Update on Cyrodiil Performance & Upcoming AOE Tests

scottii

Guild Wars 2 also uses the same engine as TESO - HAVOK.

JasLaguna wrote: »

PvE and PvP will never be able to be balanced the same way. If they all have the exact same skills, at least one will always be out of balance.

That said, please take a look at how Guild Wars 2 addressed this issue by balancing each separately. In PvE, specific skills that would break PvP would get adjusted ONLY within a PvP setting. In PvP, that same skill could have any of the following modifiers: added effect, subtracted effect, different cooldown timer, or an increased or decreased resource cost. If it needs to be tweaked in PvP, then tweak it just for PvP.

And as in Guild Wars 2, the tooltip would point out and highlight whatever difference it would have within a PvP setting, so that it was clear and easy to understand what the difference would be between the two settings.

I certainly don't want ESO to become GW2, and I love ESO immensely more, but GW2 got it right when they split PvE and PvP game balancing. In the vast majority of games that don't do this, balancing for one wrecks it for the other and it's always been like that in every game that keeps PvE and PvP skills the same. When you have a game with both PvE and PvP elements, being able to adjust the skills based on which setting you're in frees you up to balance the one without destroying the other.

The fact that this test is being conducted ONLY within Cyrodiil both proves my point and demonstrates my example! DO THAT! Adjust it within Cyrodiil! However you need to! But leave it out of the PvE realm!

Tommy_The_Gun

"During the times that any of these tests are active, we will be awarding double Alliance Points for anyone active in Cyrodiil."

I appreciate that, but the problem is that people who "spam AOE" (aka "Ball Groups") - sit on millions & billions of AP. They don't care about that "bonus" (for them it is almost nothing). So, basically, groups that cause biggest stress on the server in PvP will not participate in the live test. They will suspend their play for the duration of this test. In other words - you will not get "true" results.

AP bonus means that PvP-ers wont play, but PvE players (usually "low APM" players) will go to Cyro instead - to level up their AW skill lines and get skill points from AW ranks - just to never play PvP again on their mains.


I would suggest to give "something" different, more valuable (Transmute Crystals, Crown Gems, or even Crowns), that will attract PvP players more. AP is the last thing they want / need (as said before, they have millions of them).

Dusk_Coven

Tommy_The_Gun wrote: »

"During the times that any of these tests are active, we will be awarding double Alliance Points for anyone active in Cyrodiil."

I appreciate that, but the problem is that people who "spam AOE" (aka "Ball Groups") - sit on millions & billions of AP.

Not sure that AP boosters even really need AoEs. From what I understand they just ride back and forth mostly.

Ragnarock41

I hope you can reach a sensible solution. No matter how good or bad the solution you find, the feedback will be definitely negative regardless of if you try to fix this or not, so you all better be mentally prepared. That being said, even implementing something very limiting like friendly fire would be a better solution than very large cooldowns, considering this will require the entire class roster rebalanced around it.

x48rph

Thank you for the candid post about what is going on and what your trying to do to correct it. I know the currently list of tests are not going to be popular and I would hope that at the end of the day, it can be fixed without something so drastic but even if it can't I still will support you doing so because while people might keep pointing out that the fluid and dynamic combat in pvp is a hallmark of eso, combat in cyrodil is anything but fluid and dynamic thanks to the ridiculous lag. Fighting to get abilities to go off and feeling like your in a slide show are much, much worse issues then having things be a little less dynamic with aoes. I trust you will take a long hard look at classes and abilities before anything goes live permanently and adjust accordingly. Thank you for actually focusing on the issues in cyrodil for a change.

SpiderKnight

This is great news. I'm going to be there just for the entertainment of watching templars struggle.

Jaimeh

Stop destroying combat in the name of performance. ESO's fast-paced combat is one of the greatest things about the game; please do not introduce any more cool-downs, this will be a slippery slope, and will be terrible for gameplay. Why not explore other ways to fix the lag that's associated with a big population in one area. Every update we're given these reasons why lag exist (APM, AoEs), and it's disconcerting to see that your solution is essentially to slowly deconstruct the combat system bit by bit (all the skill standardization, and how the classes are basically have blended into each other), because you can't figure out other ways to fix your coding and hardware issues. You're killing the game in the name of fixing it. There's some poetic irony right there. Please reconsider, ZOS devs, surely there must be another way to go about this.

Lake

Radaney wrote: »

World of Warcraft has this exact same issue. Large scale PvP in World of Warcraft is actually impossible because of the amount of server commands, the server simply breaks.

Upgrading the servers has not even worked for Blizzard either so I'm going ZOS the benefit of the doubt here.

Blizzard also introduced GCD to try and make it better and it didn't work.

...

People saying 'just upgrade the servers' are just uneducated crybabies who have no idea what they're talking about, this server stability with overwhelming server commands is not exclusive to ESO, and ESO even has it worse than WoW because WoW does have cooldowns and still gets these issues. This is not an easy issue to fix.

Yeah, Blizzard has a gargantuan amount of resources & yet one single Twitch streamer holding an "event" can still crash an entire server.

waterfairy

What about people like myself who already have a problem maintaining resources in pvp? The meta teams can spam abilities, the rest of us normal players can't. Instead of limiting things to help server performance you should get more servers and lower the max number of players on each one

EntropicLynx

This may have already been mentioned (I've only read the first handful of replies to this thread, rather than all 15 pages of them)...

There is a fundamental difference between various types of AoE abilities.

There are Point-Blank abilities that happen targeted on the caster themself.
There are Target Area abilities, a.k.a. Ground Target abilities that are targeted at a location the caster directs them to.
There are Target+Nearby abilities that are targeted on the enemy they are cast on and affect others within a radius of that first target.
There are AoE abilities that have no "over time" effect to them and are designed to be potentially "spammable" as long as resources are available.
Ther are AoE abilities that have an "over time" effect to them and are designed to be only used when the effect timer has expired or is about to.

You *CANNOT* treat all of them equally by introducing a 3-second GCD for all AOE abilities. If you do this, not only will you see the high-speed and dynamic nature of combat in ESO (something that is, for many of us, the primary reason we enjoy this game rather than just about any and every other MMO on the market), but you would also drastically alter the nature of how character builds are approached, and not in a good way. What you will almost definitely find is that people will deliberately avoid AoE abilities that were specifically designed and intended to be spammable in favor of AoE abilities that have an "over time" effect as their primary component. This will be problematic in general, but will absolutely *CRIPPLE* classes that have their main spammable ability as an AoE ability.

If you actually push these kinds of "tests" to the live servers, you will find yourselves receiving a *** ton of hate mail from anyone who doesn't realize it is a "test" situation, and also from many who simply want you to know that they will not accept such changes. You will also find that you will lose customers over this, even if, ultimately, you decide not to keep these changes.

Instead, you need to look as much as possible at processing optimizations, hardware upgrades, and changes to skills that allow making things easier for computation and processing without fundamentally changing the very nature and feel of the combat system itself.

Here's one suggestion, that may have already been made: Make single-target abilities more powerful so that there is a *positive* incentive to use them, rather than a punitive disincentive against AoE.

Madhatten512

ZOS have you even considered practical changes like just off the top of my head LOWERING GROUP SIZE IN CYRO, making heals and buffs only stack inside the group, and incentives and/or penalize faction stacking that really causes a lot the issues. Might be a good place to start before you rip apart the combat.

0jtpk0

Writing to support the common dissent: these strategies go directly against what makes this game's combat fun.


To me, these strategies seem to indicate that ZOS is unwilling or unable to dedicate the necessary (likely heavy) resources to address performance issues in Cyrodiil without compromising the core mechanics of the game. If so, it seems we will either have the flow of combat severely diminished (like as proposed here) or will simply continue to endure poor performance for the remaining life of the game (assuming the combat changes actually do provide significant performance boosts).

Personally, if forced to choose, I would take keeping combat fluid over performance boost at the cost of fluidity any day of the week. At least now I can find times off peak hours where the game performs decently and get the best of both worlds from time to time.

HankTwo

Looks like they are finally killing their own game... There are myriads of playstyles these changes would destroy, and without heavily reworking a lot of the available aoe skills, a singificant amount of them would become useless. If I have the time I will make a more in depth reply, but I can already say that I'm very dissapointed by these proposed 'solutions' for the lag.

AbraXuSeXile

On greyhost eu i can think off top of my head 5 guilds who generally dont play at same time, maybe 2 on at same time.

Are you saying 2x10 is killing your servers?



Really

Tammany

Sugaroverdose wrote: »

Said a guy who wasn't in game even a year

Ever heard about google or forum browsing ?

It's pretty easy to find old topics and look how people disliked aoes in 2014.for example.
And since then general opinion constantly changes "we like", "we dont", "pls nerf aoe limit", "pls dont".
And literaly every patch topic was full of people with "Oooh you ruing the game, don't or i gonna leave".

muh

I'm pretty sure someone else here mentioned it already, but here we go.

Why reinvent the wheel? Yes, unrestricted or uncapped AoEs create a lot of stress on servers. It's nice of you to admit or recognize that it is unsustainable for your infrastructure. But I don't think your ideas are workable solutions for this problem.

Every other MMO with RvR/zerg style PvP areas, that come to my mind, have target caps on their AoE abilities for that exact reason.

If we look at Sap Essence from Nightblade, it deals damage in an AoE and heals friends in an AoE.
Imagine a 50vs50 zerg, a Nightblade uses Sap Essence and because both groups engaged in some serious group hug business they're all in range. That's what PvP is all about right? Friendship and love.

Anyway... As it is currently, the Server has to find all of those 50 enemies in range and apply damage to them and then find all of your 50 friends in range and heal them. So for simplicities sake 100 calculations for that one skill per second.

With a target cap of, let's say 8 enemies and 12 friends, Sap Essence only looks for enemies in range until it found 8 and then doesn't care about the other 42 it could hit. Then it looks for friends in range and stops after it found 12 and doesn't care about the other 38 anymore. Boom down to 20 calculations for that skill per second.

Reduction of 80% server caluclations for this unlikely scenario. No need to gut the player experience around AoE abilities 6 years into the lifecycle of the game.

bigdavid11b16_ESO

If anybody remembers. the day that nobody could group. is the day that cyrodiil had zero lag. Maybe you all should have a little ponder as to why. The reasoning is that there were no people spamming abilities that were linked to group utility
example : PURGE

Why can we not just start at once place. Group size. Lower it to a max of 12 and go from there.

Feric51

Ixtyr wrote: »

I can't speak for everyone, but I can't imagine anyone who actually wants to play the game like this - am I understanding correctly that this means every single skill with any form of AoE check is put on a 3s GCD? If so...Honor The Dead? Check. Sap Essence? Check. Virtually any buff or debuff skill that targets a player other than yourself or your immediate target you're fighting? Check. Deep Fissure? Check. Even skills like Funnel Health, which is a single-target skill but then has an AoE range check for its accompanying heal? How does that work? Where do AoE proc sets (damage, buffs, heals) factor into this? If my set procs, does that reset my CD and make me wait even longer for my next skill? If every single one of these skills is now on a cooldown boxing each other out, this is a terrible band-aid. . .and again, I don't understand what else you could be learning from this "test" if it's not just a test to see whether or not players accept it, rather than actually gaining new insights into areas for optimization.

Based on this logic, we'd also see skills like Mage's Fury/Wrath (...to the target and xxx damage to other enemies nearby); Force Pulse (Up to 2 nearby enemies....), Whirlwind (obviously), Shrouded Daggers, etc etc.

So many of these are spammed by one class or another, that the hyperbole about "blanket nerfs" of previous updates will finally be realized.

JTD

Sugaroverdose wrote: »

JTD wrote: »

Here's a couple of thoughts..
1. Cap AOE to a max amount of targets like in most mmorpgs focussed on pvp.

Whole PvP community asked for AoE caps removal during all those lags for years, and here's hero with solution of lags, rofl.

And did it improve? Is it better or worse?

Sugaroverdose

Tammany wrote: »

Sugaroverdose wrote: »

Said a guy who wasn't in game even a year

Ever heard about google or forum browsing ?

I know what exactly people talked here for years and it's effectively opposite of your claims.

muh wrote: »

I'm pretty sure someone else here mentioned it already, but here we go.

Why reinvent the wheel? Yes, unrestricted or uncapped AoEs create a lot of stress on servers. It's nice of you to admit or recognize that it is unsustainable for your infrastructure. But I don't think your ideas are workable solutions for this problem.

Every other MMO with RvR/zerg style PvP areas, that come to my mind, have target caps on their AoE abilities for that exact reason.

If we look at Sap Essence from Nightblade, it deals damage in an AoE and heals friends in an AoE.
Imagine a 50vs50 zerg, a Nightblade uses Sap Essence and because both groups engaged in some serious group hug business they're all in range. That's what PvP is all about right? Friendship and love.

Anyway... As it is currently, the Server has to find all of those 50 enemies in range and apply damage to them and then find all of your 50 friends in range and heal them. So for simplicities sake 100 calculations for that one skill per second.

With a target cap of, let's say 8 enemies and 12 friends, Sap Essence only looks for enemies in range until it found 8 and then doesn't care about the other 12 it could hit. Then it looks for friends in range and stops after it found 12 and doesn't care about the other 8 anymore. Boom down to 20 calculations for that skill per second.

Reduction of 80% server caluclations for this unlikely scenario. No need to gut the player experience around AoE abilities 6 years into the lifecycle of the game.

AOE caps have literally nothing to server performance issue, it was lagging with them as hell, so your proposal have nothing to real solution

Earthewen

We would then go through each class and ensure that there are viable builds for each and make adjustments as necessary.

Who would decide what is viable? And who would decide what adjustments are to be made? One of the things I love about ESO is that we can come up with our own builds. I happen to not run a meta build. I'm not sure I like someone who doesn't know me and how I play to tell me what build is viable ...

EtTuBrutus

These impact certain classes FAR more than others. Fine line of destroying the game and balance.

Earthewen

muh wrote: »

I'm pretty sure someone else here mentioned it already, but here we go.

Why reinvent the wheel? Yes, unrestricted or uncapped AoEs create a lot of stress on servers. It's nice of you to admit or recognize that it is unsustainable for your infrastructure. But I don't think your ideas are workable solutions for this problem.

Every other MMO with RvR/zerg style PvP areas, that come to my mind, have target caps on their AoE abilities for that exact reason.

If we look at Sap Essence from Nightblade, it deals damage in an AoE and heals friends in an AoE.
Imagine a 50vs50 zerg, a Nightblade uses Sap Essence and because both groups engaged in some serious group hug business they're all in range. That's what PvP is all about right? Friendship and love.

Anyway... As it is currently, the Server has to find all of those 50 enemies in range and apply damage to them and then find all of your 50 friends in range and heal them. So for simplicities sake 100 calculations for that one skill per second.

With a target cap of, let's say 8 enemies and 12 friends, Sap Essence only looks for enemies in range until it found 8 and then doesn't care about the other 42 it could hit. Then it looks for friends in range and stops after it found 12 and doesn't care about the other 38 anymore. Boom down to 20 calculations for that skill per second.

Reduction of 80% server caluclations for this unlikely scenario. No need to gut the player experience around AoE abilities 6 years into the lifecycle of the game.

Maybe we should remove the ability to macro in the game and some of the less functioning addons. That might help out with the too many calculations and information bottlenecks that seem to be a problem.

OtarTheMad

I think this will be interesting. I can't wait to get out there and test.

Vaoh

Reminds me of the initial introduction of AoE Caps.

You people have an MMO developed initially by a fantastic combat lead Nick Konkle, and yet all you can do is continue to homogenize, bandaid fix, rework/ruin core mechanics, and otherwise drive the gameplay into the ground every year. Absolutely disgusting.

Even I didn’t expect you all to think of directly adding cooldowns onto Skills. My interest in this game is sinking more than it ever has.

p00tx

It's not skills creating the lag. Players have been telling you for god knows how long now. I encounter horrendous lag when I'm by myself, riding through the middle of nowhere. It's your silly environmental graphics changes that trigger some of the worst lag. Snow, rain, fog, doors coming down, keep shapes changing after resources are gained or lost...all cause massive lag spikes and disconnects, no matter who is casting what in the vicinity. There is a reason you can barely ride through the Northern regions of Cyrodil. The whole thing is covered in snow and snowfall. Changing abilities won't make a bit of difference. A mass of players just riding through an area can cause massive slow downs. The issue is somewhere in your graphics.

Newflash: PvPers don't care about weather or keeps changing shape. They don't need roleplay elements in there. Just keep it simple and it would go a long way to fixing some of the more prevalent issues.

MadMattatatta

I would like to see, as posted in a previous response, a list of aoes and there durations posted before/during this test. after all this affects some class identities(jabs last 1 sec) vs players just spamming group aoe/heals (healing springs lasts 8 secs). it doesn't seem right to take away/ "adjust" a classes identity, individual aoe/class spam able, because some one doesn't know how to heal or use a group aoe properly.

Rianai

A lot of the performance deterioration can be connected to game updates - with the "performance improvement" update 25 as infamous (recent) highlight. And yet, not the slightest sign of reconsideration or attemps to revert whatever went wrong. Just the same old "we have already made a lot of improvements and will continue to do so" that's been part of almost every update for the past years with no noticeable positive impact on the game whatsoever.

Making mistakes is not a problem - as long mistakes are identified as such, lessons are learned and solutions are found.

But this whole story reminds me of throwing more and more flammable material into a smoldering fire until it turns into an inferno. And then, instead of admitting mistakes and trying to remove/reduce those added nutriments, fighting the fire by burying the entire area under tons of dirt. Yes it will eventually damp down the flames - and choke everything else that might have been still alive.

Cinbri

Sugaroverdose wrote: »

Well after observing this topic for some time, i believe that game just need to reverse all of battle mechanics changes to 1.5 patch state, with soft caps and everyone will be happy

Remember that 10% increase on stats that noone needed. It was far better with 2000 hp, not current 20000hp. Those totally unnecessary 0.001% calculations they added "for precise" dont even have real value (like if you actually notice difference between 20000hp and 20009hp) apart from drastically increase amount of calculations for every character with every stat, i.e. lag.

Wandering_Immigrant

My initial reactions to this;

Very interesting. I wish they would extend the tests to console so I could be a part of it.

This will most definitely improve performance, so expect one of these, or some variation of the concept to become permant.

Test 1) This is far too drastic to end up being the answer. It would require a full rebalancing of the game, all classes all abilities even single target would need to be adjusted, because if you can't cast any AOE your ST would have to make up the difference. This is just testing in the extreme to get a baseline. It'll be a rough week in Cyrodiil, but with a potentially major change on the horizon it's best to be as comprehensive as possible.

Test 2) This will be worse for some classes than others, but I think with an adjustment to the strength of affected abilities it could be a functional solution. Keep in mind none of this is ideal, but I could work around hard-hitting abilities with a cool down.

Test 3) This one would probably be the consensus player favorite. Sustaining is easy, especially if you're just sitting back laying down AoE/HoT, so it'll have minimal impact on combat. But because of that it won't do much for performance either, in fact it even adds an extra calculation in ramping cost. So don't get your hopes up here.

Test 4) Harder to swallow than test 2, but still much more functional than test 1. I could live with it if test 2 doesn't have optimal results, but I wouldn't like it.

I would like to see a 5th test run focusing on only ground AoE/Hot as I feel those are the biggest culprits here. Something along the lines of an individual half duration cool down with individual full duration ramping costs. So for example a ground AoE with a 12 second duration would have a hard cool down of 6 seconds followed by 6 seconds of ramping costs if recast before the effect ends. People cast these and the crowd moves so they cast them again, and then again. The server has to calculate for each cast and then continue recalculating each time someone runs in or out of the circle, that has to be a huge strain. Forcing these to be placed more strategically rather than spammed until you get it right I think would be the best combination of performance and gameplay if it worked.

Final thought:
I smell a Nightblade nerf coming in the future. Sad days. We finally start getting a little attention and now they're nerfing the most common counter to Cloak.