Is current performance problem related to the new system you introduced into the game?

Czekoludek

As we can see in patch notes for Markarth release:

ZOS_GinaBruno wrote:

Buffs and Debuffs

All Major and Minor Buffs and Debuffs have been hard coded to go through server validation sooner and more effectively as cached values, rather than abilities that exist on your character. This new system will also properly ensure that no two same buffs or debuffs can ever stack in effectiveness, rather than needing to be individually spot checked on each application. It will also prioritize duration-based sources rather than permanent ones, so you do not need to fear losing their effects. While doing so, we have also properly reigned in the power of these effects into our Set Bonus Efficiency (SBE) standard; Minor effects are now worth 2 set bonuses (previously between .4 and 4.2) and Major effects are worth 4 (previously between 2.14 and 13.3). The new values can be found below.

Many ppl were shocked why you want to move even more calculations to the servers when they are in such bad condition (because of TOS I cannot properly describe them as it would be definietely be vulgar). Since the change buffs are often not working properly and other game issues like lags and crashes are even more tedious then before.
Are you sure this new system is not responsible for recent issues? [snip]
Can we get a honest answer what went wrong and more important, why did you push this patch in that state on live?
@ZOS_GinaBruno you promised us better communication (multiple times) so maybe an eso live with tech team or long and honest post? It would be nice if ZOS treat their customers like humans and not like walking bags of money. A little decency would be nice.

[Edited to remove Bashing]

Thechuckage

Most likely upper corporate has given instructions for the PR side of things. Anyone can see the severe lack of communication on any other issues as an indicator. Most likely (IMO) let the dissenters yell into the void until they are tired. Plenty of other people to keep paying subs and buying from the crown store. And if the unhappy people just happen to wander off with their complaints...............

zvavi

Their idea is not wrong though, what they did should have made performance better (that way is definitely more efficient) thing is, I don't trust their programmers. Like seriously. Background story to why I don't trust them (for those that already read the background story in another thread, it got updated! Yey!):

lets take the recent group finder bugs as a reference.
group finder had a bug. that it would auto change the role of yourself to normal (affecting essentially those that queue for vet), even when you are doing exclusively vet content. Instead of fixing it, they have changed group finder, to auto register any cp player into the vet role, no matter what dungeon he gets into, normal included. This created a new bug. The one we know of low level characters not being able to join a normal dungeon, in a normal instance, but since group leader when group was queued was forced into vet role, all the low level characters weren't able to join a normal instance, blocked by "group" setting. now instead of fixing that, they have meddled with the group finder grouping itself. creating longer than expected queue times for all normal dungeons. Which was "fixed" again, and now we have 2 modes for each person, the individual one, and the group one, so now every dungeon you jump in through group finder will report normal (vet dungeons too, which is back to bug in square 1!!!) But now when you finish the dungeon, it puts you in your individual mode you were before it (Which doesn't even work half the time).

I mean, how hard is it to change your mode, into the one of the instance you are heading to? Vet dungeon? Make group setting vet. Normal? Make group setting normal. Shouldn't be hard. Makes sense. Fixes all related bugs.

Wing

radiating regen is still in the game, they dont care about performance.

eKsDee

The problem with trying to performance test an MMO is that nearly nothing impacts performance flatly (ie the same in all situations), rather performance impacts will scale to the number of players concerned, and can often depend on context. For this reason, potential performance-worsening features like this either need to be stress tested in a fake setting with bots (ie testing if performance just gets worse as more players are involved), or needs to be pushed to the live servers to let the players do the testing (ie testing is performance gets worse as players actually play with and around the new feature), and unfortunately the latter is the better test, and so these things will almost always hit the live servers before major performance issues are discovered.

The problem is the fact that Zenimax, for whatever reason, is either unable or unwilling to roll back changes like this when issues are discovered, they pile on a multitude of other changes on top of the problematic ones which only makes rolling those problematic ones back even harder, and any fixes are generally pushed back to the next quarterly patch, which can be 2-3 months out.

Patch needs to hit live for these issues to be discovered, okay, not great at all, but understandable. A week or two goes by, issues are discovered, great. PR is hurt by letting issues go through to live, but there's nothing Zenimax can really do about this, so they just take it. This is generally what happens, but this is where what should happen and what actually happens deviates.

What should happen is this. Zenimax goes back and internally determines what would hurt PR more: keep the patch live and let players just deal with the issues while they figure out a fix, or roll the patch back so players aren't forced to deal with the issues while they figure out a fix in the background. Whichever hurts PR the least is what they do.

What really happens is this. Zenimax keeps the patch live no matter what, forcing players to deal with the issues while Zenimax seemingly figures out a fix in the background, which may or may not be pushed to the next immediate patch the moment it's implemented, or may be pushed back to the next quarterly patch.

They can absolutely roll patches back, any source control platform (basically a platform that remotely keeps track of the history of a project, both to more easily allow multiple devs to work on the same project, and to roll the project back to an older version if a change didn't go as planned) allows for this, so they can roll it back to how it was before they merged the PTS build with the live build, and go from there. So they can absolutely roll the problematic patch back, they just don't.

The problem with rolling the problematic patch back is the fact that it can potentially stall the release of the next content update, but that should already be considered when they determine what would hurt PR more, and, IMO at least, if the issue is impactful enough to roll back a major content update, it's impactful enough to stall the release of the next content update. Just my opinion, your's may be different and that's okay.

So they could definitely be doing things better, but unfortunately problematic patches being pushed to live is just part of MMO development. Only way to fix that is to get an adequate amount of players on the PTS and testing the new features enough to discover issues, and all I have to say to that is good luck.