I want to share my thoughts on this, I’ve been so conflicted over the years on this feature.
In all honesty I wish all indoor platforms would send 0 for distance/speed because you’re not actually going anywhere. I did that for just a little bit of time many years ago and we were flooded with support tickets 
.
The other idea we’ve played with is to use a basic physics model and then on flat terrain estimate speed/distance.
The three issues I’ve had with this are:
- We’re not going to be able to make this completely accurate. I’m positive that people will be upset with the way we implement it. IE: on 200 watts I actually go 19 mph not 18 mph, please update the model.
- People are used to the way it is. There are a good amount of people who will get upset if we change it.
- Does it make anyone faster?
I know the answer to #1 and #2 is probably build out a physics model where people can tweak their inputs and terrain type and rolling resistance until they get something they like.
But, distance/speed tracking goes against our belief system of what makes you faster. So if we’re going to spend the effort, why not focus on the efforts that we believe will make you faster.
But…I’ve gone as far as asking universities querying us about projects to have them build us a microservice that accepts a .FIT file, height and weight and returns the fit files with virtual speed/distance. No one has accepted that project yet though :-D.
If anyone wants to build this and open source it (running it on Azure Functions would make it easy for us to integrate) that would be awesome. Maybe other companies could leverage it too and we’d all output the same numbers (which would be awesome).