The Truest Solution to ESO Server Performance

Markytous

Many years ago, before the fear of data breach or hackers ruining the gameplay, the game's packet calculations (your healing, damage, character spreadsheet, etc.) were handled by your Elder Scrolls Online client! The game started development in 2007, being Zenimax Online Studios' first project, our beloved ESO! Some time between 2018-2019 (I can no longer recall specifics), our server calculations were reconfigured from Client checks to Server checks to increase security of the game. While this prevented a sizable amount of player tampering with code and calculations, it came to the detriment to the server performance. You see, the game was designed to be played with Client-sided packet calculations in mind over a decade before the change. We've never managed to restore performance back to what it was in 2017 and before. It has nothing to do with your sets, skills, procs or what have you! It is simply that the calculations per player in instances has become much more complex, increasing the load that the server has to handle at any given moment.

Cyrodiil can be enjoyed once again in the way we did long ago! The only solution I can propose is that we work together to make Client-Side Server Calculations a reality again. This is no easy task because of the security issues that may come with it. I am willing to put up with ANY and ALL Anti-Cheat Softwares to enjoy my beloved ESO and Cyrodiil again as I did in those years; wide eyed with wonder exploring the beautiful and treacherous landscapes of Cyrodiil. My aim for this is not to be controversial. I love ESO and I want to see it return to its former glory!

Now, Vengeance? Thats super fun. I don't propose scrapping that at all. But imagine if our Alliance-Locked CP Cyrodiil Server and Vengeance Server could run side-by-side without a hitch! No lag, no performance issues, just straight up fun! We know what works, guys. We played it! Client-Sided Server Calculations can do it. There is no shame in returning to it if we can do so in a safe and secure way. I hope this message reaches all of you well. If nothing results from it, at least I can say that I also did my very best in contributing towards progress in one of my absolute favorite games of all time.

LadyGP

I hate to be the person to bring this down but this won't happen. Zos hasn't really been able to handle cheaters for a while and this would make the issue 100000 times worse. I'm with you... but... they just won't do it.

They havent even handled the AFK macro people killing cyro right now... no way they will bring anti cheat in.

Markytous

LadyGP wrote: »

I hate to be the person to bring this down but this won't happen. Zos hasn't really been able to handle cheaters for a while and this would make the issue 100000 times worse. I'm with you... but... they just won't do it.

They havent even handled the AFK macro people killing cyro right now... no way they will bring anti cheat in.

I know... I know... But I just want to say it to at least get the solution out there for more friends out there. If I say "I know its not going to happen" it discourages it from happening... even if its not going to happen to begin with... lol It is a dream and I've been prepared to face the face that it will remain a dream. Maybe if we all share our desire to turn back the clock but make different steps something new can happen. Heck, the calculation changes were done quite impressively fast! Perhaps its not that complicated to change them back. I know block and sprint got changed to communicate with the server differently. So many memories lol

AngryPenguin

Markytous wrote: »

Many years ago, before the fear of data breach or hackers ruining the gameplay, the game's packet calculations (your healing, damage, character spreadsheet, etc.) were handled by your Elder Scrolls Online client! The game started development in 2007, being Zenimax Online Studios' first project, our beloved ESO! Some time between 2018-2019 (I can no longer recall specifics), our server calculations were reconfigured from Client checks to Server checks to increase security of the game. While this prevented a sizable amount of player tampering with code and calculations, it came to the detriment to the server performance. You see, the game was designed to be played with Client-sided packet calculations in mind over a decade before the change. We've never managed to restore performance back to what it was in 2017 and before. It has nothing to do with your sets, skills, procs or what have you! It is simply that the calculations per player in instances has become much more complex, increasing the load that the server has to handle at any given moment.

Cyrodiil can be enjoyed once again in the way we did long ago! The only solution I can propose is that we work together to make Client-Side Server Calculations a reality again. This is no easy task because of the security issues that may come with it. I am willing to put up with ANY and ALL Anti-Cheat Softwares to enjoy my beloved ESO and Cyrodiil again as I did in those years; wide eyed with wonder exploring the beautiful and treacherous landscapes of Cyrodiil. My aim for this is not to be controversial. I love ESO and I want to see it return to its former glory!

Now, Vengeance? Thats super fun. I don't propose scrapping that at all. But imagine if our Alliance-Locked CP Cyrodiil Server and Vengeance Server could run side-by-side without a hitch! No lag, no performance issues, just straight up fun! We know what works, guys. We played it! Client-Sided Server Calculations can do it. There is no shame in returning to it if we can do so in a safe and secure way. I hope this message reaches all of you well. If nothing results from it, at least I can say that I also did my very best in contributing towards progress in one of my absolute favorite games of all time.

We don't have to imagine anything. We had great or near great performance from 2015-2019. If ZOS could deliver good performance ten years ago, they can do it today.

Why they aren't is the question to put to ZOS. We know they can do it because we've seen them do it. (yes, they delivered decent to good performance with server side calculations) Cyrodiil performance used to IMPROVE during MYM events.

These performance issues are all ZOS and nobody else can fix them except ZOS. People just need to stop running cover for bad business practices. And to be clear, vengeance mode is an example of bad business practice. Instead of fixing their game ZOS is changing the game we have into something totally different.

Markytous

AngryPenguin wrote: »

Markytous wrote: »

Many years ago, before the fear of data breach or hackers ruining the gameplay, the game's packet calculations (your healing, damage, character spreadsheet, etc.) were handled by your Elder Scrolls Online client! The game started development in 2007, being Zenimax Online Studios' first project, our beloved ESO! Some time between 2018-2019 (I can no longer recall specifics), our server calculations were reconfigured from Client checks to Server checks to increase security of the game. While this prevented a sizable amount of player tampering with code and calculations, it came to the detriment to the server performance. You see, the game was designed to be played with Client-sided packet calculations in mind over a decade before the change. We've never managed to restore performance back to what it was in 2017 and before. It has nothing to do with your sets, skills, procs or what have you! It is simply that the calculations per player in instances has become much more complex, increasing the load that the server has to handle at any given moment.

Cyrodiil can be enjoyed once again in the way we did long ago! The only solution I can propose is that we work together to make Client-Side Server Calculations a reality again. This is no easy task because of the security issues that may come with it. I am willing to put up with ANY and ALL Anti-Cheat Softwares to enjoy my beloved ESO and Cyrodiil again as I did in those years; wide eyed with wonder exploring the beautiful and treacherous landscapes of Cyrodiil. My aim for this is not to be controversial. I love ESO and I want to see it return to its former glory!

Now, Vengeance? Thats super fun. I don't propose scrapping that at all. But imagine if our Alliance-Locked CP Cyrodiil Server and Vengeance Server could run side-by-side without a hitch! No lag, no performance issues, just straight up fun! We know what works, guys. We played it! Client-Sided Server Calculations can do it. There is no shame in returning to it if we can do so in a safe and secure way. I hope this message reaches all of you well. If nothing results from it, at least I can say that I also did my very best in contributing towards progress in one of my absolute favorite games of all time.

We don't have to imagine anything. We had great or near great performance from 2015-2019. If ZOS could deliver good performance ten years ago, they can do it today.

Why they aren't is the question to put to ZOS. We know they can do it because we've seen them do it. (yes, they delivered decent to good performance with server side calculations) Cyrodiil performance used to IMPROVE during MYM events.

These performance issues are all ZOS and nobody else can fix them except ZOS. People just need to stop running cover for bad business practices. And to be clear, vengeance mode is an example of bad business practice. Instead of fixing their game ZOS is changing the game we have into something totally different.

I am highly satisfied to agree with you on this. While Vengeance is fun to me, if we restored the servers back to Client-Side Calculations we wouldn't need Vengeance at a "test" and it could simply just be a funny game mode on the side. Of course, without performance issues, Vengeance would never exist. I would obviously choose the classic Cyrodiil as they were before the server code changes over the current timeline with endless "Year of Performance" years haha. In that regard, I can see Vengeance as an advertisement of sorts; "We're doing something, guys!". I agree with them that they are doing something to make Server-Sided Calculations work with Vengeance, but the Client Sided ways of the past are the truest solution. No need to rebuild the game's calculations, skills, and etc,.

Conclusion: Client-Sided Server Calculations will save this game. It is now a question of if Zenimax wants to save it or not.

AngryPenguin

Markytous wrote: »

AngryPenguin wrote: »

Markytous wrote: »

Many years ago, before the fear of data breach or hackers ruining the gameplay, the game's packet calculations (your healing, damage, character spreadsheet, etc.) were handled by your Elder Scrolls Online client! The game started development in 2007, being Zenimax Online Studios' first project, our beloved ESO! Some time between 2018-2019 (I can no longer recall specifics), our server calculations were reconfigured from Client checks to Server checks to increase security of the game. While this prevented a sizable amount of player tampering with code and calculations, it came to the detriment to the server performance. You see, the game was designed to be played with Client-sided packet calculations in mind over a decade before the change. We've never managed to restore performance back to what it was in 2017 and before. It has nothing to do with your sets, skills, procs or what have you! It is simply that the calculations per player in instances has become much more complex, increasing the load that the server has to handle at any given moment.

Cyrodiil can be enjoyed once again in the way we did long ago! The only solution I can propose is that we work together to make Client-Side Server Calculations a reality again. This is no easy task because of the security issues that may come with it. I am willing to put up with ANY and ALL Anti-Cheat Softwares to enjoy my beloved ESO and Cyrodiil again as I did in those years; wide eyed with wonder exploring the beautiful and treacherous landscapes of Cyrodiil. My aim for this is not to be controversial. I love ESO and I want to see it return to its former glory!

Now, Vengeance? Thats super fun. I don't propose scrapping that at all. But imagine if our Alliance-Locked CP Cyrodiil Server and Vengeance Server could run side-by-side without a hitch! No lag, no performance issues, just straight up fun! We know what works, guys. We played it! Client-Sided Server Calculations can do it. There is no shame in returning to it if we can do so in a safe and secure way. I hope this message reaches all of you well. If nothing results from it, at least I can say that I also did my very best in contributing towards progress in one of my absolute favorite games of all time.

We don't have to imagine anything. We had great or near great performance from 2015-2019. If ZOS could deliver good performance ten years ago, they can do it today.

Why they aren't is the question to put to ZOS. We know they can do it because we've seen them do it. (yes, they delivered decent to good performance with server side calculations) Cyrodiil performance used to IMPROVE during MYM events.

These performance issues are all ZOS and nobody else can fix them except ZOS. People just need to stop running cover for bad business practices. And to be clear, vengeance mode is an example of bad business practice. Instead of fixing their game ZOS is changing the game we have into something totally different.

I am highly satisfied to agree with you on this. While Vengeance is fun to me, if we restored the servers back to Client-Side Calculations we wouldn't need Vengeance at a "test" and it could simply just be a funny game mode on the side. Of course, without performance issues, Vengeance would never exist. I would obviously choose the classic Cyrodiil as they were before the server code changes over the current timeline with endless "Year of Performance" years haha. In that regard, I can see Vengeance as an advertisement of sorts; "We're doing something, guys!". I agree with them that they are doing something to make Server-Sided Calculations work with Vengeance, but the Client Sided ways of the past are the truest solution. No need to rebuild the game's calculations, skills, and etc,.

Conclusion: Client-Sided Server Calculations will save this game. It is now a question of if Zenimax wants to save it or not.

Except calculation were server side 2015-2019 and performance was, mostly anyway, pretty good. That means ZOS can, if they chose to, deliver good performance again today. Server technology has come a long way in the last decade.

Vengeance is a distraction. ZOS needs to just fix the game they originally created. We know they can do it because they've done it in the past.

Stamicka

Yea I would definitely put up with something like Riot’s Vanguard for better performance. A lot of people have privacy concerns with that sort of thing though.

Technically Xbox could afford to have more client side calculations since you can’t download 3rd party programs on them. It seems like ESO on Xbox still handles most things on the server side anyway. Maybe there’s more of a reason they choose not to go that route.

Unfortunately, I think Vengeance is a dead end. I saw lag creeping back in simply with the introduction of weapon skill lines. In its current state it’s still a bit too simple to be fun long term and they really need to add more complexity. More complexity seems to mean more lag though, and that puts us right back to the performance we have now. I’m hoping they figure out a new approach.

Arrow312

Firstly, yes, they would definitely have to change something. Vengence is not the great saviour that many believe him to be. The campaign is far too simple and just gets boring in the long run. During the event, battles for castles sometimes took a very long time. How long should they last in Vengence? Then the effect that such a campaign has on the rest is often not taken into account. Upgrades, sets and many other things are no longer needed in PvP. So why should people who only play PvP for the most part still play PvE, let alone spend a lot of money on crowns? As an addition, this campaign could find some supporters. As the only one, it will be worse rather than better. But that's just my opinion.

randconfig

Markytous wrote: »

Many years ago, before the fear of data breach or hackers ruining the gameplay, the game's packet calculations (your healing, damage, character spreadsheet, etc.) were handled by your Elder Scrolls Online client! The game started development in 2007, being Zenimax Online Studios' first project, our beloved ESO! Some time between 2018-2019 (I can no longer recall specifics), our server calculations were reconfigured from Client checks to Server checks to increase security of the game. While this prevented a sizable amount of player tampering with code and calculations, it came to the detriment to the server performance. You see, the game was designed to be played with Client-sided packet calculations in mind over a decade before the change. We've never managed to restore performance back to what it was in 2017 and before. It has nothing to do with your sets, skills, procs or what have you! It is simply that the calculations per player in instances has become much more complex, increasing the load that the server has to handle at any given moment.

Cyrodiil can be enjoyed once again in the way we did long ago! The only solution I can propose is that we work together to make Client-Side Server Calculations a reality again. This is no easy task because of the security issues that may come with it. I am willing to put up with ANY and ALL Anti-Cheat Softwares to enjoy my beloved ESO and Cyrodiil again as I did in those years; wide eyed with wonder exploring the beautiful and treacherous landscapes of Cyrodiil. My aim for this is not to be controversial. I love ESO and I want to see it return to its former glory!

Now, Vengeance? Thats super fun. I don't propose scrapping that at all. But imagine if our Alliance-Locked CP Cyrodiil Server and Vengeance Server could run side-by-side without a hitch! No lag, no performance issues, just straight up fun! We know what works, guys. We played it! Client-Sided Server Calculations can do it. There is no shame in returning to it if we can do so in a safe and secure way. I hope this message reaches all of you well. If nothing results from it, at least I can say that I also did my very best in contributing towards progress in one of my absolute favorite games of all time.

I absolutely agree. VERY FEW TIMES I would take the risk to my privacy with a kernel-level anti-cheat program, but I would do it for ESO. @ZOS_GinaBruno @ZOS_Kevin Please ask the devs if they would be willing to consider moving calculations back to the client to fix server performance and giving us an anti-cheat program.

Zodiarkslayer

I think the recent changes to the client have actually made all performance issues even worse.
Mine freezes for about 15s everytime I shut the game down. Customer service is not helping.

If anything, they should revisit their stance on modernizing UI and graphics. The no optional changes caused my CPU to draw much more power and cause a lot more strain on the system overall. The visual improvements however are underwhelming, when I am feeling generous, imperceptable when I am honest.

Operativ

Oh, no, no, no. This is a very short-sighted idea from a technical perspective. Cheaters are omnipresent, and the only real solution to the performance is to optimise server-side calculations by refactoring certain code snippets, and improve the server hardware along with its usage.

No MMORPG should have any sort of client-side authority. That's single player games' territory. No MMORPG should have to use proprietary anti-cheat solutions (Easy Anti-Cheat, BattlEye, or something so horrible like Vanguard). At best, they're glorified spyware that don't stop the vast majority of dedicated cheaters.
Let me ask you a question - would you hand the keys of your house over to a monitoring agency for "additional security"? The trick is that not every "security employee" that would come by your house would be honest. Some would definitely steal a bit of this, and a bit of that. The same with your personal data, as well as serving as a potential admin-level backdoor to modify your Operating System files.

Take a look at Ravenwatch Cyrodil. It's lag-free, because "ballgroups" don't run there. The problem is the amount of stacking calculations as well - like tens of instances of damage shield & healing on the same person, multiplied many times over as there are many group members in a "ballgroup".

If better latency is desired, then Zenimax should direct (or hire) their architects at the server-side configuration to refine it. I could even help ZOS with that, but I don't have many hours to invest into more commercial tech projects.

. . . . . . . . . . . . . . . . . . . . . . . . . . .

While we're at it, the client-side could use performance improvements to utilise the power of modern configurations, as well as use Vulkan (or even DirectX 12) instead of the current DirectX 11 implementation; and DirectX 10 should be deprecated and unsupported if it already isn't.

randconfig

Operativ wrote: »

Oh, no, no, no. This is a very short-sighted idea from a technical perspective. Cheaters are omnipresent, and the only real solution to the performance is to optimise server-side calculations by refactoring certain code snippets, and improve the server hardware along with its usage.

No MMORPG should have any sort of client-side authority. That's single player games' territory. No MMORPG should have to use proprietary anti-cheat solutions (Easy Anti-Cheat, BattlEye, or something so horrible like Vanguard). At best, they're glorified spyware that don't stop the vast majority of dedicated cheaters.
Let me ask you a question - would you hand the keys of your house over to a monitoring agency for "additional security"? The trick is that not every "security employee" that would come by your house would be honest. Some would definitely steal a bit of this, and a bit of that. The same with your personal data, as well as serving as a potential admin-level backdoor to modify your Operating System files.

Take a look at Ravenwatch Cyrodil. It's lag-free, because "ballgroups" don't run there. The problem is the amount of stacking calculations as well - like tens of instances of damage shield & healing on the same person, multiplied many times over as there are many group members in a "ballgroup".

If better latency is desired, then Zenimax should direct (or hire) their architects at the server-side configuration to refine it. I could even help ZOS with that, but I don't have many hours to invest into more commercial tech projects.

. . . . . . . . . . . . . . . . . . . . . . . . . . .

While we're at it, the client-side could use performance improvements to utilise the power of modern configurations, as well as use Vulkan (or even DirectX 12) instead of the current DirectX 11 implementation; and DirectX 10 should be deprecated and unsupported if it already isn't.

"At best, they're glorified spyware that don't stop the vast majority of dedicated cheaters.", that's not true? Dedicated cheaters only go to up to the kernal level, a rare 1% go beyond that using hardware and modified/published drivers for Windows to bypass easy anti-cheat.

So a kernal level anti-cheat will stop 99% of hackers. For the other 1% you can just use server logic to validate calculations periodically to confirm if the behavior of a player's client has been modified, and gather further evidence via in-game player reports, also I think Windows can/does remove drivers used for hacking purposes once they're aware of them...

I'm no expert though, but from my understanding, that's how the best modern day anti-cheats work. The biggest concern would be how much time and resources they'd have to invest into making this happen, and of course, concerns for privacy and performance for older machines from the community.

Though there's infinitely many vulnerabilities to exploit in any number of drivers to remake their cheats and keep going, so maybe the resource investment in modifying the game to detect cheaters long-term would be a concern, but I can't see how it would be more expensive the buying up more physical space, paying for cooling for a massive room of hardware for the server, replacing the hardware/maintenance, and so on just to keep the server performance as it is currently and stripping the game of features like strict pet limits and population limits in popular zones, all of which makes the experience for the players worse over time, meaning less players invest in the game, meaning even more money lost over time.

So either way you look at it, moving the calculations back to the client would be the best long-term solution, and they likely didn't have the more advanced anti-cheats we have today back when they made the decision to move calculations to the server in the first place. Of course, that's just speculation, but it makes sense to me

Renato90085

change all 0.3sec/1hit skill be 2s/1hit in game

randconfig

Renato90085 wrote: »

change all 0.3sec/1hit skill be 2s/1hit in game

Pretty sure those abilities are single target, so the calculations are minimum compared to an AoE like Wall of Elements that has to iterate over all nearby players to see if they're close enough to take damage, in addition to then performing the damage calculation for each player hit.

I'd say one of the worst offenders is ball groups casting AoE heal over time effects that also stack with others casting an identical AoE heal over time effect, causing an massive number of calculations for the server.

Elric_665

The Second they try to force me to install some Kernel-Drivers for a Anticheat, is the Second I cancel my Sub and uninstall the Game.

drkfrontiers

I will uninstall the game the day there is any surveillance on my pc required to play a game on, to support part of a game I spend 1% of my time with.

Markytous

Operativ wrote: »

Oh, no, no, no. This is a very short-sighted idea from a technical perspective. Cheaters are omnipresent, and the only real solution to the performance is to optimise server-side calculations by refactoring certain code snippets, and improve the server hardware along with its usage.

No MMORPG should have any sort of client-side authority. That's single player games' territory. No MMORPG should have to use proprietary anti-cheat solutions (Easy Anti-Cheat, BattlEye, or something so horrible like Vanguard). At best, they're glorified spyware that don't stop the vast majority of dedicated cheaters.
Let me ask you a question - would you hand the keys of your house over to a monitoring agency for "additional security"? The trick is that not every "security employee" that would come by your house would be honest. Some would definitely steal a bit of this, and a bit of that. The same with your personal data, as well as serving as a potential admin-level backdoor to modify your Operating System files.

Take a look at Ravenwatch Cyrodil. It's lag-free, because "ballgroups" don't run there. The problem is the amount of stacking calculations as well - like tens of instances of damage shield & healing on the same person, multiplied many times over as there are many group members in a "ballgroup".

If better latency is desired, then Zenimax should direct (or hire) their architects at the server-side configuration to refine it. I could even help ZOS with that, but I don't have many hours to invest into more commercial tech projects.

. . . . . . . . . . . . . . . . . . . . . . . . . . .

While we're at it, the client-side could use performance improvements to utilise the power of modern configurations, as well as use Vulkan (or even DirectX 12) instead of the current DirectX 11 implementation; and DirectX 10 should be deprecated and unsupported if it already isn't.

Then they should either do that or rebuild the game where this is happening. Tweaking the amount of code flying around in the same exact infrastructure is yet another band-aid fix that doesn't return the game to the tech that was tested for years and worked for years. Something, ANYTHING, needs to happen. Also Ravenwatch is lag-free... until a certain amount of players appear and play normally. This should never be the case. A fully population locked Cyrodiil (and IC) instance should NEVER experience any latency issues of any kind up to its absolute capacity (that is the function of the server capacity). Truth is, Grey Host's server capacity is still too high realistically because the current packet infrastructure probably can't handle a server cap of 80 players. ZOS is doing what is realistic about allowing players to engage in the content while mitigating the negative response to lag and setting the increasingly lower server pop caps. We've got to the point where if the pop caps got any lower it would be obvious (it already is) and Vengeance's much higher pop cap is a "cat out the bag" situation now.

But I agree. Something needs to happen with regards to getting server tech up to speed. GW2, WoW and others can support massive battles and they were designed to. ESO was designed to support them, too. However with the switch to Server-Sided calculations we are now playing on an iteration of the game which didn't get the years of dev/engineer time in. It's either security, or unplayable game. I'd rather the decision to be "play functioning game or spare my computer's security" instead of "play dysfunctional game that still has cheats or... not". See what I mean? I'm not pro-anti cheat software but the hurdle the devs are sidestepping should have been resolved way back before the release of the game. The damage has been done.

LadyGP

A lot of people who have never worked in any kind of coding/development job giving very interesting (and strong) opinions on what "HAS TO HAPPEN".

I'll say this... at the end of the day... the board room is going to make the decision.

There is a reason why over the last 5 years you see game studio after game studio putting out a massive open letter appology to their fans after they shipped a game that was an absolute train wreck upon launch. All of that can be directly linked back to the leadership of that company - it all comes down to profits/timelines.

You're not going to get a rework of an engine. Actually... with the recent layoffs and cancelation of BB you probally are about 1% closer to having a new engine rework now than you were last year. If you're looking for copium it might be... leadership needs to figure out the next source of revenue for Zeni and maybe relaunching an ESO 2.0 (new engine, new graphics, modernized infastructre) might be it (it isn't but I'm just playing along for the heck of it).

Anyways, no one is going to sign off on a new engine rework. You can only do so much with the tools you have. Getting new tools, new infastructre, etc costs money and is dang near impossible to get leadership to sign off on (trust me I've been in the boardroom briefing leadership on how why our devs need XYZ and how it will undinw our technical debt and in the long run give our company hundreds of thousands of dollars). What do you think leadership said? They never like the timeline, they never like the initial investment.. they ALWAYS want the bandaid.

Who owns Zeni? you htink Microsoft (look at their trackrecord) is going to let ZoS stop content, fix the issues, bring in new rousources, work on a new engine, migrate to unreal (could you imagine if that happened oh my I'd cry happy tears), literally none of that is going to happen.

So what is a dev left to do - we are left to spend our time doing something we know is only a half butt solution but it's the only thing in our tool box because leaderhsip won't buy us a new toolbelt full of new shiny tools.

It sucks but this is the reality of development as a whole. We can sit here and list out 100 reasons why other options would be better but at the end of the day the powers that be have to make money - that is literally their only job.

This is just the harsh reality of it. The sooner you realize this the sooner alot of the decisions made over the past 7 years make a whole lot more sense.

(I have an insane amount of respect of Zos, Zos devs, the leadership at Zos, Matt etc. None of this is a knock on them or anything).

exiars10

Fully agree with @Operativ.
No to anti-cheat software as @Elric_665 wrote, too.

@Markytous
What you posted in not true history of Cyrodiil.
How do I know? Because I left the game after Summerset release (June 2018), and I was even ESO+ subscriber, and didn't even play (except daily log in) until April of this year.

Cyrodiil performance was already bad in 2018, but after Summerset it was horrible. Just check my history of messages where I reported horrible performance and "millon" bugs with my favourite PvP class - Warden who was literally broken, but not because it was OP, because you simply couldn't play normally.

Please tell me you and some other posters here how Cyrodiil was great in 2018.
I apologize if I misunderstood you.

Markytous

exiars10 wrote: »

Fully agree with @Operativ.
No to anti-cheat software as @Elric_665 wrote, too.

@Markytous
What you posted in not true history of Cyrodiil.
How do I know? Because I left the game after Summerset release (June 2018), and I was even ESO+ subscriber, and didn't even play (except daily log in) until April of this year.

Cyrodiil performance was already bad in 2018, but after Summerset it was horrible. Just check my history of messages where I reported horrible performance and "millon" bugs with my favourite PvP class - Warden who was literally broken, but not because it was OP, because you simply couldn't play normally.

Please tell me you and some other posters here how Cyrodiil was great in 2018.
I apologize if I misunderstood you.

Even in 2018 the amount of players in the servers was higher and the servers could sustain it better than today on un-refreshed servers since launch. It has been degrading but more and more code was set to be checked on the server-side from the client-side. Efforts to do this have been scattered about patch notes over the years and server population caps have seen reductions several times leading up to now. Summerset patch is 2018 and it was better then than it is today. I don't see how you can be offended by indicating that performance was better then than now. Have you checked Grey Host Cyrodiil now that you've been back? We could still make very sizeable keep sieges and battles in Morrowind and still be able to play the game.

exiars10

No way it had better performance.
Warden was literally unplayable / broken in Cyrodiil.
At Keep / Castle / bridge fights I had on average ~15-20 fps.
That was Vivec PC EU server (the main 30 day campaign).

Do I have to link old threads to show how horrible was?

Here it is my almost last post when I quit game on 16th June 2018 after canceled subscription:
https://forums.elderscrollsonline.com/en/discussion/comment/5225044#Comment_5225044

Even PvE was plagued for months in 2018 before Summerset. I reported it:
https://forums.elderscrollsonline.com/en/discussion/400325/game-performance-is-terrible-this-evening

Sorry, what you write is a rose tinted nostalgia.

And no, I don't play Cyrodiil in 2025 except Vengeance last time.

Markytous

exiars10 wrote: »

No way it had better performance.
Warden was literally unplayable / broken in Cyrodiil.
At Keep / Castle / bridge fights I had on average ~15-20 fps.
That was Vivec PC EU server (the main 30 day campaign).

Do I have to link old threads to show how horrible was?

Here it is my almost last post when I quit game on 16th June 2018 after canceled subscription:
https://forums.elderscrollsonline.com/en/discussion/comment/5225044#Comment_5225044

Even PvE was plagued for months in 2018 before Summerset. I reported it:
https://forums.elderscrollsonline.com/en/discussion/400325/game-performance-is-terrible-this-evening

Sorry, what you write is a rose tinted nostalgia.

And no, I don't play Cyrodiil in 2025 except Vengeance last time.

Oh I thought you had videos for me to watch and make comparisons with today's experience. You'd have to experience the lag and low population caps of today in order to make an informed opinion of this topic. What class is "broken" doesn't have any bearing with the performance of the server/connection and especially now that classes basically don't exist its even more irrelevant. I do have video evidence of battles from 2018 with WAY more players seen sieging than today but they are also videos which showcase gameplay that is being "altered" so it may not be good to share those ones here.

ajkb78

Arrow312 wrote: »

Firstly, yes, they would definitely have to change something. Vengence is not the great saviour that many believe him to be. The campaign is far too simple and just gets boring in the long run. During the event, battles for castles sometimes took a very long time. How long should they last in Vengence? Then the effect that such a campaign has on the rest is often not taken into account. Upgrades, sets and many other things are no longer needed in PvP. So why should people who only play PvP for the most part still play PvE, let alone spend a lot of money on crowns? As an addition, this campaign could find some supporters. As the only one, it will be worse rather than better. But that's just my opinion.

You know Vengeance is not an end state, right? It's a performance test. It already shows significant change since the initial Cyrodiil Champions PC-only test, but it is only a test: the fact that it happens to be playable is almost irrelevant.

Having said that, while it's not the end state and shouldn't be, there are quite a lot of aspects of it that have been very refreshing. While it's not a long term solution to have *no* build diversity available through sets, upgrades etc. it has been very good to see edge case builds rendered impossible in the test campaigns. (You know the ones I mean and unless you run them you almost certainly hate them as much as I do...) So I hope they find a decent balance between reintroducing some build diversity and neutering *those* builds.

Markytous

ajkb78 wrote: »

Arrow312 wrote: »

Firstly, yes, they would definitely have to change something. Vengence is not the great saviour that many believe him to be. The campaign is far too simple and just gets boring in the long run. During the event, battles for castles sometimes took a very long time. How long should they last in Vengence? Then the effect that such a campaign has on the rest is often not taken into account. Upgrades, sets and many other things are no longer needed in PvP. So why should people who only play PvP for the most part still play PvE, let alone spend a lot of money on crowns? As an addition, this campaign could find some supporters. As the only one, it will be worse rather than better. But that's just my opinion.

You know Vengeance is not an end state, right? It's a performance test. It already shows significant change since the initial Cyrodiil Champions PC-only test, but it is only a test: the fact that it happens to be playable is almost irrelevant.

Having said that, while it's not the end state and shouldn't be, there are quite a lot of aspects of it that have been very refreshing. While it's not a long term solution to have *no* build diversity available through sets, upgrades etc. it has been very good to see edge case builds rendered impossible in the test campaigns. (You know the ones I mean and unless you run them you almost certainly hate them as much as I do...) So I hope they find a decent balance between reintroducing some build diversity and neutering *those* builds.

To put it very plainly I absolutely 100% agree with you. I really just wish the Vengeance path wasn't necessary to "revitalize" PVP in this game. It's more about server infrastructure and code rules more than about balance... However the balance of Vengeance was fantastic and I was 1vXing out there when supposed "hardcore PVPers" were saying it was a no-skill zerg fest. Made me feel like maybe some players weren't as good as they thought they were in PVP? idk lol EDIT: And before anybody says I was probably abusing NB or Sorc, no! I was 1vXing on a Dragonknight! It was so satisfying.

Desiato

This is false.

The history in the forums clearly shows crippling lag began a couple of months after launch as players levelled and learned how to play. Though it ramped up quickly, it was actually granular as the record shows.

At first players blamed "the lighting patch." Some would claim "light particles were server side" which was absurd, yet this myth perpetuated. Brian Wheeler investigated and explained it was actually just players levelling and characters unlocking new skills and passives. But also, players learned how to play and the ball group play style emerged and was copied.

Learning to play was a LONG process for the ESO community. Before target dummies were introduced, only a tiny percentage of players could weave light attacks. I know this from fighting in Cyrodiil because almost no one did it. I did from the start and would constantly deal with accusations of cheating by players who couldn't fathom it.

The big cheating scandal known as Zazeer-gate happened in 2016 and performance didn't change after. They did not spontaneously re-engineer the relationship between the client and the server. This would have been EXTREMELY obvious to every single player had it happened.

What they did was change where values were stored in memory to prevent previously known methods of memory hacking to work, added obfuscation and user-mode security checks. This is what people who have analyzed ESO's memory have determined through their work.

What you've heard is a myth invented by players that is not backed up by fact.

A brief history of performance issues in Cyrodiil:
https://forums.elderscrollsonline.com/en/discussion/658524/the-glory-days-of-cyrodiil/p1

Markytous

Desiato wrote: »

This is false.

The history in the forums clearly shows crippling lag began a couple of months after launch as players levelled and learned how to play.

The big cheating scandal known as Zazeer-gate happened in 2016 and performance didn't change after. They did not spontaneously re-engineer the relationship between the client and the server. This would have been EXTREMELY obvious to every single player had it happened.

What they did was change where values were stored in memory to prevent previously known methods of memory hacking to work, added obfuscation and user-mode security checks. This is what people who have analyzed ESO's memory have determined through their work.

What you've heard is a myth invented by players that is not backed up by fact.

A brief history of performance issues in Cyrodiil:
https://forums.elderscrollsonline.com/en/discussion/658524/the-glory-days-of-cyrodiil/p1

And yet the servers were able to house more players then than now. The only difference even under these conditions would be the server pop-caps. The game has gone from more playable to less playable over time. After the server replacement that ZOS did, players were celebrating for maybe 6 months after this happened then performance dipped again after that. Of course, this was after code was recalibrated to be more server-sided and population caps had been reduced.

https://forums.elderscrollsonline.com/en/discussion/495264/server-lag-explained-server-calculations-skill-actions-are-not-the-main-reason-for-lag/p2
https://forums.elderscrollsonline.com/en-gb/discussion/586142/console-server-vs-client-calcs

Update 25 has lines that involve making combat fundamentals (Blocking) more efficient for the server. This was also done to Sprinting in another update. None of these changes have helped mitigate issues that the server is experiencing today. Point is, the gameplay has deteriorated beyond what it was before and before was more playable than now. We have played more performant Cyrodiil in the past with higher population caps already. If you believe it was all bad lag from the start, then that doesn't rule out that it has gone from bad to worse than bad.

Desiato

The first pop cap reductions happened very early. No one knows when exactly because it wasn't documented. By by late 2015/early 2016 2016/early 2017 I had guild members who had attempted to quantify the player cap in various ways and had come up with about 150 per faction. Who knows what it actually was, but it was nowhere close to 600.

Recent evidence shows 80-100.

Over time, more features were added, the TTK was constantly increased and sustain became easier and easier to the point where we are now where players and group can easily reset back to full health and resources, so fights drag on and on and on.

Play styles also changed. We went from ball groups in 2014 to AOTP and Chuck Norris in 2017 running with 2-4 24 player groups.

This was offset for a time by the publicized server upgrades, in addition to continuously lowering the player cap to the ghost-town levels of today.

The addition of a new feature last year, Scribing, had a noticable impact on server performance.

It's important to understand not all players load the server equally. A player standing on a keep wall heavy attacking will contribute very little to server load compared to a highly mobile, high APM player executing abilities on cooldown. A group of such players will be even more impactful per-player because of buff stacking.

There are no easy solutions or they would have done it by now.

Markytous

Desiato wrote: »

The first pop cap reductions happened very early. No one knows when exactly because it wasn't documented. By by late 2015/early 2016 I had guild members who had adapted to quantify the player cap in various ways and had come up with about 150 per faction. Who knows what it actually was, but it was nowhere close to 600.

Recent evidence shows 80-100.

Over time, more features were added, the TTK was constantly increased and sustain became easier and easier to the point where we are now where players and group can easily reset back to full health and resources, so fights drag on and on and on.

Play styles also changed. We went from ball groups in 2014 to AOTP and Chuck Norris in 2017 running with 2-4 24 player groups.

This was offset for a time by the publicized server upgrades, in addition to continuously lowering the player cap to the ghost-town levels of today.

The addition of a new feature last year, Scribing, had a noticable impact on server performance.

It's important to understand not all players load the server equally. A player standing on a keep wall heavy attacking will contribute very little to server load compared to a highly mobile, high APM player executing abilities on cooldown. A group of such players will be even more impactful per-player because of buff stacking.

There are no easy solutions or they would have done it by now.

Now on this I absolutely agree with you. Everything you say here I do not object. Even if we do not agree on how this all happened, we do absolutely agree that the decay has been a painful endurance run for many long-term fans of this game. And yep ball groups show up and crash the whole server while if we just have fun group ups things can be tolerable. Solutions will be even harder with Microsoft now overhead aiming at maximizing profits over producing a better game. Since the game has been profitable "as is" they would not dare touch the status quo and fixing performance in any meaningful way would likely "cost money". It really is a sad situation for all of us.

exiars10

You again rewrite history no matter how much we show you?
And yes, broken class has everything with performance as you simply didn't read anything I posted back then.

Here we go again - just two screenshots from my public Steam account:

Posted: 29 Dec, 2017 @ 1:35am
I was standing behind Faregyl Farm, killed some Flawless Conqueror, and look at performance - "glorious" 40 fps! When encountered players behind farm, it sunked to 20 fps.
https://steamcommunity.com/sharedfiles/filedetails/?id=1248082368

Posted: 31 Dec, 2017 @ 10:55am
Siege at Alessia Castle. Almost cinematic 24 fps with 130 ms, and I was just standing and shooting from ballista. Going down, and it was 15 fps.
https://steamcommunity.com/sharedfiles/filedetails/?id=1251521819

@Desiato
Thank you for your effort documenting Cyrodiil performance. I posted screenshots to show how horrible was it in the past, too. Markytous has typical rose tinted nostalgia.

Markytous

exiars10 wrote: »

You again rewrite history no matter how much we show you?
And yes, broken class has everything with performance as you simply didn't read anything I posted back then.

Here we go again - just two screenshots from my public Steam account:

Posted: 29 Dec, 2017 @ 1:35am
I was standing behind Faregyl Farm, killed some Flawless Conqueror, and look at performance - "glorious" 40 fps! When encountered players behind farm, it sunked to 20 fps.
https://steamcommunity.com/sharedfiles/filedetails/?id=1248082368

Posted: 31 Dec, 2017 @ 10:55am
Siege at Alessia Castle. Almost cinematic 24 fps with 130 ms, and I was just standing and shooting from ballista. Going down, and it was 15 fps.
https://steamcommunity.com/sharedfiles/filedetails/?id=1251521819

@Desiato
Thank you for your effort documenting Cyrodiil performance. I posted screenshots to show how horrible was it in the past, too. Markytous has typical rose tinted nostalgia.

No Desiato actually posted some video evidence which is useful for conducting this kind of conversation. There's a big difference between the the way you and Desiato are discussing. Two screenshots and yourself complaining about performance doesn't exactly prove much. Its a general concensus that performance had declined over time and the developers admit to this by replacing the old servers when they did a couple years back. Big difference.

Mojey87

i dont understand some comments on this thread, against anti-cheat!!
well if you dont cheat then what are you afraid of? what the impact that anti-cheat will do to your game except it will make it safe from any who like to take advantage over other players.