How Many, How Far Apart, For Better Server Performance?

Recremen

We all know the holistic mantra "spread out", but in the interest of player-driven behavior changes creating better server stability, it would be helpful to know exactly how spread we need to be! We were shown in a recent ESO Live that Cyro can be broken up into a grid representing chunks of terrain dedicated to some discrete part of the megaserver. Assuming that wasn't an oversimplification to help us non-engineers understand the concept, would it be helpful for players to be able to turn this grid on in the UI? Then we could start talking about how many players per grid piece can be "safely" handled, and even do some testing on our own to see how many players we can have in an area fighting organically without issue. As I recall from another recent communication, a big problem with testing PvP changes is a lack of ability to duplicate live combat scenarios. If we can give more nuanced feedback regarding server load by gaining this information about the grid, would that be of benefit to the dev team? Would it be of benefit to us players so we can better negotiate how many to bring to a given fight, what an acceptable loss is versus an escalation of force that risks more lag, etc.?

If the grid isn't a perfect analogy and I'm running to far ahead on a meaningless tangent, is there some other analogy that could help us better discern for ourselves if we're spread out enough or not? Currently our metrics tend to be "This fight is good, there's no lag, let's keep going" and "This fight is terrible, it's so laggy, but we're all here anyway so let's just keep going". If instead we could talk about the relationship between population density, location, etc., we could probably make more informed decisions!

Can any of the engineers on the dev team provide some qualifying statements/advice/metrics we can use to try to figure this out? We all of course appreciate the dedication to delivering on the "hundreds of people on screen at once" advertisement, but in the meantime it would be a pretty cool stopgap to navigate the issue with all available data. Thanks!

WRX

The grid discussion was kinda strange from what I remember, and people have different interpretations.

I would still love to see some transparency so the players who care to make this game better can make a more informed decision and difference. I'm glad to see more guilds hopping aboard this view point.

All venting aside, it would be awesome to see a response to this to give the players a little more knowledge.

KenaPKK

I'm more conceptual than this. I'm pretty satisfied with "if there is lag, move apart until there is no more lag."

Also, is any amount of information enough to motivate the careless majority to change their behavior? What we need are positive reinforcement tools to encourage spreading out -- i.e. meaningful PvP objectives spread all throughout Cyrodiil and IC, and greater rewards for fighting in small groups or in new locations.

Ishammael

Add map objectives in all of the dead space of Cyrodiil. Objectives that both count to the AvA score as well as reward players with something tangible, for example: harvest nodes of all flavors (columbine!), bonus to TV stones in the sewers, the AP bonus buff w/o needing to enter a delve, etc.

Sallington

Am I the only one that thinks groups spreading out is kind of a bullsh!t solution, and us not getting the game we bought?

No matter what, some objectives will be more important than others, and a majority of forces from either guilds or pugs will all converge on them. Their infrastructure and either handle the game they made, or it can't. Right now it can't.

Sieging keeps with a group of 12 vs another group of 12 is nowhere near as fun as a 60 vs 60 slaughterfest.

Minno

Sallington wrote: »

Am I the only one that thinks groups spreading out is kind of a bullsh!t solution, and us not getting the game we bought?

No matter what, some objectives will be more important than others, and a majority of forces from either guilds or pugs will all converge on them. Their infrastructure and either handle the game they made, or it can't. Right now it can't.

Sieging keeps with a group of 12 vs another group of 12 is nowhere near as fun as a 60 vs 60 slaughterfest.

We can be angry at them, but we are facing the realty that current networking standards for this game might not be able to handle the loads at primetime. Spreading out seems to let us do our part so long as ZOS does theirs.

Minno

KenaPKK wrote: »

I'm more conceptual than this. I'm pretty satisfied with "if there is lag, move apart until there is no more lag."

Also, is any amount of information enough to motivate the careless majority to change their behavior? What we need are positive reinforcement tools to encourage spreading out -- i.e. meaningful PvP objectives spread all throughout Cyrodiil and IC, and greater rewards for fighting in small groups or in new locations.

I agree but at the same time constant information on the nature of this issue wouldn't hurt either.

Minno

Regarding OP's post, my personal feelings on the matter force me to do a couple of things:

- home on an underdog server during my time of play (right now its Haderus sewers might switch to trueflame.).

- move to the opposite side of the map if I see a DC zerg at the first keep I enter. If zone chat says "I got chal defense" I'll go to help by hitting resources at the next enemy keep or stay on reinforment lines.

I'm only one person, but I do what I can. I'm sure other groups have a similar strategy.

Sallington

What I would love is if they would have a "test" PvP campaign where they could get data on how certain skill are effecting performance. Maybe disable caltrops for a weekend. Maybe take out AOE caps for a weekend. Just try some different things for once instead of internally "working on it" without any feedback.

Recremen

Sallington wrote: »

What I would love is if they would have a "test" PvP campaign where they could get data on how certain skill are effecting performance. Maybe disable caltrops for a weekend. Maybe take out AOE caps for a weekend. Just try some different things for once instead of internally "working on it" without any feedback.

That would be cool, but if I recall they said that isn't doable. :-( Something about all campaigns on Live drawing from the same ability data sets but, obviously, different campaign scoring rules, etc. So disabling Caltrops on one server would mean disabling them on all.

Sallington

I'm all for some drastic changes to test performance on the live servers. What's the worst that happens? They say "Woops!" and revert the changes?

Dreyloch

I would love it if we could truly help the Devs and engineers make the lag go away. However, if your running around with your guild/raid/small group. Are you really going to shy away from a keep battle just for the sake of reducing lag? Having knowledge of how many players are in that grid area? Especially if it's your keep, or if you know your faction has less people?

Sadly I think most players would just find a way to use it as a tactical tool. Any raid I've ever been in is looking for 2 things. Either gaining keeps for position, or just plain old good fights.

Minno

Dreyloch wrote: »

I would love it if we could truly help the Devs and engineers make the lag go away. However, if your running around with your guild/raid/small group. Are you really going to shy away from a keep battle just for the sake of reducing lag? Having knowledge of how many players are in that grid area? Especially if it's your keep, or if you know your faction has less people?

Sadly I think most players would just find a way to use it as a tactical tool. Any raid I've ever been in is looking for 2 things. Either gaining keeps for position, or just plain old good fights.

While true, if (iin my case I'll use DC names) I see VE says they have ales defended, going there with more people probably makes zero sense from both a strategic and performance standpoint. You're better off hitting anotger keep or help defend other areas.

Yes people will venture to cross swords. They will hit a keep for fights. But if you see a group claim they can defend or you notice more bodies than the task for attack/defend, take the initiative and move to another place. That action might start another area for battles to help reduce people going to the same location.

Minno

Recremen wrote: »

Sallington wrote: »

What I would love is if they would have a "test" PvP campaign where they could get data on how certain skill are effecting performance. Maybe disable caltrops for a weekend. Maybe take out AOE caps for a weekend. Just try some different things for once instead of internally "working on it" without any feedback.

That would be cool, but if I recall they said that isn't doable. :-( Something about all campaigns on Live drawing from the same ability data sets but, obviously, different campaign scoring rules, etc. So disabling Caltrops on one server would mean disabling them on all.

Correct. They can't change one without the others. But I believe they stated that when PTS hits, for us to test pvp there and start discussing issues.

FMonk

Sallington wrote: »

I'm all for some drastic changes to test performance on the live servers. What's the worst that happens? They say "Woops!" and revert the changes?

The worst that could happen is that the test changes horribly break the game, and then they don't revert them (because they've never reverted any changes), and we're stuck with an even worse game until the "next major update" (which may not even fix anything.

Takllin

FMonk wrote: »

Sallington wrote: »

I'm all for some drastic changes to test performance on the live servers. What's the worst that happens? They say "Woops!" and revert the changes?

The worst that could happen is that the test changes horribly break the game, and then they don't revert them (because they've never reverted any changes), and we're stuck with an even worse game until the "next major update" (which may not even fix anything.

This ^

I think it's actually impossible for them to revert code, because of how poorly written/held together it is.

Recremen

Minno wrote: »

Recremen wrote: »

Sallington wrote: »

What I would love is if they would have a "test" PvP campaign where they could get data on how certain skill are effecting performance. Maybe disable caltrops for a weekend. Maybe take out AOE caps for a weekend. Just try some different things for once instead of internally "working on it" without any feedback.

That would be cool, but if I recall they said that isn't doable. :-( Something about all campaigns on Live drawing from the same ability data sets but, obviously, different campaign scoring rules, etc. So disabling Caltrops on one server would mean disabling them on all.

Correct. They can't change one without the others. But I believe they stated that when PTS hits, for us to test pvp there and start discussing issues.

I do hope we get better testing on this round of PTS, but the continued lack of incentives makes me think this will continue to be unlikely. Something as simple (or at least, something that sounds to me as simple) as letting AP earned on PTS count towards live servers could probably get a lot more people testing. But when we're faced with "spend hours of fighting for objectives that matter, or spend hours fighting for objectives that won't count once the servers reset", the choice is pretty clear for most people. :-/

Recremen

Also, which ZOS dev do I tag in that would have the highest chance of being able to provide insight on this topic? @ZOS_BrianWheeler ? Someone else?

Sallington

Recremen wrote: »

Also, which ZOS dev do I tag in that would have the highest chance of being able to provide insight on this topic? @ZOS_BrianWheeler ? Someone else?

KenaPKK

Recremen wrote: »

Also, which ZOS dev do I tag in that would have the highest chance of being able to provide insight on this topic? @ZOS_BrianWheeler ? Someone else?

You can reach out to them, but no ETA on a reply.