Upcoming siege changes in next major update

Ishammael

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

Takllin

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

Isn't this the perfect justification to get rid of AoE caps? Wheeler's explanation for not doing this is that it adds more calculations to the server, which would cause more lag. AoE caps does exactly this, just reduces damage instead of increasing it.

It's as if they want to exhaust every single possibility besides AoE caps, and even then I don't know if Wrobel wants to get rid of them...

BTW, well done on last week's ESO Live @ZOS_BrianWheeler that was in depth content that we don't often get to see or hear about.

frozywozy

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

Manoekin

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

Takllin

Manoekin wrote: »

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

This ^

ZOS_BrianWheeler

This is simply stating that adding scaling damage to Siege weapons means adding more LOS checks, which we are not looking to do as that's counter to the current efforts to reduce server stress.

frozywozy

Manoekin wrote: »

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

I'm not sure what you mean by removing the scaling damage from AOE caps without removing AOE caps themselves.

Manoekin

frozywozy wrote: »

Manoekin wrote: »

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

I'm not sure what you mean by removing the scaling damage from AOE caps without removing AOE caps themselves.

You could remove the scaling part and increase the AOE cap to a specific number, say 12 or any other number. If every time you cast an AOE ability it's causing double or triple the amount of LOS checks it should because of the damage scaling from AOE caps, that's significant. If it simply just "hit the target" and that was it, damage boom, you wouldn't have double or triple the checks or damage calculations. I can imagine what magicka det is doing, as it scales both up and down depending on the numbers hit.

Takllin

ZOS_BrianWheeler wrote: »

This is simply stating that adding scaling damage to Siege weapons means adding more LOS checks, which we are not looking to do as that's counter to the current efforts to reduce server stress.

True, and maybe we're incorrect in doing so, but we are looking at this and saying that AOE caps works in a very similar fashion, so why not use this justification for removing them?

Manoekin wrote: »

frozywozy wrote: »

Manoekin wrote: »

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

I'm not sure what you mean by removing the scaling damage from AOE caps without removing AOE caps themselves.

You could remove the scaling part and increase the AOE cap to a specific number, say 12 or any other number. If every time you cast an AOE ability it's causing double or triple the amount of LOS checks it should because of the damage scaling from AOE caps, that's significant. If it simply just "hit the target" and that was it, damage boom, you wouldn't have double or triple the checks or damage calculations. I can imagine what magicka det is doing, as it scales both up and down depending on the numbers hit.

Yeah it's essentially removing the scaling portion, and then the discussion becomes do you cap the number of people an AOE can hit, or do you leave it uncapped like siege.

frozywozy

Manoekin wrote: »

frozywozy wrote: »

Manoekin wrote: »

frozywozy wrote: »

Ishammael wrote: »

ZOS_BrianWheeler wrote: »

Siege at it stands now just hits and area and does damage, and is done.

Can you make all AoE abilities act like this?

It used to be like that and we would have people dealing aoe damage through walls.

We don't want this back.

I think the point is that scaling damage increases the number of LOS checks, and that they would like to see the scaling damage gone from AOE caps working as how siege does now. Thus, a significant drop in the number of LOS checks the server has to calculate, let alone the extra damage calculations.

I'm not sure what you mean by removing the scaling damage from AOE caps without removing AOE caps themselves.

You could remove the scaling part and increase the AOE cap to a specific number, say 12 or any other number. If every time you cast an AOE ability it's causing double or triple the amount of LOS checks it should because of the damage scaling from AOE caps, that's significant. If it simply just "hit the target" and that was it, damage boom, you wouldn't have double or triple the checks or damage calculations. I can imagine what magicka det is doing, as it scales both up and down depending on the numbers hit.

If we really have to go there... But from what I'm concerned, lord @fengrush thread about aoe caps was pretty clever. Remove them aaaaaaaaaand done.

ZOS_BrianWheeler

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

Takllin

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

Ah okay, I figured this might have been the case but somewhat hopeful it would be the same.

WRX

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

Disappointing to be honest.

Thanks for all the clarification.

jrkhan

@DeanTheCat
I think the los check must be to determine if targets in the red circle are in los with the center of the circle (e.g. If you are on the inside of a keep wall, you don't take damage from siege hitting the outside of the wall)

So, I could be way off on this @ZOS_BrianWheeler , but from what I'm reading it sounds like:
1) get list of targets in same general area as where seige hit
2) of those targets, determine which are within a certain radius squared of the center of seige
3) do a los check for each remaining target to see if they have los to center of siege circle (not behind a wall)
4) apply damage/debuffs

I don't see why you couldn't count the targets to be damaged in 3, and pass the total to 4 without additional los checks.
That total should be all you'd need for damage to scale similar to magika detonation..
Of course, it would depend on how exactly it's implemented and it might not be possible to explain without diving further in to the code then you'd care to do on this forum

Merlight

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

So it's the detonation-like scaling that is costly? That's yet another reason to get rid of AoE ability damage reduction (the 6/24/30 rule), because then you'll be able to remove the costly scaling from detonation and end up with not one, but plenty of zerg-buster abilities. I know it's not your domain, but I haven't seen @Wrobel acknowledge that AoE damage reduction alone is causing issues (and I'm not talking about its obvious unfairness towards small groups here). Magicka Detonation looks like it was designed specifically to overcome that reduction -- that's like fixing a leaky bucket by pouring more water in. I wish he would let go of that crutch already.

ZOS_BrianWheeler

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

In the most direct way to put it, after an LOS check is completed, damage according to the 6/24/30 rule is applied without further LOS checks because the server already knows who it's going to hit and all it needs to do at that point is sort out who to apply the rule to.

That sorting could be a pain point on the server, or could not be a pain point on the server as it's all part of what we're digging into regarding Cyrodiil performance. It is "one more thing the server has to do", but so are LOS checks, so are any AE, Siege included, so are heals (smart or not), and damage so there's no silver bullet here, but hopefully this makes things clear regarding the difference between the AE Cap (falloff) vs. Scaling damage.

Manoekin

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

Thanks for the clarification

Takllin

ZOS_BrianWheeler wrote: »

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

In the most direct way to put it, after an LOS check is completed, damage according to the 6/24/30 rule is applied without further LOS checks because the server already knows who it's going to hit and all it needs to do at that point is sort out who to apply the rule to.

That sorting could be a pain point on the server, or could not be a pain point on the server as it's all part of what we're digging into regarding Cyrodiil performance. It is "one more thing the server has to do", but so are LOS checks, so are any AE, Siege included, so are heals (smart or not), and damage so there's no silver bullet here, but hopefully this makes things clear regarding the difference between the AE Cap (falloff) vs. Scaling damage.

Thank you for the in depth explanation.

Not to keep bringing this off topic, but I wonder, if it would be possible to use the PTS to test AoE caps? Create two campaigns, one with, and one without. Or enable AoE caps for one week, and disable them the next? See how the performance goes.

Obviously this is @Wrobel department, so more of just a general question.

Starshadw

ZOS_BrianWheeler wrote: »

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

In the most direct way to put it, after an LOS check is completed, damage according to the 6/24/30 rule is applied without further LOS checks because the server already knows who it's going to hit and all it needs to do at that point is sort out who to apply the rule to.

That sorting could be a pain point on the server, or could not be a pain point on the server as it's all part of what we're digging into regarding Cyrodiil performance. It is "one more thing the server has to do", but so are LOS checks, so are any AE, Siege included, so are heals (smart or not), and damage so there's no silver bullet here, but hopefully this makes things clear regarding the difference between the AE Cap (falloff) vs. Scaling damage.

Well, as a data point, I've half a dozen Aussie friends who have, through a process of elimination, narrowed down the massive AOE spam in Cyrodiil as their pain point when it comes to horrible performance during PvP. It's awful, because most of them have simply stopped playing, since they can't enjoy PvP at all. Their ping rates jump into the thousands any time they find themselves up against a crowd of enemies all spamming AOE skills (which is pretty much what PvP has become at this point - and sadly, I don't see removing AOE caps as making that less likely at all). This is precisely why I begged you folks to stop closing down campaigns and compressing the PvP population onto fewer campaigns. Fewer campaigns = even MORE enemies all around, all spamming AOE.

And again, I truly wish that you'd remove the magicka det skill entirely. If at some point in the future you find a way to fix it so that 1) it works as you originally intended it (as a zerg-buster, as opposed to a zerg-enhancer) and 2) doesn't add to the server load issues, then by all means, bring it back.

Ishammael

ZOS_BrianWheeler wrote: »

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

In the most direct way to put it, after an LOS check is completed, damage according to the 6/24/30 rule is applied without further LOS checks because the server already knows who it's going to hit and all it needs to do at that point is sort out who to apply the rule to.

That sorting could be a pain point on the server, or could not be a pain point on the server as it's all part of what we're digging into regarding Cyrodiil performance. It is "one more thing the server has to do", but so are LOS checks, so are any AE, Siege included, so are heals (smart or not), and damage so there's no silver bullet here, but hopefully this makes things clear regarding the difference between the AE Cap (falloff) vs. Scaling damage.

Thanks for this info and the continued discussion. Very much appreciated!

As a (sort of) side note: if you ever need us (player groups) to go onto PTS in order to spam AoEs for a measurable effect of some proposed change to improve performance, we will gladly do it.

frozywozy

Starshadw wrote: »

ZOS_BrianWheeler wrote: »

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

In the most direct way to put it, after an LOS check is completed, damage according to the 6/24/30 rule is applied without further LOS checks because the server already knows who it's going to hit and all it needs to do at that point is sort out who to apply the rule to.

That sorting could be a pain point on the server, or could not be a pain point on the server as it's all part of what we're digging into regarding Cyrodiil performance. It is "one more thing the server has to do", but so are LOS checks, so are any AE, Siege included, so are heals (smart or not), and damage so there's no silver bullet here, but hopefully this makes things clear regarding the difference between the AE Cap (falloff) vs. Scaling damage.

Well, as a data point, I've half a dozen Aussie friends who have, through a process of elimination, narrowed down the massive AOE spam in Cyrodiil as their pain point when it comes to horrible performance during PvP. It's awful, because most of them have simply stopped playing, since they can't enjoy PvP at all. Their ping rates jump into the thousands any time they find themselves up against a crowd of enemies all spamming AOE skills (which is pretty much what PvP has become at this point - and sadly, I don't see removing AOE caps as making that less likely at all). This is precisely why I begged you folks to stop closing down campaigns and compressing the PvP population onto fewer campaigns. Fewer campaigns = even MORE enemies all around, all spamming AOE.

And again, I truly wish that you'd remove the magicka det skill entirely. If at some point in the future you find a way to fix it so that 1) it works as you originally intended it (as a zerg-buster, as opposed to a zerg-enhancer) and 2) doesn't add to the server load issues, then by all means, bring it back.

You see it the other way around. Compressing people into less amount of campaigns = Encouraging organized ball groups to NOT stack one unique competitive campaign making it a nightmare every primetime.

The best, as I said in the past, would be to have 2 or 3 campaign with reduced max pop by 20% and different max group sizes to fit everyone's playstyle. See how the queues go at primetime and adjust and add more campaign ONLY when needed (when all campaigns are max pop).

This way, we would not have 3-4 major ballgroups of 24+ per faction on Azura Star. Those guilds would spread across 3campaigns and it would reduce the amount of aoe fights happening at once.

Darlgon

Takllin wrote: »

Not to keep bringing this off topic, but I wonder, if it would be possible to use the PTS to test AoE caps? Create two campaigns, one with, and one without. Or enable AoE caps for one week, and disable them the next? See how the performance goes.

Obviously this is @Wrobel department, so more of just a general question.

Possible? Probably.. However.. effective? No.. probably not. If you look at the PvP community vs the game players as a whole, its pretty small. And, to make them willing to put aside the characters they hand-crafted and are trying to progress and go play on PTS with autoleveled templates that makes everyone in Cyro gods? Extremely unlikely.

Takllin

Darlgon wrote: »

Takllin wrote: »

Not to keep bringing this off topic, but I wonder, if it would be possible to use the PTS to test AoE caps? Create two campaigns, one with, and one without. Or enable AoE caps for one week, and disable them the next? See how the performance goes.

Obviously this is @Wrobel department, so more of just a general question.

Possible? Probably.. However.. effective? No.. probably not. If you look at the PvP community vs the game players as a whole, its pretty small. And, to make them willing to put aside the characters they hand-crafted and are trying to progress and go play on PTS with autoleveled templates that makes everyone in Cyro gods? Extremely unlikely.

You severely underestimate the willingness of the community to help ZOS get this fixed. It's not asking people to give up their characters for a week. We can set aside a few hours to test it.

Yes, it's a very different server than what live uses, but it's the best benchmark test we can get.

jrkhan

Still think I'm missing something.

Are you saying that at the time siege damage is calculated the server does NOT know who it's going to hit?
Unlike other abilities where it does know which targets are hit?

If it does know who it's going to hit, we're taking about multiplying 'base siege damage' + 'targetsHit' * 'perTargetSeigeBonus' right?
So, similar to an inverted 6/24/30 rule, except no sorting necessary.

Unless I've been misunderstanding exactly what was meant by a magika detonation like effect..
Did you mean everyone hit with seige would be debuffed with something like magika detonation that would then explode?
Because THAT sounds like it would increase LOS calls...

Is it possible there was some miscommunication somewhere?
I'm still not seeing why more LOS checks are necessary if we're just talking about scaling up damage based on targets hit - especially since scaling down via 6/24/30 does not.

And echoing other comments, greatly appreciate your going in to this level of detail.

Bfish22090

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

What good would increasing the damage of magicka detionation past 5 targets do if its scaled back down after 6 targets because of aoe caps

Manoekin

Bfish22090 wrote: »

ZOS_BrianWheeler wrote: »

AOE Caps and scaling damage as it's built in Magicka Det. function differently on the server level.

What good would increasing the damage of magicka detionation past 5 targets do if its scaled back down after 6 targets because of aoe caps

It will do more damage than it does now against groups tightly stacked with large numbers.

Takllin

That's also a question for @Wrobel as Wheeler doesn't handle mechanics.

Darlgon

Takllin wrote: »

Darlgon wrote: »

Takllin wrote: »

Not to keep bringing this off topic, but I wonder, if it would be possible to use the PTS to test AoE caps? Create two campaigns, one with, and one without. Or enable AoE caps for one week, and disable them the next? See how the performance goes.

Obviously this is @Wrobel department, so more of just a general question.

Possible? Probably.. However.. effective? No.. probably not. If you look at the PvP community vs the game players as a whole, its pretty small. And, to make them willing to put aside the characters they hand-crafted and are trying to progress and go play on PTS with autoleveled templates that makes everyone in Cyro gods? Extremely unlikely.

You severely underestimate the willingness of the community to help ZOS get this fixed. It's not asking people to give up their characters for a week. We can set aside a few hours to test it.

Yes, it's a very different server than what live uses, but it's the best benchmark test we can get.

No.. You over-estimate it.. by a LOT. After having tested the game builds on PTS for other major updates.. Cyrodiil is a wasteland, except for the guys hanging around BRK mill or wherever, dueling and yelling in zone chat to stop taking keeps because it ruins their balance of buffs. The only time we had huge participation was the "play with the devs stuff" which in turn lead to.. 10 minutes of playtime... and over an hour of server crashing. The official explanation? "Well.. the PTS is a smaller server, so it could not handle the load. We got some good data tho."

Merlight

ZOS_BrianWheeler wrote: »

@Merlight Detonation like scaling is more costly than Siege weapon damage server side, however removing that scaling isn't the same as removing the 6/24/30 rule on the server as they're hit the server with requests differently.

Magicka Det is built with several individual "lego blocks" that also do individual LOS checks for applying the initial cast, the multiplicative bonus for targets in the area, and lastly to apply the damage upon explosion.

The 6/24/30 rule is more ingrained into the ability system and technically any "lego block" can get that rule applied to it, and the LOS rules for each "lego block" in an ability determine how many LOS checks there are. For clarification, any ability can have LOS checks at the beginning of casting, end of casting, both beginning and end, or no LOS check at all. The important thing to note is the 6/24/30 rule does not add additional LOS checks.

Thanks for the explanation. I didn't mean they were equal. My line of thinking was this: if the 6/24/30 falloff didn't exist and all 60 took full damage, then Magicka Detonation wouldn't need that costly scaling to be powerful. It seems to me that you (or whoever designed Magicka Detonation) are struggling with the falloff, just like players.

Takllin

Darlgon wrote: »

Takllin wrote: »

Darlgon wrote: »

Takllin wrote: »

Not to keep bringing this off topic, but I wonder, if it would be possible to use the PTS to test AoE caps? Create two campaigns, one with, and one without. Or enable AoE caps for one week, and disable them the next? See how the performance goes.

Obviously this is @Wrobel department, so more of just a general question.

Possible? Probably.. However.. effective? No.. probably not. If you look at the PvP community vs the game players as a whole, its pretty small. And, to make them willing to put aside the characters they hand-crafted and are trying to progress and go play on PTS with autoleveled templates that makes everyone in Cyro gods? Extremely unlikely.

You severely underestimate the willingness of the community to help ZOS get this fixed. It's not asking people to give up their characters for a week. We can set aside a few hours to test it.

Yes, it's a very different server than what live uses, but it's the best benchmark test we can get.

No.. You over-estimate it.. by a LOT. After having tested the game builds on PTS for other major updates.. Cyrodiil is a wasteland, except for the guys hanging around BRK mill or wherever, dueling and yelling in zone chat to stop taking keeps because it ruins their balance of buffs. The only time we had huge participation was the "play with the devs stuff" which in turn lead to.. 10 minutes of playtime... and over an hour of server crashing. The official explanation? "Well.. the PTS is a smaller server, so it could not handle the load. We got some good data tho."

Can't remember a time when the community was so invested in a change that went into PTS for testing...

I can see that it's not worth discussing further with you, you view neither them nor the community highly.