TR feature request: Strava suffer score to TR tss conversion

For those of us riding outdoors without power meters, I was thinking that TR could crest an algorithm to convert the Strava suffer score of your ride (or what ever they call it these days) to TR tss. This should be pretty simple to accomplish and more accurate than using RPE to assign tss.

I don’t think Strava documents exactly how they calculate suffer score (or “Relative Effort”) anywhere. So then its not so easy. Its not difficult to derive a TSS-like measure from heart rate data so TR could do that (thats what I did for


I have typically found the Strava suffer score to be significantly less accurate than RPE. I’m not sure what they base it on, but they don’t seem to do a very good job.


We considered the possibility of using HR to estimate TSS, but we ultimately decided that that self reported RPE was a better approach since heart rate can vary based on a lot of different factors and that could ultimately decrease the accuracy of your data.

On, the r-squared between my TSS and HR-load is 0.82. This is with rides outside [Powerpod], inside [KK Road Machine] and inside [Stac zero], none of which have been cross-calibrated. There are a couple of rides in which the TSS and HR-load differ hugely, but in general, the HR-load seems to me to give a pretty good estimate of TSS – though, of course, it would be nice to know the confidence intervals around David’s equation.

Using RPE, suppose that I am uncertain whether to set the ride as tempo or as uncomfortable, a difference between RPE = 5 or 6. The difference in estimated TSS is 14 per cent.