WvW Performance and Map Cap Testing

[Yuffi.2430]

Having been the sole player on a map (pre-linking) I can absolutely guarantee that reducing the numbers of players sufficiently will stop lag - but this approach does make the game unplayable.The big question has to be why lag occurs. Even if it's server performance there will be a reason why the server can't cope, so it's worth trying a few experimental weeks to gain data. We had golem week (which some players enjoyed and others hated but everyone knew was temporary), so some you could run some weekly investigations.

But rather than experiment with WvW borderlands and EB, why not open up EotM again and give players an good incentive to use it while you test ideas and take measurements. You'll still have "normal" metrics from the borderlands and you can trial all sorts of things in EotM - changing population caps, reducing visual effects, ban minis, reduce condi/boon to one stack and one tick, even banning stealth in case lag is caused by large numbers of invisible perma-stealth thieves spamming random skills: whatever you need to try, all on existing one map that's already part of/associated with WvW.

[SWI.4127]

I was skeptical, but I have to admit there wasn't any noticeable lag with the lower map caps. Queues were a problem considering it was reset night, and I know it's not a long-term solution, but Anet admitted as much. So from my personal experience from 1 night, it seems to have made a difference.

[Custodio.6134]

Well, as ANet already stated, this is a temporary test, and not even close to a solution. It is more like, putting a band-aid on a cut-off hand. It reduces the symptoms (lag/skill-lag) but does not fix the initial problem(s) causing this behaviour.

There are multiple ways to adress these kind of problems (at least by the thinking of an end-user without programming/server-administration- or any related knowledge). To name a few:improving the servers overall (processing-power, internet-connection etc)improving the client (code-optimization, reduction of needed traffic etc)while reducing the maximum player-count per map, add a new (or maybe multiple new) map(s) in WvW where the people can split up. this would perhaps help for both, reducing server-lags while lowering the queues (if the maps are properly designed and worth playing the same as playing the current borders/ebg).

make Edge of the Mists really worth playing again. This may need a full redesign of EotM though, but some important points to consider in this process would be (imo):Full PIPs in EotM,a revamp of the map, to make use of the Warclaw, perhaps gliding tooa better way EotM influences the Mist War other than just dropping some supplies to the borderlands

also: as players we don´t really know how the employees from ANet, that are working on all game-related stuff, (especially WvW) obtain their information of what is going on in WvW. But, as these issues are existing for YEARS now, i conclude they don´t really play the game/mode they are working on very often and/or are understaffed.

Additionally, although this might be a bit off-topic, i consider "WvW-Performance" being more than just lags. The problems are much more profound in the Servers and the Client itself. The Lag-Issues (Rubberbanding and Skill-Lags) are a Server-Sided issue. But the other huge issue is the Client-Performance and Hardware-Utilization. The Game doesn´t even close take advantage of current hardware. This is a state, that already exists since release. I wonder why it took ANet so much time to finally take action.

My assumption: ANet didn´t (or perhaps still doesn´t) really care for the game-mode we (as the people playing WvW as their main game-mode) love so much. And still we play WvW, perhaps because there is no real alternative game ppl could consider to leave GW2 for. And the latter is both, a huge advantage and at the same time a huge risk ANet is taking.

[Solanum.6983]

I'm okay with reducing the player cap as a band-aid fix until the real issues are sorted for now, I just hope that it wont take too long but between the whole Alliances situation and "4-6 weeks" balance patches I'm kind of skeptical.

[SexyMofo.8923]

The lag is still there, and the game feels emptier. Your solution will only kill this game mode further. I’m sure your performance engineer has given you tons of possible options and you decide to choose the cheapest one.

[Sovereign.1093]

@"Yuffi.2430" said:Having been the sole player on a map (pre-linking) I can absolutely guarantee that reducing the numbers of players sufficiently will stop lag - but this approach does make the game unplayable.The big question has to be why lag occurs. Even if it's server performance there will be a reason why the server can't cope, so it's worth trying a few experimental weeks to gain data. We had golem week (which some players enjoyed and others hated but everyone knew was temporary), so some you could run some weekly investigations.

But rather than experiment with WvW borderlands and EB, why not open up EotM again and give players an good incentive to use it while you test ideas and take measurements. You'll still have "normal" metrics from the borderlands and you can trial all sorts of things in EotM - changing population caps, reducing visual effects, ban minis, reduce condi/boon to one stack and one tick, even banning stealth in case lag is caused by large numbers of invisible perma-stealth thieves spamming random skills: whatever you need to try, all on existing one map that's already part of/associated with WvW.

Here here. Make eotm great again. :3

[Baldrick.8967]

@suialthor.7164 said:Addressing lag is nice. How about looking into ways to avoid hour plus queues?

Edit: I've been 1 in queue for nearly 40 mins with a discord of people still wanting inEdit 2: I've been 1 in queue for 60 mins with a discord of people of still wanting in. At this rate some people are on track for multi hour queue.

Now if only they hadn't just stolen the name Guild Wars and actually followed on from Guild Wars. We could have had instanced battles, guild battles of various sizes, fighting for your faction, etc. ...

As for reducing population cap, complete cop out. Wasn't there something in the original blurb about 100's of people a side having epic battles- and now we're down to what, 50 or 60?

How about instead they look at all the garbage that clutters up the screen, crazy amounts of condis, boons, strips, blah blah and strip some of that computation out so the servers can cope better.

Or how about you actually use a TEST SERVER and allow players to test things before general release like other companies do.....

[kathy.8291]

Is this AnOTHER way to Kill WvW. When you are capped out and cant run with your guildies will stop people from playing WvW that they enjoy. Alot of us DONT like to pve and just press no 1 skill. This will NOT solve the lag problem in WvW. Everyone should just turn down their graphic card to None and see if its your internet. I have NO lag even on reset and we are qued.

[briggah.7910]

I played 1 hour in EBG on NA and had no lag. There have been plenty of resets that I didn't get to play because I couldn't get into a map fast enough before this change. The queues were still +20 when I logged off. Still was huge blobs running around that rotfstomped the smaller zergs. During the week it will go back to ebg being queued at prime time and 3 other dead maps.

[Blocki.4931]

@Phlogistn.7893 said:If you reduce the size of players on the map with already full servers and links to those full servers that are almost full I can guarantee you that you’re going to have some very (more) angry players. Many players are already waiting hours to get on at reset. All maps are already queued.

This is simply a really lousy hotfix to a problem that’s been ongoing for quite some time.

And yeah, Alliances.

This Is Not The Way

And the very obvious solution is to not overstack servers or leave them if you are not okay with waiting. Almost like there is a drawback to stacking.

[Tiawal.2351]

More testing needed, but there was already less server lag this reset, though the scourge "balancing" also contributed to this.

While huge fights could be "epic", we had enough time in almost 8 years to see that the engine used by Guild Was 2 can't do it.Lowering numbers can't please everyone: some rather have 120 in a single squad, with no skills working because server lag, and would still "enjoy" their time (the social part on voice chat), while others would love a 15 vs. 15, truly competitive lag free game mode instead.If all 3 sides have about 40 players in one spot (in a 3 way fight; for a keep for example) and they aren't "lazy" pressing buttons (like many people are, so not every group will put same load on the servers), then the servers will end at 100% CPU load capacity easily. I would recommend just 25 each side, but because the cost involved maybe around 40 is the best compromise.

Reducing boon & condi durations, as well as other skills that contribute much to heavy server side load, would allow more numbers to fight without server lag. These changes would help even the PvE mode, but WvW really needs a skill & trait rework, focused on reducing server lag.

[blackgamma.1809]

i mean, Amazon servers uses SMR hard drives that are designed to be slower

just dont cheap out on servers, and makes troubleshooting heaps easier

[Noah Salazar.5430]

Can we somehow support game, or donate for better servers? ther culd be collection displayed in forum wher evryone can donate 5,10,15,20,25 euro for poll and when we hit mark (whos showed for evryone) you culd buy better servers for some time?

[keenedge.9675]

You could consider weighing the cost of staff time adjusting an underperforming server versus the cost of a higher tier AWS home for WvW instance maps.

[Moira Shalaar.5620]

@"Blocki.4931" said:And the very obvious solution is to not overstack servers or leave them if you are not okay with waiting. Almost like there is a drawback to stacking.

This response completely ignores the effect that bandwagoning has on those players who were already on a server that gets hit by relink server stacking. My guild and several allied guilds are all on the same server. Our queues were more or less livable prior to this link yet now I am waiting 60-90 minutes to play with my friends (and no, not only on reset night). That long of a wait is entirely because of the stacking that happened post link. Your suggestion is that I should leave my server since I hate waiting that long to play WvW, but leaving my server with all my friends is tantamount to leaving the very reason to play WvW.

As I understood it part of the purpose of the long announced and never fulfilled Alliances feature was to be able to reduce or handle the bandwagoning effect. Until the Alliance vaporware materializes, Anet needs to find a different solution to server stacking other than your "if you don't like the wait, move" solution that primarily penalizes everyone who was perfectly happy with their server community.

[DemonSeed.3528]

I think one thing we may not have considered is that Anet may have done other things under the hood besides just lowering map caps - I'm curious if they've done server adjustments in conjunction with the caps and slowly start easing the population back in during the testing. It's just a week long test so I'd rather the frustration with queues if they can generally smooth out server performance. If they are only doing the map caps with nothing else I would be very surprised (or worried). If that fails then worst case scenario is to start hammering at skills that affect server load balance. It will be awesome if they were able to crunch this one out and makes me hopeful for future, as I'm firstly fighting lag instead of players.

[Caedmon.6798]

So far i notice my FPS seems more stable in blob fights. Normally there was this hitch when a blob just appeared, and the occasional hitches during fights,that seems gone. I was also able to hold over 60 Fps in a 2 way blob fight just now,fps seems overall more stable. But ping wise i dont think i notice much difference yet. Still suffered from some delays yesterday.

[Dana.4760]

"" I'm glad you are asking for opinions and worrying about a wvw issue. GW2 is a great game, unmatched and with people dedicated to wvw that we have great interest in that you work better and observe from within the needs of the game. I think that in terms of performance, servers should be improved and programming should be optimized, experts should invest in that and we could collaborate. We love to play zerg v zerg v zerg and without lag .Invest in mcm, hire someone who plays regularly and knows the communities, and the servers. Will be worth. "" Me alegro que pregunten opiniones y que se estén preocupando por un problema de wvw. GW2 es un gran juego, inigualable y con gente dedicada a wvw que tenemos gran interés en que funciones mejor y observen desde adentro las necesidades del juego. Creo que en temas de rendimiento deberían mejorar los servidores y optimizar la programación , los expertos deberían invertir en eso y nosotros podríamos colaborar. Nos gusta jugar zerg v zerg v zerg y sin lag. Inviertan en mcm , contraten a alguien que juegue habitualmente y sepa de las comunidades, y los servidores. Valdrá la pena. ""

[zealex.9410]

@ASP.8093 said:

@zealex.9410 said:Its sad pvp and wvw cant have a more substantial split from pve balance so you guys could change the function of skills and even have a pvp/wvw trait page.

Maybe if cmc tries to talk to the balance team to reduce the boon and condie bloat :/

If it's boon/condi/triggered-ability bloat that's triggering performance issues (which is a theory I find highly credible, certainly), I think you'd want to also reduce them in PVE, tbh. Because the big event-driven maps are also going to be prone to similar issues.

Knowing how rude this might sound id say that this isnt as necessary because in big ow maps ppl tent to only press their aa.

That aside, i dont see a reason not to change it in pve but that would also require maybe at look at all the endgame content additions made post hot as less skilled groups could struggle too much to do them. Fo me the encounters would turn jarder which id love but im not speaking for anyone.

[Camdal.6197]

I experienced more lag this reset than earlier. (it seemed as if fewer people triggered the lag, ie, two smallish blobs was enough, rather than a 3 way).

Q's were notably longer, but that cannot have been just 7 less people per map. I think it was people logging in to check if there was a difference.

[Deaths Sensei.5317]

@suialthor.7164 said:Addressing lag is nice. How about looking into ways to avoid hour plus queues?

Edit: I've been 1 in queue for nearly 40 mins with a discord of people still wanting inEdit 2: I've been 1 in queue for 60 mins with a discord of people of still wanting in. At this rate some people are on track for multi hour queue.

Sorry, but this is a side-effect of stacking a server with wvw guilds.

[Deaths Sensei.5317]

@Moira Shalaar.5620 said:

@StARlORD.2736 said:Another thing to worry about is there are some big servers paired with big servers, for e.g.: JQ + BP, Mag + SoR and FC + NSP. No disrespect, but ya guys need to pair servers wisely, not blob servers + blob servers and blob servers + medium to high servers. This is why skill lags happen, because of this. And pugs are also a wild card.

SoR was Medium population until being linked to Mag, after that the bandwagon hit us hard. Since the change from Medium to Very High and then back down to High all happened post link announcement, that at least can't be blamed on the link matching formula.

They should lock server transfers 1 week before relinks, and for 3 weeks after (1 month in total), could help with bandwagoning

[Alexander.7694]

@Cal Cohen.2358 said:The key feedback that we’ll need from all of you is to get a sense of how the maps feel with fewer players.

Maps feel empty and less fights happening, queues longer to enter as expected. I can't comment on lag, seems same to me.

We have to consider that the main blob that might experience lag is not the total number of people on the map, some are roaming and doing other stuff, so most likely the cap on number of players won't affect lag much. If we have lets say 50 people on a map, they won't all be in the same place. If we lower them to 40, could be that the main blob fight is let's say 30 in both cases. The roamers and idle won't affect. My 2c

[suialthor.7164]

@Cal Cohen.2358 said:The key feedback that we’ll need from all of you is to get a sense of how the maps feel with fewer players.

When we spend more time in queue than actually playing it is hard to give feedback. So far the queues are worse than usual. Last night the queues for all battlegrounds except Edge of the Mists were multi hour. I was first in queue on one map and it took a little over an hour to get into it.

Today (early afternoon) one map is full (team mates and opponents) with a multi hour queue, and the other maps are either empty of team mates or empty of opponents. This isn't fun, so I closed the game. Which I suppose is one way to reduce lag.

[Svarty.8019]

@Helicity.3416 said:

@Mini Crinny.6190 said:

@Helicity.3416 said:Nice.

Even longer queues for what will

almost certainly

not give a measurable performance increase.

so you would rather them do nothing and people to continue to complain about it?

I would rather they actually look at what causes it (it's boons and condis) than attempt the lowest effort possible "fix" (really, it's boons and condis) to attempt to make things work better (it's the boons and condis, look at the boons and condis) yes.

If I understood Raymond's explanation on the livestream, ways different skills/effects interact with one another.