Road Grand Tours / Strava / Trainer Road power anomaly

OK, so this is very niche, but bear with me.

I did a race on Road Grand Tours, which then automatically synced to Strava. It gave me an average of 260 watts and a normalised power (taken from intervals.icu) of 270. Note that this tied in pretty closely with the back-up I recorded on my Wahoo Bolt.

However, when the ride synced from Strava to the Trainer Road calendar, the normalised power was boosted to 277, and the average power bumped all the way down to 230.

Strava: Follow Martin on Strava to see this activity. Join for free.
Trainer Road: Log In to TrainerRoad

(Excuse language; I got dropped by the lead group on the first lap and ended up solo-ing the rest.)

Any thoughts on what might be causing this? I don’t think it can be anything to do with counting zeroes as there were so few of them.

I had pretty much exactly the same in relation to a race on RGT today. Strava and RGT say 267 watts average, Trainerroad says 230. Odd.

Thanks - worth saying that 2 friends who did the race with me had the same issue too.

Yeah same here, not in races but the other rides I’ve been doing on there. TBH I don’t expect TR to actually do anything about power discrepancy caused by a 3rd party app as it’s not their problem but interested in the replies all the same.

Yes I’ve also had the same issue, I’ve also noticed when I export the files so I can bring them into Garmin connect that it doesn’t even register power there at all!

I originally thought it might be linked somehow to the drafting benefit you receive?

It’s possible, but my understanding is that the drafting benefit is really just an instruction that tells the program to make it need fewer watts to go the same speed.

But if there’s an issue with Garmin Connect sync too, then it does point to it originating at RGT’s end. Might drop them an email.

Yeah I think it might be at RGTs end, you can export the original or email it to yourself, it’s a .fit file but it fails when trying to import to Garmin. So all I can do is export a gpx from strava which will import but without any power data.

Sounds like the same issue I had, RGT and Strava reported 241W average but TR reported 210. I haven’t looked at the export - does it maybe include the beginning/end where you idle, warm up, and cool off?

No, looks like just the race to me. Unless that information is somehow hidden in the file. Though that wouldn’t explain the increase in normalised power either…

That’s a very good point. Did you submit an issue to TR yet? I guess I should later.

1 Like

No, will do.

I reported it tonight after it happened again (that’s now every ride on RGT).

I think I figured out the issue however. If you look at the FIT file using https://www.fitfileviewer.com/ you can export it to a CSV. Then you can sum up the power and get an average from it. Based on my FIT file, it seems TR takes the total accumulated power over your ride and divides it by total time in seconds, which in my case lined up perfectly for the number TR reported (215W). However the FIT file actually had fewer entries than total ride duration in seconds (it occasionally skips an entry for a second, the total difference was around 400 entries for today’s 3600 second ride). If you divide the total accumulated power by total number of entries (i.e. get the actual average), you end up with the number that both RGT and Strava report.

Bug filed with the explanation, hope they can fix it quickly. You can easily confirm above with your RGT or TR FIT exports. I haven’t yet taken a random TR workout to find out if there is a similar pattern with skipped readings from TR or Zwift. It is possible they don’t drop for a split second or maybe back-fill and pad the numbers. So this may also be an issue in RGT or in how RGT handles missed entries. But definitely a problem with TRs calculations.

4 Likes