Remove AoE caps so the game becomes playable

Once again had to endure over 15 consecutive minutes of 5 sec skilldelay because of a zergfight on the other side of the map. Remove AoE caps already so this doesn't happen (and no the warclaw has absolutely nothing to do with the delay before anyone tries to say so).

why removing aoe cap wont make laggier? what's ur technical logic behind it?

@SkyShroud.2865 said:why removing aoe cap wont make laggier? what's ur technical logic behind it?

alright, I could be completely wrong in this, so little disclaimer before you hear the rest.I asked this question on reddit a while back. the response I got is that servers can calculate things like damage easily, but not who gets that damage. when calculating the who part, the server has to do an area check on everyone near the skill. so multiply that by 1000 times per second and it can add up to a lot.

What's the logic here, that with no AoE caps the zergs will just run ele dps or something and instakill the entire enemy zergs, thus creating less lag since fights are over quickly?

Because that's the only logic I see with this. Otherwise, as mentioned above no AoE cap would create a hell of a lot of more lag than how it is now.

Or nerf aoe, spam hard so every battle isn't a literal aoe spam fest,that will probably solve the crazy amount of constant instant multiple damage calculations as well as draw people into the mode. All 6 of my buds I talked into trying it within three hrs were like f this garbage lmao and left cuz the scourge aoe spamming. Dont blame them either,I'd have left by now if there was a better mmo out there.

@Stand The Wall.6987 said:

@"SkyShroud.2865" said:why removing aoe cap wont make laggier? what's ur technical logic behind it?

ermmm...it indeed makes logical sensegiven that the server has no need to "decide" the targets and simply just target everything in the radius, it indeed lead to less cpu time

but how you know the lag is due to it and not other factor like the network?

@"SkyShroud.2865" said:given that the server has no need to "decide" the targets and simply just target everything in the radius, it indeed lead to less cpu timeUhm... of course it has to "decide" the targets. How else would it know players got hit in the first place?

From a layman point of view the TL;DR of AoE is this;

If a skill has 5 man cap and there are 270 potential targets (~90+90+90 since AoE can have both damage and boons), you start checking distances and can end early if 5 targets on each pulse. If a skill has no cap, you have to keep checking all 270 for distance regardless if 5, 10, 20 or more already found.

Take one wild guess which of these would exponentially start to hog down the server the more AoE gets put down. Even if done the other way around (from each players perspective, kitten if I know how programming works) then you can still optimize it by cutting any distance algorithm off at AoE cap reached.

Its really the same theory on how culling work. You may argue that well, if an engine knows it has to dumb render the entire scene full depth thats way simpler math than determining which plane is hidden by what other plane and should be faster, right? Well... yes. But actually no. Only from a certain point of view.

I'd rather they removed half the boons and conditions from the game and skills, capped their durations, and put less stress on servers that way.

Can also see people "one shotting" themselves off retal. xD

@"Dawdler.8521" said:sniphey dawdle a question,if an aoe was changed from 5 targets to 20, what would that do to lag since the server has to send out less "no" packets? am I thinking about this backwards?

@SkyShroud.2865 said:but how you know the lag is due to it and not other factor like the network?I strongly doubt it's network related, packet streams do not pile up exponentially as players clash up. And people would notice if having a better net connection lead to less lag.

Personally, I would be fine with aoe caps being removed, even if just so there's a good reason to not stack up and hug your commander in zergs.

@Stand The Wall.6987 said:

@"Dawdler.8521" said:sniphey dawdle a question,if an aoe was changed from 5 targets to 20, what would that do to lag since the server has to send out less "no" packets? am I thinking about this backwards?Not sure what you mean with a "no" packet. Again, I'm no programmer but checking who's

not

in distance of an AoE field combined with an AoE cap seems backwards and not something that could be optimized for speed. But I dont know, dont even know if the GW2 engine use distance in such simple way and just steps through a player array to find targets. Its not that easy since you dont know the order of which it finds players (ie it could randomly end up the same, or 4x slower). Many players might not even be valid for calculation true (you can actually see how the engine map player dots on the minimap for example, it cuts down on the refresh when they are too far away. Pointless info, the server dont need to send you that so often). But that's WAY larger than AoE so one cant apply "they are too far anyway" to AoE. Anyone within your view will be moving fast and smooth and could step on your AoE at any time (well, ideally).

Somehow I hope that an Anet dev would come in here and just say "no you're wrong, this is how it works..." :p

@Dawdler.8521 said:

@"SkyShroud.2865" said:given that the server has no need to "decide" the targets and simply just target everything in the radius, it indeed lead to less cpu timeUhm... of course it has to "decide" the targets. How else would it know players got hit in the first place?

Well, I was thinking from the pure dps perspective. Upon looking a this thread again, I remembered there is also dot dps. This means that the server has to keep track of the dot over a greater number of people. This start to question if it really reduce cpu time. Of course, the boons too, whose boons take priority, whose boons should overwrite whose boons.

@Skotlex.7580 said:

@"SkyShroud.2865" said:but how you know the lag is due to it and not other factor like the network?I strongly doubt it's network related, packet streams do not pile up exponentially as players clash up. And people would notice if having a better net connection lead to less lag.

That is the real mystery. The lag people usually do complain about is "skill" lag but where does this "skill" lag come from? We can still move during "skill" lag thus is it really server lagging or is it skill packet got choked somewhere?

@SkyShroud.2865 said:

@Skotlex.7580 said:

@SkyShroud.2865 said:but how you know the lag is due to it and not other factor like the network?I strongly doubt it's network related, packet streams do not pile up exponentially as players clash up. And people would notice if having a better net connection lead to less lag.

@Dawdler.8521 said:

@Stand The Wall.6987 said:

@Dawdler.8521 said:sniphey dawdle a question,if an aoe was changed from 5 targets to 20, what would that do to lag since the server has to send out less "no" packets? am I thinking about this backwards?Not sure what you mean with a "no" packet. Again, I'm no programmer but checking who's

not

in distance of an AoE field combined with an AoE cap seems backwards and not something that could be optimized for speed. But I dont know, dont even know if the GW2 engine use distance in such simple way and just steps through a player array to find targets. Its not that easy since you dont know the order of which it finds players (ie it could randomly end up the same, or 4x slower). Many players might not even be valid for calculation true (you can actually see how the engine map player dots on the minimap for example, it cuts down on the refresh when they are too far away. Pointless info, the server dont need to send you that so often). But that's WAY larger than AoE so one cant apply "they are too far anyway" to AoE. Anyone within your view will be moving fast and smooth and could step on your AoE at any time (well, ideally).

idk what I mean by that either lol. yeah getting told whats what by a dev would be nice, but some dreams would be crushed for sure.

@Dawdler.8521 said:

@Stand The Wall.6987 said:

@Dawdler.8521 said:sniphey dawdle a question,if an aoe was changed from 5 targets to 20, what would that do to lag since the server has to send out less "no" packets? am I thinking about this backwards?Not sure what you mean with a "no" packet. Again, I'm no programmer but checking who's

not

in distance of an AoE field combined with an AoE cap seems backwards and not something that could be optimized for speed. But I dont know, dont even know if the GW2 engine use distance in such simple way and just steps through a player array to find targets. Its not that easy since you dont know the order of which it finds players (ie it could randomly end up the same, or 4x slower). Many players might not even be valid for calculation true (you can actually see how the engine map player dots on the minimap for example, it cuts down on the refresh when they are too far away. Pointless info, the server dont need to send you that so often). But that's WAY larger than AoE so one cant apply "they are too far anyway" to AoE. Anyone within your view will be moving fast and smooth and could step on your AoE at any time (well, ideally).

@Ben Phongluangtham.1065 - Could you tell us if we are right or wrong? I do think if the servers would not literally die that removing the AoE caps might actaully get zergs to ball up less, but if it's just not feasible tech wise we could stop going down this rabbit trail.

@Dawdler.8521 said:

@SkyShroud.2865 said:

@Skotlex.7580 said:

@SkyShroud.2865 said:but how you know the lag is due to it and not other factor like the network?I strongly doubt it's network related, packet streams do not pile up exponentially as players clash up. And people would notice if having a better net connection lead to less lag.

I have never read about skill queue, where the source?!

@"Dawdler.8521" said:

Somehow I hope that an Anet dev would come in here and just say "no you're wrong, this is how it works..." :p

Well, we have had devs state specifically that the AoE cap is a technical limitation of the game's engine and not a gameplay decision.. So for whatever reason the AoErng is more efficient than real AoE.

What I don't get is why add so much AoE when you know damn well it's gonna bork the servers... Can scourge not AoE? Why make a class that highlights your games weak point?

@SkyShroud.2865 said:

@Dawdler.8521 said:

@SkyShroud.2865 said:given that the server has no need to "decide" the targets and simply just target everything in the radius, it indeed lead to less cpu timeUhm... of course it has to "decide" the targets. How else would it know players got hit in the first place?

Exactly so the server has to calculate all of the AoEs and who they affect. Boon applications and stacking of them or if they're corrupted to conditions. Condition application and stacking of them via AoEs and if they're cleansed or converted back to boons. Healing AoEs based on healing power and who they affect and if they are poisoned. Crowd Control and who that affects and if they have stability or stunbreaks. Stealth and if they're marked or revealed. Teleports and where they end up during all this. Toughness on each character and how much power damage it reduces. If the players are affected by invulnerabilities or evades. If the characters are blocking and also if the person attacking them has unblockable. Reflection of projectile attacks and who it's reflected back to. Quickness and Alacrity and its effect on speed/recharge rates. Disabling conditions and their effect on movement if not cleansed. I could go on but the game will definitely lag during all these calculations exponentially.

@LetoII.3782 said:

@"Dawdler.8521" said:

Somehow I hope that an Anet dev would come in here and just say "no you're wrong, this is how it works..." :p

It is Anet quality.....skill lag happens due servers botlenecking one side to deliver in info in queue back...Scourges and firebrands are a bad design due how game works, still they are needed classes for most gw2 players due how bad they are, enxe why gw2 is Gimmick over skill, it’s like Diablo3, with my necro lvl 1, I started directly in torment 7 not due skill but how broken class is carrying me trough the game.Similar process happens in gw2, overperforming classes, low skill for the aoe combat players don’t need to know how to burn their aoe, due low cd and how low risk/ high reward when stacking the classes Anet wants players to stack.

Just think how next elite specs aoe spam will need to overpweform the current ones....

Removing aoe cap also would kill the servers, 1st anet need to remove aoe and then increase on skill that have to be aoe strong against stacked zergs o. Higher cd to increase the difficulty gameplay with thisbplayers would have to learn how to play rather then search for a stacking spam Zerg build.

@Dawdler.8521sometimes in a zerg fight, a skill wont turn on even if i press the button lots over a minuteother skills will work but one wontis that lag or a bug?

You can't simply remove the 'aoe cap' to achieve what you want. You have confused the AOE cap and the area search. It is understandable why someone non-technical might make the mistake, it's because your brain does an area search so effortlessly in RL. Let's explain how the game might have to attempt an area search:

The game has data for the position of every player. But, the game doesn't understand that player A and player B are right to some position where a skill went off (without special techniques). For us as humans, we can see straight away that player A and B are obviously in position, but the game doesn't think like a human.

In the most naive method, the game would have to go through the list of players on a map, and calculate how far away they are from a point on the map. The ones who are within AOE range, are the ones that receive the effect of a skill at that location. That's a slow operation (if done like that). Actually, if the game decides to stop searching after reaching the AOE limit, that's an optimisation (ie, AOE caps help performance).

There are ways to make the search faster, for example if you divide the map up into a grid and keep a list of players in each grid cell, then you know you only need to search through those grid cells. There are more fancy ways to do similar things too. But if all the players are in the same grid cell, then that's the same problem all over again !

@SkyShroud.2865 said:

@Dawdler.8521 said:

@SkyShroud.2865 said:

@Skotlex.7580 said:

@SkyShroud.2865 said:but how you know the lag is due to it and not other factor like the network?I strongly doubt it's network related, packet streams do not pile up exponentially as players clash up. And people would notice if having a better net connection lead to less lag.

This is important for two reasons:1) prevent client based cheating that would modify skills and2) prevent huge damage spikes for pressing a skill 10 times in a row until the server is able to respond with "skill now on cooldown".

If the server is overloaded with processing requests, then every mechanic that needs calculations and synchronisations from the server will behave in weird ways.

Fun Fact: During the first HoT beta the modified daredevil dodge rolls were coded as skills (and not actial dodge rolls) and suffered the same lags like normal skills. The spec was unplayable because of that issue.

Silly question, but say if you had 50 guys, you cast an AoE on them that affects 10, wouldn't the server have to calculate who was in range for the 50 regardless of who it decided got hit? In other words, regardless of if the skill had a cap of 10 targets or no cap, it would have to calculate distance for each opponent anyway to know who was in range of the skill. The only part it would save would be the allocation of who got hit, but it then leads to the question of if you hit a boon corrupt on 50 people as opposed to 10 would processing the extra targets take up the computer time saved by not assigning targets?

Edit; Personally I'd rather remove all large AoE from the game, and have everything affect single targets or have the effects bounce from one target to another instead like engi pistol 3 or thief shortbow auto. Would be a lot better in terms of visual spam and server time etc imo.