ZwiftPower restoration update [March 2021]

You appear to have significantly overestimated any fixes to the ZwiftPower site, I’m still finding about 1 race in 3 is failing to have some or all of its ‘fit files’ processed.
Most recently yesterday afternoon’s Road to Sky ride where no-one in Category A, B, or C has had their data properly reported:
ZwiftPower Event ID=1946028

As for a “constructive” discussion, I think most would agree that your bullet point glibly saying you’re not asking for a “blue sky wishlist for a faster, stronger, able-to-leap-tall-buildings ZwiftPower” is the pretty much the definition of non-constructive since long-suffering users would almost universally accept a site that simply works in the most basic sense by processing all ride data, processing that data correctly, and doing so in a somewhat timely manner. We’re far too realistic to be expecting anything else or better and I imagine many are annoyed at even reading such a barbed point, I recommend you remove it!

It seems that for some riders events are not completing correctly in Zwift Power:

This happened to me twice, for the WTRL Duathlon Bike leg specifically 2 weeks in a row. No other events were impacted and I always get my data perfectly, often with fit file as well so it does not appear to be an issue on my end.

Here’s an example:
This is LIVE data filtered by my name.


Results view for the same event, again filtered by my name:

Event: Zwift Power - Zwift Duathlon League | WTRL - Opt 4 Bike
I did complete this event but it never showed up in my profile as a “Finished” event. This means the data from the event is not used in calculating my category and power numbers.
This seems to be a widespread issue. It did not happen to me this past week but I have seen reports that it is ongoing. Hopefully it can be rectified and these events actually marked as Finished so we can get recognition for them.

Are these courses always the same newer ones? Also, have you noticed if it’s unique to event type, i.e. race vs. group ride?

This is one I mentioned in the top on the original post. It’s one of the biggest issues in terms of reliability of data, so we are looking into it. Thanks for the detail.

Thanks for your thorough feedback. I appreciate the detail you’ve given in outlining the use cases you would have for ZwiftPower and the data it provides.

The questions of historical race data and more robust ways to get, interpret, and present that data is a double edged sword. On the one hand it’s very useful to see over a given period of time how certain riders are progressing/regressing in their results, and to spot areas that don’t line up with the overall trend of their performance. However, we have seen in the past that with more specific data as it relates to these users, people become the target of aggressive comments and behavior, and we do not wish to promote that on our platform in any way.

I don’t have a promise for you that we will or won’t do anything, but I can tell you that we are aware of the need for data as well as the need for privacy from a racer’s perspective.

I was attempting to concisely communicate that we are aware that the data isn’t always perfect, but that it’s better than it has been in the last few months. We aren’t chalking it up to being “fixed” by any stretch, and I agree with you that the core function of ZwiftPower is not what long time users are used in its current state.

When we say we’re not looking for the blue sky discussion, we simply mean that we all have things we would love to see in terms of what ZwiftPower can or can’t do; this isn’t unique to the community, we talk about it at the office quite a bit. This could be finishing features that were started and never realized, implementing completely new features or ways to show data or aggregate points, etc.

What we want to focus on in this thread specifically is to make sure that our understanding of what work to get ZwiftPower to simply working state is aligned with what you as a community are using it for. It would be easy for us to say that X feature needs fixing, but if Y feature is the bigger deterrent to actually using the site, we want to prioritize that.

1 Like

Would it be possible in the interests of clarity to help users understand what to feedback and what not to feedback on ( and therefore you getting a long list of wish restated issues ) would be for you to list out the issues that zwift has in backlog , those that are still to be verified or accepted , those already marked as known , in progress and/or done and indeed things that are marked wont fix or closed .

I would gladly contribute if it helped get issues into verified /accepted state if I saw this.
I expect the last 2 statuses are going to be ones you are going to be reluctant to post (especially here in this forum :smile: ) . Then again maybe you should look at how customer focused really you are if you dont , and you have seen how it goes , dont state it and we , the user community will make up the answer ourselves on your behalf !

I know you posted two areas known in the Original post , but maybe an update and more refined list now we have been feeding back for a week or so ?

1 Like

@xflintx You didnt acknowledge the post about restoring post raice discussion threads that was lost when the Zwiftpower forum was closed down and that functionality was not ported into this forum. This whilst not mission critical has had a big impact on the community and social aspect of post race. Is that an accepted issue Zwift are working to resolve ?

Ergo this ;

2 Likes

I don’t know where to search for this and don’t want to start a separate thread but is there a reason why the initial data posted after a race is way off? Example: Shortly after a race where I’ve done some pretty big sprints, Zwiftpower might show that my 15-second sprint was something like 4.8 w/kg, which seems way too low. When I check back later in the day, or maybe the next day, it will almost double.

I don’t think there are any users of Zwift & ZwiftPower unaware that the site isn’t perfect, but I also don’t believe anyone here still has any expectation of perfection given the distance between it and reality, so that ‘blue-sky’ comment comes across as a barbed attack on users who in actuality just want a company whose sole focus is pretty much ‘software development’ to do the most basic of basics: A working app and a working data site.
We’ve not had that yet.
Issues with the basics of data processing on ZwiftPower have been an ongoing issue since at least September/October 2020. 6-months later we seem no closer to any kind of solution (given that solving the issues necessitates understanding the causes of the issues, something that seemingly still hasn’t been identified at all!)
If Zwift doesn’t have developers capable of diagnosing problems with pre-existing code (and use of Zwift’s own API) on something as basic as a website over that time period perhaps you should simply write a brand new site which competent developers could do in far less than 6-months…

However since Zwift also seems incapable of even identifying itself which data hasn’t been processed properly (even though it’s identified on ZwiftPower as an incomplete data process) and needs telling about incidences then in addition to the one I mentioned yesterday the following events also haven’t had large swathes of rider data processed beyond a random snapshot:
ZwiftPower Event ID=1347855
ZwiftPower Event ID=1588153
ZwiftPower Event ID=1588171
ZwiftPower Event ID=1790127
ZwiftPower Event ID=1840905
ZwiftPower Event ID=1889570
ZwiftPower Event ID=1904449

2 Likes

Short answer is that (as I understand it) the initial results are pulled from “live data” direct from the event - if you dive into your power curve on any event straight after completion you’ll be able to see the power curve doesn’t appear to have any difference from 1 - 15 seconds or so. Post race there’s a behind-the-scenes task to pull down .fit files from Zwift for each rider in the event and do the necessary number crunching to fully update your power averages for each timeframe (which is why you’ll see things looking different -especially for short hard efforts- after a little while).

2 Likes

Do you happen to know why they do the random “snapshot” in the first place? It’s not accurate or seemingly particularly helpful for anything, or am I just missing some obvious reason why it’s a good/needed thing? (Wouldn’t be the first time, lol!)
Would also be a lot easier to spot where data hasn’t been genuinely processed for rides/riders if it was simply ‘blank’ until the FIT file was genuinely processed.

The short answer is that we have basically everything regarding ZwiftPower in the backlog, and we are just looking to see what’s most adversely affecting your current usage of ZwiftPower.

This thread isn’t for cherry picking on our end, either, it’s just to gauge the relative impact of these issues.

As for your question about the forums, it’s valid. However, I can tell you that the way our forum platform currently is built, that feature isn’t really viable within this space. I believe firmly in the post-race discussion, as I’ve had many good chats post-ride/race in person myself and always looked forward to them. We definitely want to have something like that on our forum, we just have to bring it back in a good way. The first thread we made a while back was clearly not the way to do it.

This is correct.

Let me preface all of this by saying this: I don’t write code here at Zwift. I have a good understanding of how the site works thanks to your collective comments and feedbacks, as well as well poking around on the backend.

From what I can put together this “snapshot” comes from how Live Data works. Really, as far as I can tell, missing .fit files are related to Live Data. That is to say, when live data works till the end of the race, the results seem to be more robust, and when Live Data doesn’t work as we would expect, then we have more issues with how the post-race results are displayed and how accurate they are.

Both of these things come from the way ZwiftPower does (or does not) communicate with the Zwift API for events. By default, when an event is created to display results, ZwiftPower will use Live Data during the event. When this communication stops for whatever reason, then you get your snapshot. The snapshot in and of itself is not necessarily a feature, more of an inadvertent result of that communication stopping. Ideally, then, when an event is finished, even when the Live Data hasn’t worked, you should still have your .fit file upload.

The part I don’t understand fully myself is twofold:

  1. I don’t know what causes the communication to the events to breakdown and to produce that “snapshot”
  2. I am not 100% sure that missing .fit files are the result of the breakdown of Live Data

These are more or less the two points I made in the original post, which is that we’re aware of them and they’re also the biggest issues (in my eyes) that we need to address with the site after taking care of some stuff in the backend.

1 Like

Hi Flint,

Thanks for asking for feedback here. Today I spotted an error on Zwiftpower which may not affect many “serious races” but impacts Zwift’s mass-participation events like the ToW and TdZ.

When an event has multiple route choices, Zwiftpower only seems to list the “A” route in the events list, and only shows the “A” profile in the course description. This made me take a few moments to question if I was looking at the right event today.

I feel this could be better if it did something like listing all the routes consecutively, or simply said something like “Multiple Routes”.

I’m biased about this because Zwift’s showpiece events were my first taste of Zwift competition and ZwiftPower, so I feel they may attract more Zwift and ZP newcomers, who you want things to work right for. I appreciate this may not be a priority - it probably depends on whether you see an increase in ZP registrations and visits around showpiece events like the TdZ and ToW or not!

Hope this helps.
Paul

So why did you close the Zwiftpower forums if you didn’t have a proper alternative in place?

So, wouldn’t the sensible first step be to simply disable the “Live Data” functionality altogether for the time being?
That way you can see whether FIT File processing issues are all then completely resolved, or not.
If they aren’t then you know there’s another root cause of the file processing issues.
If they are then you get the site back to proper file processing and can run diagnostics and code checks on the Live Data functionality, get it fixed, reactivate it for some ZHQ events to live test, ensure it’s actually working and then work on re-introducing it to the site and events as a whole.
Currently neither function is working which is the worst possible position. If you don’t know ‘for certain’ what effect one is or isn’t having on the other then you won’t get anywhere.

Hey everyone,

Thanks to all of you who took the time to detail out the areas of the site you’re still feeling aren’t giving you the race experience you’d like.

For some clarity, here are the next steps for us here at the office:

  • I will take all the information here and compare to what we already have checked in to work on or not. If we don’t have any work related to your reports, I will make it.
  • These reports are then fed into the correct pipelines based on work and perceived priority. I will certainly give your feedback when I check the work in, though I cannot promise how they will be prioritized or when they will be done.

What’s important is that they exist and that means our team will have visibility on it for the time being.

I’m going to close the thread as I feel we have reached a good point to do so.

Thank you again!

5 Likes