I realise this is an old-ish thread, but glad I found it, as it addressed my observation that my relative effort scores had about halved (for similar workouts) after changing from max HR-based to Custom HR zones. This was for activities with very similar proportions of time spent in each zone and very similar overall times. I had only tweaked the zones by a few HR beats. But the relative effort scores had about halved.
My theory is that it is do with Max HR value itself. When HR zones are set using max HR, then the max HR exists as a set number and is used in whatever calculation Strava uses for relative effort. In my case 170. But when setting custom zones, there is no max HR set - (you can’t set the upper end of z5 to a maximum figure.)
If the algorithm for relative effort uses not just the % time in zones, but also max HR, perhaps as a denominator for part of the calculation, then it being a bigger number would cause the relative effort score to be much lower. (eg it wouldn't surprise me if the calculation uses the ‘default’ of 220 when using custom zones - because no max has actually been set.)
Anyhow, to fix the activities, I changed back to HR zones based on a max HR. Then changed activity type on each of the affected activities over the last couple of weeks to a different type (I used ‘gravel’, but it doesn’t matter), and then back again to ‘ride’. This caused the RE scores for those activities to be recalculated, back to similar values to what I’d expected based on prior activities. And so the Fitness and Freshness graph, which had shown an odd declining trend, looks as I’d expect again, with consistently calculated data.
I’ve also now tweaked the max HR down a beat or two to get the strava calculated zones closer to my desired custom zones, which are within a beat or so, close enough...
I would encourage Strava to look at this - it feels like an oversight in the RE calculation algorithm, maybe due to what I’ve suggested. Allowing users to also set a max HR value when in custom zones, would likely solve it.
I realise that is some senses this doesn’t matter - because the RE score in itself is just a number and it’s more the trend in it that’s important But when I’ve got several years worth of data all calculated on the same basis, it would be preferable not to have a significant discontinuity in that trend!
Anyhow, I thought I’d post all this just in case it helped anyone else who was wondering why RE scores changed after moving to Custom HR zones, along with my work-arounds, as well as hoping Strava would address it.
