What makes GW2 combat in WvW fun? (and how I feel recent changes go in the wrong direction)

[JusticeRetroHunter.7684]

17 hours ago, DeceiverX.8361 said:

I understand computational complexity because I have a degree in CS and work on embedded systems as my day job.  Based on no evidence of ANet's server code, but simply an understanding that skill effect application is a very trivial problem to solve, I believe the lag problem is likely caused on the data validation layer of the game, which is a consensus algorithm as noted and requires significantly more computational resources.  I also suspect there are ways to improve the existing algorithm for WvW, because I do not believe it was developed with the current game state of boons and timed-effects like mass conditions in mind.  I have a slight hunch on its current implementation, which I believe the lag is caused by limitations of some specific lock designs often used in parallel processing when writing concurrent real-time systems.  And that this validation layer is utilized game-wide, and the original developers did not account for the use case of the current day WvW setting for how it handles processing.  And rightfully so; writing the code for this is extremely difficult, time-consuming, and can easily break a lot of stuff.  Post-game-release, this is likely not something to be fiddled with if at all possible, and there are no guarantees other solutions would still be truly sufficient.  I actually think ANet's programming staff are largely very capable, and so I suspect a solution with any meaningful performance improvements to be non-trivial.

I still disagree with your assessment that it's the source of anything pertinent to the boon-dependent state of the game.  Smallscale has the same issues with nearly an identical meta and the skill target argument does not apply.  Boon builds and supports themselves simply over-perform on a numerical stat-level and while theoretical unlimited target caps would likely influence how large groups of players interact by reverting to pirate ship metas, I do not believe this would have any impact on the build choices utilized in any meaningful way, because even when distributed away from stacked groups, the builds often used are almost identical conceptually.  My roaming guild ran 4 supports to one DPS like seven years ago because it was optimal even in 5v5.  5x Cele ele was optimal in sPvP back in 2013.  None of this is new; boonstack builds have *always* been OP.  Just concentration makes them very easy to maintain their performance and peak efficacy with, and unlike in 2013, a single necro can't just corrupt once and end the gravy train for a full 10 seconds.

But we can agree to disagree.

I think it’s very obvious to understand that these problems are in vastly different problem spaces from each other and that the basis of this target cap issue is something that exists beyond just the scope of the game…which is why it is the source of these problems.

Just to say this again: the problem of target cap is a geometric problem between packing infinite things into infinite density in a finite area. It increases exponentially with growing number of players and that has a significant effect on how skills interact with each other (damage dispersal), how they get balanced AND how it get computed by a computer. Anet never said that their computational complexity issues were caused by boons or stats.

Boons and Stats alone is just some separate problem in a separate problem space with completely different parameters…none of which at base level grow exponentially in problem space.

The point of this is to talk about why people group up and ball together right? How would deleting concentration alter why people crowd together? And the answer to that can’t be anecdotal…it has to be some logically provable fact or reason. You already alluded to why this can’t be the true source somewhat yourself already in your previous comment: this game existed for a time without concentration and without much boons and people still stack as optimal strategy.

 

Edited April 29 by JusticeRetroHunter.7684

[DeceiverX.8361]

3 hours ago, JusticeRetroHunter.7684 said:

I think it’s very obvious to understand that these problems are in vastly different problem spaces from each other and that the basis of this target cap issue is something that exists beyond just the scope of the game…which is why it is the source of these problems.

Just to say this again: the problem of target cap is a geometric problem between packing infinite things into infinite density in a finite area. It increases exponentially with growing number of players and that has a significant effect on how skills interact with each other (damage dispersal), how they get balanced AND how it get computed by a computer. Anet never said that their computational complexity issues were caused by boons or stats.

Boons and Stats alone is just some separate problem in a separate problem space with completely different parameters…none of which at base level grow exponentially in problem space.

The point of this is to talk about why people group up and ball together right? How would deleting concentration alter why people crowd together? And the answer to that can’t be anecdotal…it has to be some logically provable fact or reason. You already alluded to why this can’t be the true source somewhat yourself already in your previous comment: this game existed for a time without concentration and without much boons and people still stack as optimal strategy.

 

There are two falsehoods here:

The problem scope isn't infinite and when optimizing for performance problems in computation, you can't make assumptions about infinite sizes.  At most, you have a theoretical maximum of 210 participants in a fight if all three servers have every player on the map engage simultaneously.

- Your rant about player "density" going on about collisions and people occupying the same space literally does not matter.  Again, the calculation for whether or not targets are in a given radius and having things affect them and modifying fields and things within areas is *trivially* easy both programmatically and computationally.  Basic orbital mechanics problems and playing around with gravitational fields and collision detection with hundreds of entities was one of my CS101 projects in college.  In a slow interpreted language.  The lag has *nothing* to do with "infinite density" of players physically standing on top of each other.  It's why lag literally did not happen on objectively worse hardware years ago despite the same number of players being in WvW doing the same blobbing before major shifts into boon/condition application in the game, even when you had AC's with 50 target cap before they reduced them.  It's why the lag sometimes starts before the blobs even hit each other while they're stacking up.

 

The second falsehood is a total misunderstanding of two major reasons people stack up and how leadership realistically works at this scale.

- People have always stacked for spreading boons predictably and easily.  Whether it's 5 or 50 in a squad, whether it's by blasting combo fields or spamming Tomes, you need proximity to party members supply boons.  Generally, builds with poor sustain and ranged/AoE damage or boon denial have poor access to AoE defensive boons.  Builds being supported and healed this way are more or less invincible against non-damage-dealers, AKA, a traditional "frontline" which used to CC and provide supplemental damage.  But because of boon uptime and so much support making even a squishy DPS immune to a Minstrel Firebrand's paltry damage and unable to be CC'ed or affected by impairing conditions, there is way more strategic value in simply not leaving the DPS at all; all it does is serve to potentially cause a missed boon, invuln, or heal, which remember, without, is instant death to that player and probably the whole group over time because there won't be enough damage to drop the other group.  People stack *because* of boons, not in spite of them.  Real warfare tactics involve tons of split charges and flanks, feints, and ambushes of huge numbers of soldiers instead of a sprawling mass.  Because even diverting attention plays a huge role in winning an engagement.

By stacking, you eliminate any variability.  Every group now performs at peak every second, and there's no risk of people faltering.  Nobody can accidentally run out of range or get left behind.  Nobody worries about getting pulled away or immobilized or the pin getting sniped.  Stand still, don't move, and keep the support going.

- This extends to the greater psychology of reliable play, wherein if you have fifty people to lead, you can simply have them stack on your pin as a commander, and every single subgroup will have their optimal play rolling at all times.  There's no scrambling of groups trying to find their party members.  There's no chaos of someone missing a stab cast and getting pulled in while four people try and dive after a mesmer pull to get their healer or DPS out.  Stacking makes leading and surviving just easier across the board because boons matter more than any positioning.  Yes, uncapping targets *might* assist with this by trying to get people to split up, but this assumes the huge stacked group isn't going to fire or focus back.  Realistically, how many Wellomancers, OWP Barrage Soulbeasts and and staff eles will be able to throw down multiple Barrage casts and Meteor Showers and similar AoE's on a group of 50 without being similarly focused and blown up as soon as they get in range to start casting?  Every meta this has happened in, specifically during boonhate metas like what happened with Scourge/SpB at launch, the groups just disengage as soon as an offense is mounted, or they stand at range and blow up the first person who gets impatient.  This isn't PvE where the boss stands largely still and gets beat on over time.  When offensive boonhate metas take over, you have to consider the large group acting as reacting with the leader who does move quickly and does avoid the AoE immediately.  They actively avoid getting into mass scale aggressive plays because it poses too much risk.  This is why ANet keeps boons over strips and corrupts if for no other reason than the impacts to the willingness for people to fight at large-scale.  It results in gameplay where large groups just run around buffing and avoiding each other until someone says "screw it," tries to advance, and (usually) dies.  Why?  Because they don't have a boon advantage since their reliability from the first point is shot and there's chaos on the field.

As soon as there's a forced need to cause people to spread out, you're changing the tactic, but not the builds or reasons why people build that way.  The boons will still dominate.  It'll just be a cluster of a mess for a while as groups try and cope with having to change leadership strategies and coordinate differently.  But the builds stay the same.

[Verdict is Vengence.6912]

On 4/23/2024 at 6:03 AM, ZTeamG.4603 said:

you can always be in for a rude awakening from a class you never see or would initially write off as "low tier"

This is why I play meme builds. You have no idea how many havoc players I've ended with glass rifle warrior - a warrior with no survivability, all offensive signets, everything put into [Killshot], protecting my guys with sniper one-shots.

Gankers won't let people return to battle from waypoint? Killshot.

Uncatchable perma-stealth thieves skulking around a freshly captured objective? Killshot.

Bunker style, boon heavy, players tanking everything thrown at them while conditioning people to death? Killshot.

Wvw opens up when you go against the grain. I've become in my own head a protector of the roads and anti-ganker all because of the meme builds I choose to play. I'm instinctively doing the "R - Roleplaying" in MMORPG. But in a zerg, I'm just another cog in the machine, following a tag and their tactics. Which is fine and still very fun but less fulfilling.

[JusticeRetroHunter.7684]

7 hours ago, DeceiverX.8361 said:

- Your rant about player "density" going on about collisions and people occupying the same space literally does not matter.  Again, the calculation for whether or not targets are in a given radius and having things affect them and modifying fields and things within areas is *trivially* easy both programmatically and computationally.  Basic orbital mechanics problems and playing around with gravitational fields and collision detection with hundreds of entities was one of my CS101 projects in college.  In a slow interpreted language.  The lag has *nothing* to do with "infinite density" of players physically standing on top of each other.  It's why lag literally did not happen on objectively worse hardware years ago despite the same number of players being in WvW doing the same blobbing before major shifts into boon/condition application in the game, even when you had AC's with 50 target cap before they reduced them.  It's why the lag sometimes starts before the blobs even hit each other while they're stacking up.

Density matters on the computational side because there’s no cutoff for the number of players that can occupy a finite region. Pretend there was player collision, and said collision was 120x120 units. Then in a 120 radius circle you can only pack 4 people into that region ergo spells with 120 radius will only target up to 4 people.

when you don’t have collision there’s no upper bound on how many people there can be in the circle, therefor no upper bound on how many targets a spell can hit, therefor no upper bound on how many computations you’d need to do for a given spell.

That holds true for any computer program: you always have to have some meaningful discrete cutoff. If you could, you would have made a hyper computer, solved p=np, won a millennium prize and solved quantum mechanics lol.

Edited April 30 by JusticeRetroHunter.7684

[ZTeamG.4603]

On 4/29/2024 at 11:12 PM, Verdict is Vengence.6912 said:

Wvw opens up when you go against the grain. I've become in my own head a protector of the roads and anti-ganker all because of the meme builds I choose to play. I'm instinctively doing the "R - Roleplaying" in MMORPG. But in a zerg, I'm just another cog in the machine, following a tag and their tactics. Which is fine and still very fun but less fulfilling.

Agreed 100%.