Zwiftpower FIT Files for TTT

Yesterday I found myself alone in the pen then. After a while,when everyone from my team had already flown away, I could see the people in the pen. So I decided to do a FTP test and make it count in Zwiftpower. I did it, increased the W/kg quite a lot, from 4,55 to 4,75, and finish quietly the ttt to register the result. Well, there’s only live data in the results, and live data says that I did not this effort… I downloaded the fit from the activity and it works and the data is consistent with my ftp bump (which I saw in the screen).

Really disappointing.

1 Like

Part of an ongoing problem still being addressed. Follow the thread below for updates and drop an email off to zwiftpower@zwift.com with event details, your profile details so they can make sure its fixed up when they address the root cause.

1 Like

Thanks to Dave Kinshott in the WTRL TTT group on Facebook, I got a hint about why this happens…

I looked at how far people in my team went after crossing the finish line in yesterday’s TTT. Finish line was at 24.6 km and all 8 riders finished the TTT:
Ryan 24.7 km = green
Gary 24.6 km = green
Andy 24.7 km = green
Mario 24.6 km = green
Mark 26.2 km = blue
Anders 39.5 km = blue
Kim 42.9 km = blue
Timmy 46.0 km = blue

Then I looked at the previous TTT. Finish line at 42.4 km…
Mario 42.6 km = green
Anders 42.6 km = green
Andy 42.7 km = green
Gary 42.7 km = green
Dave 42.9 km = green
Mark 44.1 km = green
Robert 47.8km = blue
Kim 56.4 km = blue

4 March. Finish line at about 30.0 km.
Andy 30.0 km = green
Gary 30.0 km = green
Timmy 30.4 km = green
Anders 30.8 km = green
Simon 31.0 km = green

11 Feb. Finish line at about 29.0 km.
Ryan 29.8 km = green
Mark S 29.8 km = green
Dave 29.3 km = green
Mark C 31.4 km = blue
Anders 31.9km = blue
Simon 37.3 km = blue
Kim 41.8 km = blue

This gives us a workaround (ride an extra few km to get fit file analysed).

But Zwift could look into resolving the issue completely so it works also for those who don’t want to ride several extra kms (and to avoid that people are gaming the system by stopping immediately after the finish, e.g. since they don’t want to have their data analysed and get promoted to another category…)

2 Likes

Interesting…

Though that could still just reflect a timing issue. If the issue is caused by too many people finishing a ride at the same time (leading to ZP getting overloaded and dropping the results), then those who ride further may be saved simply because they are delaying submitting their results until a time when there’s less incoming data.

Would be interesting to see if simply hanging around for 10 minutes before submitting, without actually riding any further, produces more reliable results.

1 Like

To save people the effort of an extended ride I’ll make it clear this bug is not related to distance travelled and the above data is simply a misleading anomaly.

1 Like

I have no reason to doubt you but if you are so certain please explain your reasoning…

If it was an anomaly, I believe it would not survive the analysis of all those 28 rides, would it? Then I guess the hypothesis would need to fail quite soon if I analyse a few more of those… I even try to make it blinded by first checking the distances and afterwards checking the colour of the lightning bolt.

So let’s take a few more samples from the two TTTs ahead of 11 Feb:

4 Feb, 28 km:
Casper 28.1 km = green
Anders 28.7 km = green
Mark S. 29.0 km = green
Matthew 30.1 km = green
Dave 34.5 km = blue
Ryan 38.4 km = blue

28 Jan, 32 km:
Matthew 33.0 km = green
Timmy 33.1 km = green
Gary 33.4 km = green
Mark S. 34.5 km = blue
Mark C. 37.3 km = blue
Andy 40.4 km = blue
Anders 45.2 km = blue

Hypothesis still not rejected…how many more samples do we need? :wink:

Edit: Well, let’s do another one from the ZRL race yesterday (6 Apr) in Douce France, 24.2 km:
Andy 24.35 km = green
Evert 24.47 km = green
Jack 24.50 km = green
Scott 24.53 km = green
David 27.1 km = blue
Anders 29.54 km = blue

Well since James says it’s not about “distance ridden after finish line”, it’s probably some variable with high covariance with that. For example “time recorded after finish line”, “start delay time” or more likely a combination of those two. More about that after Thursday’s TTT…

2 Likes

I agree it seems to work; I’ve tried it last few events I thought I might struggle to get the .fit file for… I am guessing it’s more to do with having a bit of time separation between the event finish line and saving the session but just my guess. Either way the info SHOULD help Zwift fix it sooner rather than later!

Same here, ended the ride straight after the WTRL TTT on Tuesday and only have a green bolt. I will have to try and ride a few extra k’s and see if it works.

Having this discussion, in another group, shows a potential link between the Lag reported by Zwift power and the fit file processing. Any ideas?LatencyLagWtrlTTT|690x478

I haven’t had a TTT fit file process in weeks.

Idk what it is, but I’ve resigned myself to never having that data

Well, zwiftpower typically shows a “lag” in the TTT results until it has processed the fit file. So I’m not convinced that this lag is the cause of the issue, maybe it’s more an indicator of whether it used live data or a fit file (just like the green and blue bolts).

Yesterday we did an experiment in my team. We had a start delay of 12.5 minutes. So I told the others that after we cross the finish line, we continue riding for some minutes and finish at about 11, 12, 13, 14 and 15 minutes and afterwards check if we find any correlation between the fit file analysis and whether we did less or more than the start delay.

But it failed:
Finish time 41:30, distance was 27.5 km
Gary - 28.4 km - 46:04 = green (his headset didn’t work so he couldn’t hear my suggestion on riding longer…)
Ryan - 33.3 km - 53:04 = green
Andy - 35.1 km - 54:17 = green
Mario - 31.5 km - 54:37 = green
Anders - 33.2 km - 57:21 = green
Timmy - 34.3 km - 57:23 = green
Dave - 32.7 km - 58:06 = green

So even if we rode a few kilometers extra and rode for up to 16 minutes after the finish line, none of us got the file processed. A lot more than our start delay. Weird! :man_shrugging:

I noticed that in the team finishing right after us, Herd of Lions, 4 of 8 had a blue bolt. Their start delay was 8.5min, I believe, and among those who got the blue bolt, the guy who rode shortest time did about 12min after the finish banner.

After all…now I’ve feel I’ve lost a clue about how and why this happens - but I believe we know for sure that continuing to ride for a good while after the TTT will increase the chance of having the fit file processed in the end. :wink: