Update on Cyrodiil Performance

stefan.gustavsonb16_ESO · March 2015

tiamak wrote: »

...
Or in general, maybe the much more regulated environment of a console game is in favour of delicate game mechanics like those in Cyrodiil?

If things work like I think they might (and I'm only guessing, but hopefully in an educated way), the console platforms might do just fine without all the server side safety checks to defend against hacked clients. Consoles do not allow running unsigned code, and that could be the salvation for ESO PvP. Without all the extra CPU load on the server side, without the increased network traffic and without the increasing sensitivity to latency that we have seen since update 1.2.3, their original "naive trust" networking model might just be able to handle the kind of large scale battles we had back then, and which they are still advertising.

If the console versions work better, I do not expect the PC/Mac version to receive any fixes. We are probably too few players left on that platform to motivate a complete overhaul of the software architecture for the server just for us.

Navaya · March 2015

I remember when they said that 200 players on screen at the same time wouldn't be a problem, now 50 players at the same time seem to be to much.

stefan.gustavsonb16_ESO · March 2015

Navaya wrote: »

I remember when they said that 200 players on screen at the same time wouldn't be a problem, now 50 players at the same time seem to be to much.

They still say that:

http://www.elderscrollsonline.com/en-gb/game-guide/the-alliance-war

The phrase "hundreds of players onscreen at once" is repeated three times in that short presentation, clearly making it a main selling point. Even if it might have been true before patch 1.2.3, and even if it might be possible in the upcoming console versions and maybe some time in the distant future for the PC/Mac version, I think it borders on false marketing to keep stating it that prominently in their presentation without even mentioning the huge and persistent problems.

Cinbri · March 2015

Move to haderus to play in small group long time ago and now just can't coz epic lags. It is lagging even when noone online in campaign.

belugian · March 2015

if imperial city is also a pvp zone like darkness falls was, should nt it lower the population in cyro ( since it ll be instanced ) and therefore reduce the lag ?

or, just an idea out of the blue, Imagine some new domination objectives ( hold several points in the same time dispatched on the map ) that would, let's say : fortify existing keeps or make siege engines stronger or open new pve areas or quests etc. would'nt it split the population according to your wishes if they re so hard to take they require a large group to get them ?

kaidanfanlava · March 2015

Is ZOS just ignoring all of this? Is there any way of knowing whether they are even aware of our problems...?

Also, are there any plans for the introduction of the Imperial City, and that stupid-ass dragon from moving from there?

stefan.gustavsonb16_ESO · March 2015

kaidanfanlava wrote: »

Is ZOS just ignoring all of this? Is there any way of knowing whether they are even aware of our problems...?

Your first clue should be that this thread was indeed started by Paul Sage from ZOS. Your second clue would be that this forum is pretty much filled with rage about the lag in Cyrodiil. Yep, pretty sure they are aware of it. Whether they are going to come up with a solution is anyone's guess, but I bet they are not ignoring it.

Eldarth · March 2015

I suspect animation canceling is a definite contributor to lag in Cyrodil.

For animation canceling, the client computer likely has to send 1)initial action/target, 2) Canceled/New action/target, 3) animation cancellation all affected clients, 4) damage update for (1), 5) damage update for (2).

That's at least 3-5 network interactions between ALL affected clients and servers in the space that should (without animation cancelling) take 1-2 network interactions. Pretty obvious you've at the very least doubled your network overhead for every player within the combat range.

Also allowing macroing to perform animation canceling further maximizes packet traffic from clients to occur faster than a person can press a key.

Animation canceling is an exploit regardless of ZOS "acceptance" of it.

Imagine you're in a boxing ring and you know your opponent slight dips his left shoulder whenever he throws a right punch. Now, imagine you're in that same ring and you don't see your opponent move at all -- you simply spontaneously get a bloody nose and a headache.

Or perhaps I move my hand a tiny bit -- and suddenly you're hit by a critical fire spell.

If there is no animation to telegraph what combat action or spell someone is casting then there is no way you can cast an interrupt or block what you can't see.

I'm pretty confident that animation canceling is a large contributor to lag in Cyrodil - especially now that it's been encouraged by ZOS.

Kevinmon · March 2015

Where the hell is the dev response to this thread?

wraith808 · March 2015

Eldarth wrote: »

I'm pretty confident that animation canceling is a large contributor to lag in Cyrodil - especially now that it's been encouraged by ZOS.

Other than the fact that animation canceling has been in the game through good times and bad with lag in Cyrodiil.

And your supposition as to what's happening with it is suspect at best.

Eldarth · March 2015

wraith808 wrote: »

Other than the fact that animation canceling has been in the game through good times and bad with lag in Cyrodiil.

And your supposition as to what's happening with it is suspect at best.

And how is it suspect?

Prior to ESO Live #11 it was considered a possible exploit. Since ESO Live #11 it was endorsed by ZOS and even macroing was approved. I'm pretty sure since then it has exploded in it's use especially in the zerg clusters.

stefan.gustavsonb16_ESO · March 2015

Kevinmon wrote: »

Where the hell is the dev response to this thread?

Easy now. This is a dev thread, and its very existence is a lot more than we have ever had about this problem before. Sure, not a word from them since the initial post, but I reckon they wouldn't be able to say more than "no news, sorry".

wraith808 · March 2015

Eldarth wrote: »

wraith808 wrote: »

Other than the fact that animation canceling has been in the game through good times and bad with lag in Cyrodiil.

And your supposition as to what's happening with it is suspect at best.

And how is it suspect?

Prior to ESO Live #11 it was considered a possible exploit. Since ESO Live #11 it was endorsed by ZOS and even macroing was approved. I'm pretty sure since then it has exploded in it's use especially in the zerg clusters.

It's suspect because of the fact that you're sending the same commands that you would be under any circumstances- it's just the timing that's different. It's not rocket science.

Eldarth · March 2015

wraith808 wrote: »

Eldarth wrote: »

wraith808 wrote: »

Other than the fact that animation canceling has been in the game through good times and bad with lag in Cyrodiil.

And your supposition as to what's happening with it is suspect at best.

And how is it suspect?

Prior to ESO Live #11 it was considered a possible exploit. Since ESO Live #11 it was endorsed by ZOS and even macroing was approved. I'm pretty sure since then it has exploded in it's use especially in the zerg clusters.

It's suspect because of the fact that you're sending the same commands that you would be under any circumstances- it's just the timing that's different. It's not rocket science.

Exactly. Timing. With animation canceling you are sending 2-3 times as many network transactions in the same period of time. It's not rocket science. It's grade school math.

wraith808 · March 2015

Eldarth wrote: »

wraith808 wrote: »

Eldarth wrote: »

wraith808 wrote: »

Other than the fact that animation canceling has been in the game through good times and bad with lag in Cyrodiil.

And your supposition as to what's happening with it is suspect at best.

And how is it suspect?

Prior to ESO Live #11 it was considered a possible exploit. Since ESO Live #11 it was endorsed by ZOS and even macroing was approved. I'm pretty sure since then it has exploded in it's use especially in the zerg clusters.

It's suspect because of the fact that you're sending the same commands that you would be under any circumstances- it's just the timing that's different. It's not rocket science.

Exactly. Timing. With animation canceling you are sending 2-3 times as many network transactions in the same period of time. It's not rocket science. It's grade school math.

They are the same actions that you would normally send. You don't actively cancel the animation. The animation is cancelled by the timing of the next action. Your talk about cancelling an animation being an action is wrong. The cancel is a function of the timing.

Eldarth · March 2015

wraith808 wrote: »

They are the same actions that you would normally send. You don't actively cancel the animation. The animation is cancelled by the timing of the next action. Your talk about cancelling an animation being an action is wrong. The cancel is a function of the timing.

Correct. It is the same "actions" that you would normally perform. The difference is you are performing 2-3 of them in the same time that you could only perform 1 without animation canceling.

You are "allowing" 2-3x the networking traffic with animation canceling.

Imagine if you replace "animation" with "cooldown." Same effect.

Or imagine you have to drink two 8 ounce glasses of water in 10 seconds.
Now, with animation canceling you have to drink five or six 8 ounce glasses of water in the same 10 seconds.

That is the networking bandwidth issue with animation canceling.

wraith808 · March 2015

Eldarth wrote: »

wraith808 wrote: »

They are the same actions that you would normally send. You don't actively cancel the animation. The animation is cancelled by the timing of the next action. Your talk about cancelling an animation being an action is wrong. The cancel is a function of the timing.

Correct. It is the same "actions" that you would normally perform. The difference is you are performing 2-3 of them in the same time that you could only perform 1 without animation canceling.

You are "allowing" 2-3x the networking traffic with animation canceling.

Imagine if you replace "animation" with "cooldown." Same effect.

Or imagine you have to drink two 8 ounce glasses of water in 10 seconds.
Now, with animation canceling you have to drink five or six 8 ounce glasses of water in the same 10 seconds.

That is the networking bandwidth issue with animation canceling.

Do you not like animation canceling? I notice that you say that it is an exploit, even though they've said it's not, and continue to do so? I just ask, because it would help me to understand why you phrase it as such, and talk against it in other threads also, and so I don't continue down this line with you if that's your perspective.

DHale · March 2015

Animation cancelling has become a part of the game. ULT generation is all tied to animation cancelling/weaving. I do not thing the server can hadle it but it is not going away. I light attack my fire staff all the time. It is not an exploit at all...now lol.

Eldarth · March 2015

wraith808 wrote: »

Do you not like animation canceling? I notice that you say that it is an exploit, even though they've said it's not, and continue to do so? I just ask, because it would help me to understand why you phrase it as such, and talk against it in other threads also, and so I don't continue down this line with you if that's your perspective.

No I don't. I think it's stupid. It virtually eliminates any tactical spell interuption or shield usage. How can you know to interrupt a spell or block an attack if you don't even see it telegraphed by your opponent.

Let's just go the rest of the way and have cooldown canceling as well.

And with macroing that alleviates having to have any skill -- just push a button and your computer fights for you at the fastest possible speed.

kelly.medleyb14_ESO · March 2015

How about a "Capture the Flag" scenario in the Quest Hubs. At the same time add more guards and make triple the quest XP rewards in those hubs. Give an incentive for holding the hubs of more points the longer you hold it.

Maybe .5 point increase worth per hour. I bet that would spread people out!!!

Also there should be larger ticks for TAKING a keep or node than defending it. As it is MUCH more difficult the reward should be that much more.

Thanks.

Kenny_McCormick · March 2015

biggest problem are the losers that have no skill and would prefer to make our lives miserable by spamming Meteor to lag out the server. the meteor spam is real. and it sucks. please look into this

maryriv · March 2015

If you want to spread people about the best solution is to greatly incentive quests. Currently the quest turn in is very pitiful and not worth doing them not to mention they are often on the other side of the map.

ThinkerOfThings · March 2015

So I have given this some thought, and sorry if I am repeating anything, however as others have pointed out, other meaningful objectives would help spread players out. So thinking about fun objectives that are rooted in the Elder Scrolls Universe I realized that Cyrodiil needs more Daedric Artifacts. It doesn't make sense to have everyone running around with these legandary weapons, so what if they were tied to an objective in Cyrodiil and available briefly but had a meaningful impact in PvP.

So to expand on this, what if we had shrines to 15 of the Daedric Princes ( I know there are 17 but Jyggalag is dormant at the moment and Molag Bal is doing his own thing ) which would leave 5 shrines in the vicinity of each of factions borders. Maybe once every x Minutes one of these shrines could become active ( with no more than like 3 active at a time ). While active any faction could come to the shrine and "worship" by occupying an area. After occupying the area for so long a randomly chosen player in the vicinity becomes the favored of that Daedra granting them a completely different skill bar and some nifty buffs to help keep them alive.

In order for the players to maintain their new champion status, a number of players would have to stay and guard the shrine, while others would have to stay with the champion. A buff would be given to both those players that stayed with the champion and those that stayed at the shrine, providing some form of diminishing returns the greater the number of players in the vicinity of the champion or the shrine so that it would encourage groups of certain sizes to split off. One to guard the shrine and those to follow the new Daedric Champion. Overall I see the champion having power almost in scale or on par with an emperor, however Daedric Power is fleeting and as such, easier to lose than Emporership.

When the shrine is no longer occupied by the faction that created a champion or the champion dies, the shrine goes dormant and x number of minutes later a new shrine activates somewhere on the map. If the newly active shrine was not announced to the campaign, this would require players to form scouting parties to figure out which ones may be occupied. Additionally there could be achievements gained for becoming favored of each of the different Daedric Princes and one for getting them all.

I am working on coming up with general ideas for each of the princes skills, some of which I thought would be awesome if they made the player a living siege weapon, provide buffs to the group, or faction, or were greater in strength when the player was alone in enemy territory, here is some of what I have so far:

Daedric Princes:
Azura - The female Deity who maintains/draws power from the balance of night and day, light and dark
Skills: Sun beans / Moon beam skills, passive buff that increases soul gem's filling for the faction.

Boethiah - A Deity of deceit, secrecy, conspiracy, treason, and unlawful overthrow of authority
Skills: Buffs that depend on slaying whoever holds the most territory.

Clavicus Vile - A shape-shifter, who grants power and wishes through ritual invocations and pact
Skills: Summon Barbas? Extra Mana Regen?

Hermaeus Mora - The formless Daedra of knowledge and memory, seeks to possess all that is knowable
Skills: Summon Tentacled Horrors, group debuffs, ranged stamina draining attacks? Tentacles!

Hircine - The Prince of the hunt, sport, the Great Game, and the Chase
Skills: Change into an uber were wolf and wreck shop, super health and stam regen.

Malacath - The Prince whose sphere is the patronage of the spurned and ostracized
Skills: Rage mode activate? Normal / heavy attacks do siege damage?

Mehrunes Dagon - The Prince of destruction, violent upheaval, energy, and mortal ambition
Skills: Become a living siege machine calling fire from the sky. The Razor just seems to OP for PvP.

Mephala - The Prince of unknown plots and obfuscation, a master manipulator, a sewer of discord
Skills: Gets buffs from staying away from group, perma stealth, debuffs enemy forces, summons spider daedra.

Meridia - A female Deity, the patron of the energies of all living things, enemy of the undead
Skills: Wield Dawnbreaker with extra effectiveness. Extra damage vs Undead Players.

Namira - A female Deity of the "ancient darkness", the patron of all things considered repulsive
Skills: Health regen based on enemies killed. Feeding skills. Poison damage

Nocturnal - A female Deity of the night and darkness, the patron of all things secretive
Skills: Stealth skills, group stealth buffs, summon crows.

Peryite - The Taskmaster, the Daedric Prince of Pestilence, desires order in his domain
Skills: Poison damage, summon daedra.

Sanguine - the Prince of hedonism, debauchery, and the further indulgences of one's darker nature
Skills: Summon daedra, life drains.

Sheogorath - The infamous Prince of Madness, whose motives are unknowable
Skills: Cheese Mortar? Fork of doom? Summon Haskil? Wabbajack would one shot people and also seems to OP for PvP.

Vaermina - A female Deity of dreams and nightmares, a deliverer of evil omens and dark portents
Skills: Stun abilities?

ESO has alot of great potential and I hope this thought helps, if nothing else gets ideas growing for other innovations!

Messy1 · March 2015

ZOS_PaulSage wrote: »

Just to clarify. There seemed to be a misunderstanding that I was saying we are "giving up" or "blaming this on players." Absolutely not. We are looking at the issue from many angles. I certainly didn't mean to imply we couldn't or wouldn't change code to improve performance, but rather that the answer wasn't more hardware. (It is often suggested this is the best way to fix problems.) I also wasn't implying you shouldn't play naturally, but unfortunately it is true that more players right now in a smaller area causes the performance problems. I say this not to discourage behavior or try push off the issue, but rather to give completely frank information about the root cause.

Are there issues with players cheating and intentionally causing lag that ZOS knows about and is taking steps to correct?

glak · March 2015

Messy1 wrote: »

ZOS_PaulSage wrote: »

Just to clarify. There seemed to be a misunderstanding that I was saying we are "giving up" or "blaming this on players." Absolutely not. We are looking at the issue from many angles. I certainly didn't mean to imply we couldn't or wouldn't change code to improve performance, but rather that the answer wasn't more hardware. (It is often suggested this is the best way to fix problems.) I also wasn't implying you shouldn't play naturally, but unfortunately it is true that more players right now in a smaller area causes the performance problems. I say this not to discourage behavior or try push off the issue, but rather to give completely frank information about the root cause.

Are there issues with players cheating and intentionally causing lag that ZOS knows about and is taking steps to correct?

They are not putting anyone on probation or banning anyone for any length of time if that is what you're asking.

Rune_Relic · March 2015

ThinkerOfThings wrote: »

So I have given this some thought, and sorry if I am repeating anything, however as others have pointed out, other meaningful objectives would help spread players out. So thinking about fun objectives that are rooted in the Elder Scrolls Universe I realized that Cyrodiil needs more Daedric Artifacts. It doesn't make sense to have everyone running around with these legandary weapons, so what if they were tied to an objective in Cyrodiil and available briefly but had a meaningful impact in PvP.

So to expand on this, what if we had shrines to 15 of the Daedric Princes ( I know there are 17 but Jyggalag is dormant at the moment and Molag Bal is doing his own thing ) which would leave 5 shrines in the vicinity of each of factions borders. Maybe once every x Minutes one of these shrines could become active ( with no more than like 3 active at a time ). While active any faction could come to the shrine and "worship" by occupying an area. After occupying the area for so long a randomly chosen player in the vicinity becomes the favored of that Daedra granting them a completely different skill bar and some nifty buffs to help keep them alive.

In order for the players to maintain their new champion status, a number of players would have to stay and guard the shrine, while others would have to stay with the champion. A buff would be given to both those players that stayed with the champion and those that stayed at the shrine, providing some form of diminishing returns the greater the number of players in the vicinity of the champion or the shrine so that it would encourage groups of certain sizes to split off. One to guard the shrine and those to follow the new Daedric Champion. Overall I see the champion having power almost in scale or on par with an emperor, however Daedric Power is fleeting and as such, easier to lose than Emporership.

When the shrine is no longer occupied by the faction that created a champion or the champion dies, the shrine goes dormant and x number of minutes later a new shrine activates somewhere on the map. If the newly active shrine was not announced to the campaign, this would require players to form scouting parties to figure out which ones may be occupied. Additionally there could be achievements gained for becoming favored of each of the different Daedric Princes and one for getting them all.

I am working on coming up with general ideas for each of the princes skills, some of which I thought would be awesome if they made the player a living siege weapon, provide buffs to the group, or faction, or were greater in strength when the player was alone in enemy territory, here is some of what I have so far:

Daedric Princes:
Azura - The female Deity who maintains/draws power from the balance of night and day, light and dark
Skills: Sun beans / Moon beam skills, passive buff that increases soul gem's filling for the faction.

Boethiah - A Deity of deceit, secrecy, conspiracy, treason, and unlawful overthrow of authority
Skills: Buffs that depend on slaying whoever holds the most territory.

Clavicus Vile - A shape-shifter, who grants power and wishes through ritual invocations and pact
Skills: Summon Barbas? Extra Mana Regen?

Hermaeus Mora - The formless Daedra of knowledge and memory, seeks to possess all that is knowable
Skills: Summon Tentacled Horrors, group debuffs, ranged stamina draining attacks? Tentacles!

Hircine - The Prince of the hunt, sport, the Great Game, and the Chase
Skills: Change into an uber were wolf and wreck shop, super health and stam regen.

Malacath - The Prince whose sphere is the patronage of the spurned and ostracized
Skills: Rage mode activate? Normal / heavy attacks do siege damage?

Mehrunes Dagon - The Prince of destruction, violent upheaval, energy, and mortal ambition
Skills: Become a living siege machine calling fire from the sky. The Razor just seems to OP for PvP.

Mephala - The Prince of unknown plots and obfuscation, a master manipulator, a sewer of discord
Skills: Gets buffs from staying away from group, perma stealth, debuffs enemy forces, summons spider daedra.

Meridia - A female Deity, the patron of the energies of all living things, enemy of the undead
Skills: Wield Dawnbreaker with extra effectiveness. Extra damage vs Undead Players.

Namira - A female Deity of the "ancient darkness", the patron of all things considered repulsive
Skills: Health regen based on enemies killed. Feeding skills. Poison damage

Nocturnal - A female Deity of the night and darkness, the patron of all things secretive
Skills: Stealth skills, group stealth buffs, summon crows.

Peryite - The Taskmaster, the Daedric Prince of Pestilence, desires order in his domain
Skills: Poison damage, summon daedra.

Sanguine - the Prince of hedonism, debauchery, and the further indulgences of one's darker nature
Skills: Summon daedra, life drains.

Sheogorath - The infamous Prince of Madness, whose motives are unknowable
Skills: Cheese Mortar? Fork of doom? Summon Haskil? Wabbajack would one shot people and also seems to OP for PvP.

Vaermina - A female Deity of dreams and nightmares, a deliverer of evil omens and dark portents
Skills: Stun abilities?

ESO has alot of great potential and I hope this thought helps, if nothing else gets ideas growing for other innovations!

Nice idea....prefer each of their daedric relics To be honest.
Mace of molag bal
Azuras star.
Nocturnal cloak
Dagons razor
....etc etc.

Put these at the end of an impossible quest against inpossible odds.
Anyone who actually does manage to get to them....well.... they earnt them well and truly.
After one day they are withdrawn from nirn and returned to their sanctum.
You become godlike....for but one day. In which you can change the face of Cyrodiil.

The fun is not so much in the winning..although that would be awesome.
The fun comes from the sacrifice in the attempt.

stefan.gustavsonb16_ESO · March 2015

This is a very nice idea for fun and lore-related PvP play, but to be honest, it would feel quite strange to implement a large chunk of new content just to try to put a band-aid on the problem that the game has crippling bugs and doesn't work as intended.

I'm not even sure the band-aid would work at all. After the Daedric artifact was acquired, everyone would probably clump on the current Daedric Champion, friend and foe. The inability of the game to handle large groups in one location would still be a big problem, perhaps even bigger than now.

As long as the game can't handle the "massive, epic battles" for which it was designed, and which are still advertised as the main selling point for PvP, Cyrodiil will be a mess.

C3N50R3D · March 2015

stefan.gustavsonb16_ESO wrote: »

If things work like I think they might (and I'm only guessing, but hopefully in an educated way), the console platforms might do just fine without all the server side safety checks to defend against hacked clients. Consoles do not allow running unsigned code, and that could be the salvation for ESO PvP. Without all the extra CPU load on the server side, without the increased network traffic and without the increasing sensitivity to latency that we have seen since update 1.2.3, their original "naive trust" networking model might just be able to handle the kind of large scale battles we had back then, and which they are still advertising.

Has anyone confirmed it is the server safety checks that are causing the slowdowns? Have any devs commented on this? If this unlikely scenario is true then I would rather put up with a few cheaters that can be pointed out and banned by the community than a nearly unplayable PvP experience.

Sacadon · March 2015

C3N50R3D wrote: »

stefan.gustavsonb16_ESO wrote: »

If things work like I think they might (and I'm only guessing, but hopefully in an educated way), the console platforms might do just fine without all the server side safety checks to defend against hacked clients. Consoles do not allow running unsigned code, and that could be the salvation for ESO PvP. Without all the extra CPU load on the server side, without the increased network traffic and without the increasing sensitivity to latency that we have seen since update 1.2.3, their original "naive trust" networking model might just be able to handle the kind of large scale battles we had back then, and which they are still advertising.

Has anyone confirmed it is the server safety checks that are causing the slowdowns? Have any devs commented on this? If this unlikely scenario is true then I would rather put up with a few cheaters that can be pointed out and banned by the community than a nearly unplayable PvP experience.

No, I've seen nothing to this and doubt they are as a significant contributor as suspected by many.

Rune_Relic · March 2015

Sacadon wrote: »

C3N50R3D wrote: »

stefan.gustavsonb16_ESO wrote: »

If things work like I think they might (and I'm only guessing, but hopefully in an educated way), the console platforms might do just fine without all the server side safety checks to defend against hacked clients. Consoles do not allow running unsigned code, and that could be the salvation for ESO PvP. Without all the extra CPU load on the server side, without the increased network traffic and without the increasing sensitivity to latency that we have seen since update 1.2.3, their original "naive trust" networking model might just be able to handle the kind of large scale battles we had back then, and which they are still advertising.

Has anyone confirmed it is the server safety checks that are causing the slowdowns? Have any devs commented on this? If this unlikely scenario is true then I would rather put up with a few cheaters that can be pointed out and banned by the community than a nearly unplayable PvP experience.

No, I've seen nothing to this and doubt they are as a significant contributor as suspected by many.

The security checks that affetc peformance are not what you think they are.
They are normally subroutines that are run on client instead of run on server.
The more the client can do, the less network packets and tasks the server has to do.

If you have the client recording its own location, action, damage, heals etc...then the system is rapid. But then the client can dictate the gameplay and exploit the damage ect it does to other clients.
By "running on server" ZOS keeps control of all damage calcs and just tells the client what it gets.
So there can be no exploits and hacks to make people godlike.

Apparently before the 1.3 patch some have said ZOS was running client code.
When they fixed this security hole it killed the performance.
At came at the same time as the lighting patch so many assumed it was this instead.

No one can be 100% certain of this...but it makes sense how massive combat suddenly became a problem for ZOS. ZOS already confirmed there is no issues with the lighting patch and they have been through it with a fine tooth comb.