I’m not sure you’re following my point.
I’m saying that any assumptions about GE should be made by the user, not the software. kj=calories leaves those assumptions to the user. kj != calories takes those assumptions out of the hands of the user and makes them in a way that is opaque to the user
Better to leave them equal
Better to not estimate calories, and if the platform does, require the user set the multiplier.
It doesn’t really matter as long as the platform uses a consistent efficiency value any reasonable number is as likely to be wrong as a 1:1 conversion. I.e Zwift used 22.86%. Garmin is super annoying because they add basal metabolic rate estimate on top of the conversion. If you want 1 to 1 just take the kj number.
Looks like Strava is using 26.68% as their number in the newer example and 23.82% in the older ones . ( Strava only calculates kcal if the activity file did not provide it already). No idea why they changed that… But just use Kj as an estimate of kcal
That’s the whole point…TR isn’t providing it to Strava in a way that will populate kj into calories but other platforms are
@eddiegrinwald Any luck digging deeper on this question? Thanks!
I don’t know if this helps identify the source of the problem, but all my rides in TR are being pushed in from Strava. Until Tuesday, calorie numbers were essentially identical. Since Tuesday, Strava’s calorie numbers are consistently higher than TR’s calorie numbers. And Strava’s numbers are higher than what shows up in my Wahoo app (I have an Elemnt Bolt) – the Wahoo numbers are almost identical to TR’s numbers. All of this makes me think Strava is calculating its own calorie numbers even when the data being pushed to it includes calories (otherwise I would think the Wahoo numbers and the Strava numbers should be the same). In other words, seems like 100% a change on Strava’s end.
Huh, yeah I did pioneer-2 today and a few weeks ago. No ftp change in between and were within 3W avg / 10kJ for the hour according to strava. There’s a difference of over 100kcals now … from 876kcal to 996kcal. 10% is a big jump (but CICO, amirite? 
). My recent wahoo bolt rides have kcal and kJ at nearly 1:1 still.
I was going to say it’s not like I ever use the strava kcal field, but it does seem to carry thru into intervals.icu where I do like to trend weekly kcals. Seems problematic to have kcals computed differently from different data sources. It’s anew source of systemic bias. Would be in favor of TR updating their sync logic to set the kcals explicitly (assuming that’s why strava is computing it).
All I’ve been able to find is that we haven’t recently changed anything that would affect this. 
It sounds like other platforms might go about things differently, and Strava could have possibly made some changes recently as well. Have you heard anything back from them yet?
Thanks for checking!
Crickets, unfortunately. I just poked the support thread again yesterday, & will post here if I do get any info.
Just going from the size of the discrepancies it could be that Strava is now reporting “total calories” rather than “active calories”?
On outside rides with a power meter Garmin splits out “total” and “active” calories and it is the “active” that roughly equates to work done.
@eddiegrinwald Heard back from Strava support, which was just to refer to their longstanding calorie calculation (kJ estimate * human inefficiency factor). I pointed them to this thread & asked again about any changes visible from June 18.
Just adding another +1 to this thread. My Strava calorie counts suddenly got a lot higher recently. I wasn’t sure what was causing it.
I’m inclined to believe this. Best guess here is the issue is that TR has only ever submitted kJ data to Strava, not kcal data. So strava does some math on it’s end to fill it in on our/your behalf. It used to be very close to the kcal stat on the TR ride page. But Strava seemingly changed that computation recently, thus people noticing a much more significant discrepancy. I think what we’re hoping is that TR can explicitly supply a kJ field to Strava when synchronizing (as Garmin and Wahoo headunits seem to do) so they stop filling it in on TR’s behalf.
Excellent summary
Agree with your summary, @gcarver, but FWIW, re:
I’m not asking for TR to make any changes without understanding more from Strava about what knobs are turnable, & which represent durable API contracts going forward…
You or other TR folks don’t have any contacts inside Strava, @eddiegrinwald, by any chance?
Heard from Strava support again, who said they were aware of the issue & working on rolling back to the previous calculation. They pointed me at this thread, which didn’t contain any mention of the exact behavior we were describing here (increased calories relative to kJs), but the thread reports that the issue’s been fixed, & based on a quick survey of TR rides in my Strava feed from this afternoon, the old behavior does seem to be back.
I won’t see it myself until tomorrow (my ride this AM was still at the higher calorie calc). Anyone who rode in the afternoon or evening can confirm?
I’ll check later today and report back
However, I think there is still a legitimate, albeit lower priority if Strava has rolled this back, enhancement request for TR to supply the data to the Strava API in a manner similar to Garmin @eddiegrinwald
Let me know what you all find with your next activities in Strava.
However, I think there is still a legitimate, albeit lower priority if Strava has rolled this back, enhancement request for TR to supply the data to the Strava API in a manner similar to Garmin
I can let the team know that this is something people would like!
