Can updates please be tested before release?

Every update I have encountered in the past year has been riddled with bugs and glitches.

In the current update Zwift meetups no longer work.

  1. In a group zwift randomly forces riders to split off on other routes. ( Watopia, extended figure 8 reverse ) some of my group went the correct way and some veered off on a completely different course.

  2. On desert flats no matter how many watts we pull we are capped at 18-19km phr and keep bunny hopping in front of each other.

  3. On some rides, riders will float up into the sky in a straight line near the volcano.

Even on solo rides my w/kg since this update is all over the place. I’m doing a consistent wattage at 2.2w/kg but I’m bouncing as high as 8 and as low as 0. My friends have similar issue.

My last ride I did 45km hit finish and the app crashed to windows desktop - entire ride lost. All my level progress completely lost.

This is all super frustrating to get this sort of experience on a minor patch.

I think the solution is really to do some testing before release. Don’t just put code straight into production. I get one off issues with new code but this is every release and makes the app unplayable for a good week or two until things are fixed up.

Or alternatively every time you release a patch add 30 days free time to everyones account to make up for the 2 weeks we can’t play and 2 weeks as an apology thanks for waiting gift.

Cheers
Gaza

I think the chances of Zwift actually testing releases beforehand is near zero (if they haven’t done it by now then they never will).

I just wish they’d switch to a bi-monthly release schedule (every two months). Do we really need a monthly release with a one or two new frames and other minor features? Especially when it is inevitably followed up with a patch a few days later to fix all the new bugs.

At least switching to new updates every 2 months means there will be a longer period of stability and we won’t have to go through the usual fiasco every month.

Bugdates were funny the first dozen times but they’re really starting to wear out their welcome and I can’t help but think Zwift is losing customers over these updates.

6 Likes

There must be some testing, I just don’t know how stuff like “Er guys, the game literally doesn’t work on Mac” gets past it. Something is seriously wrong with the testing, the communication, the management or all of it.

As I’ve said a load of times, if the problem is actually finding this stuff then a very straightforward and instant solution is to pick a dozen volunteers around the world on different setups and give them early access to the proposed release for 24hrs. Tell them to use the game as normal and report back anything weird. Wouldn’t prevent everything of course, but would immediately highlight any showstoppers before they roll out to hundreds of thousands of paying customers. It would also allow for minor known issues to be presented with the release, rather than them being discovered by an increasingly frustrated community.

9 Likes

What do you mean stability … you mean longer periods where the instability issues are all known.

What is probably required is yes a move to less release for the ■■■■ of it on a particular day , possibly why bugs are so common on release to a more hotfix and single feature release focus to get these long term problems addressed . It seems to me that they are not factoring in an refinement and remediation required for any release so once something is released there is no retrospection and fixing issues , the team has already moved on to the next thing and have no process to look at what is deemed already delivered.

1 Like

Canary releasing to get user acceptance is a no brainer. I would be shocked if its not already done in some way , it clearly however is under resourced with numbers and/or not ensuring a good broad selection of users and set ups. Most likely ( as it is how they have done this publicly with other feature non-releases , is they give it to a set of well known or long term users and “mates” without making sure they cover the broadest set of use cases .

1 Like

Maybe, but then even that doesn’t fit with Eric at ZI publicly highlighting and demonstrating the ‘pairing screen speed lock’ bug immediately after release.

1 Like

Yes that doesn’t look good does it for assuming they just don’t have a broad enough UAT test group. Unless they are not listening to the feedback and just releasing anyway , which in the case of the latest MAC user issues is probably resulted in a very painful lesson learned (hopefully) .

1 Like

It is becoming increasingly apparent that nobody has the backbone to push back to senior management to say we need to slow the releases down to test properly as it is damaging both reputation and product.

Though, how senior management can let this happen on a monthly basis is shocking - Either dont care or are blinded by unrealistic targets or expectations.

2 Likes

The funny/sad/worrying thing about these latest rendering issues on Mac is that there were several different problems on virtually every hardware setup. Zwift were clearly scrambling to sort it out and initially tried to find a common denominator, but within a very short space of time it was obvious from the responses that it was happening on Intel integrated graphics of multiple generations, Nvidia and AMD dedicated graphics, and even eGPUs. Literally everything. So it was very obviously a much wider underlying issue than ‘Intel HD 4000 doesn’t like this’. How then, is it not spotted?

1 Like

Indeed . Whilst I was not impacted this one must have hurt Zwift right to the top . The forums probably only the tip of the tsunami of support tickets . This one surely cant have failed to trigger some sort of retrospective that should lead to improvements . We have stated the obvious not expecting to be given the gory details of whatever is happening in ZHQ but trust them to now look at this as more than just a tech issue.

1 Like

I’ve thought that numerous times over the last 18 months, remember when no one could use their iPad for a few weeks, that must have also been a big “whooops!” And surely got upper management to pay attention… but it continues to happen, so… yeah…

2 Likes

@shooj This is EXACTLY the sort of thread referenced another thread that seems to indicate that Zwift staff either don’t read or don’t care about what users say in the Forums. This sort of thread appears after EVERY release, and yet Zwift continues to put out very buggy releases every month. Surely if Zwift staff were actually paying attention to the Forum this would be the first issue to be addressed. So, you can tell us that Zwift staff read the Forum daily, but until the users see a change in behavior based on what gets posted here (especially the common/consistent issues brought up) we will continue to think that no one is actually reading or, even worse, no one cares.

4 Likes

Couldn’t agree more… Seriously Zwift´s behaviour is total lack of respect for us long term customers :roll_eyes:

1 Like

I’d say the most telling one of these kinds of bugs was the build that failed at launch because it was looking for resources in a certain QA manager’s home directory. So the real question is for the higher-ups to answer, what on earth is going on with QA at Zwift and what if anything is going to be done about it.

(Oh well, at least the updates are no longer being released on Friday evenings…)

1 Like

Dont count on that . They are going to have to hotfix this pairing screen bug at some point whenever it is QA passed . Looks at Watch …

1 Like

Fortunately I also have Zwift installed on an iPad so I should still have plenty of time to see if I can get up the Alpe in under 10 minutes…

1 Like

I, and I’m sure countless others, would happily take bleeding edge ‘risky’ releases to see new functionality first and report back bugs. It won’t covers issues related to version incompatibility between riders, but as Dave says you’ll still catch a number of bugs that currently get released to the entire userbase with gay abandon.

3 Likes

Yep I really don’t understand why they don’t approach more people to test releases. It’s clear they gave details to the press before so no reason why you can’t have some users on a test release even if they have to sign a NDA.
If they don’t have enough hardware combinations or resource to do internal QA then there are many members of the community that have multiple devices that run zwift which i’m sure would be very happy to test for them while getting the chance at a pre-look at the next update.

That and they also need a better release mechanism so they can halt/rollback a release asap if the worst case does happen and a game killer bug is released like we saw with the mac OS issues.

2 Likes

The frustrating thing is to look at the bugs squashed and see none of them prevent gameplay but the bugs created prevent gameplay for large communities. I would happily downgrade back years ago version even if that meant riding alone. It worked every time I logged in.

3 Likes
3 Likes