off the total segment length. You get a crazy average speed for the segment!! But your max speed will still only be the 27 km/h that you were going between the two.

This madness is what enabled me to claim a Strava KoM with an average speed of 70kmh (just to prove a point). Tis bonkers.

Now, this doesn't answer why it's started including your paused time in your moving time. That's another question. But what it does is illustrate that the Strava segment matching algorithm is designed to deal with real GPS data recorded outside. When you confront it with data from a trainer it can have a meltdown.

Without examining your gpx files I can't tell you exactly what's going on but the reason could be that maybe strava have tweaked their algorithm, or TACX have tweaked their recording method, or you did something slightly different ... and out pops an unexpected result.
 
Last edited:
Top Bottom