Guild Wars 2

The issue of needing to wait a little while to use a two-factor auth code should be resolved. Please let us know if you still have any time-based issues with two-factor auth.

Just in case, I’ll also remind anyone who altered or removed two-factor auth from their GW2 account that if you want to re-enable it, you also have to reset/re-synchronize your mobile app’s code for GW2.

A bug potentially affecting the last tier during 1-up/1-down matchmaking has been found, and a fix has been submitted. Until that fix can make it to the live game, we’re returning to the previous style of Glicko sort-and-shuffle. The change will first be noticed with this Friday’s reset.

This is being monitored and investigated.

There’s possibility for as much as one more instance of this problem to happen per non-Obsidian Sanctum Mist War map, per tier in NA, if it happens before this Friday’s reset. If this Friday’s reset happens first, we should be in the clear until a permanent solution can be found.

This has been resolved for now. I’ll keep an eye on NA’s WvW reset tonight in case the same issue triggers there.

We’re aware of a bug where loading screens can sometimes take 2+ minutes to complete; more often after playing in WvW for a while than elsewhere in GW2. We’re working on a fix in such a way that it can be selectively tested on Live, since it affects a vital area of GW2 behavior (loading screen timing), and is rare/difficult to reproduce in a testing environment.

In the meantime, if there aren’t queues in the maps you’re playing in, there’s a possible work-around. Switching maps— either by exiting WvW or just moving to a different WvW map momentarily— appears to get your client “unstuck” from the long loading screen behavior.

There’s been a hitch in the plans of worlds getting hitched, such that this World Link update won’t happen until one week later, on January 6th. This won’t delay/offset future world link updates.

The maintenance is complete.

WvW is available again, and today’s maintenance is complete.

mako.6493:

Not quite sure what this means for the week. I get restoring and cancelling the 5 minute round glitch. However since i cant remember what all the server glickos were at the end of the original matchup lets say FA was 1 880.000. Is that glicko rating at the end of this mathchup gonna be 1 880.000 regardless what happens the next 3 days or so? Will the glickos fluctuate during the week as normal based on score or is it frozen until the next roll at reset?

At the moment the “real” (but not publicly visible) Glicko ratings are where they would have been had the second, extra round end not happened.

After this week’s round, Glicko ratings will be updated using those “real” values and this week’s Victory Points results as inputs. Ratings will be updated in that way before the new matchups are determined.

Due to the double round ending throwing Glicko ratings significantly off from what they should have been, they’ve been restored to what they were after the end of the first of the two recent round endings; the two that were 5 minutes apart from one-another. This won’t be visible in the official leaderboards during this round. These values will be used when updating Glicko ratings when this current round ends.

The root cause of the bug was found. This shouldn’t happen again.

Edit: We’re back to using the new Victory Points for ranking the results of WvW matches.

The root cause of the bug is still being investigated. Until this can be fixed, scoring the end of the week’s round has been changed back to using accumulated War Score, instead of the new Victory Points.

We’ll have a fix for this as soon as possible, and will look into what will be done with affected Glicko ratings and EU’s current matchups. Until then, one thing that’s known is that not ranking the end of a round by Victory Points prevents this bugged behavior where a round ends, then ends again 5 minutes later.

Hexinx.1872:

So how does the game code respond if a player is in a squad in a WvW map… and logs directly out…. but the commander leaves them in the squad so the rest of us all seee this guy who is offline. Does that get removed from the calculation or does the capacity count the total squad? I would assume offline accounts have some mechanism that ‘kicks’ them from both the map and the pop count.

That’s correct. We just did a little “sanity check” to make sure no new changes had messed something up. Players that still appear in squad UI— but aren’t online— don’t count toward the map’s team counts for either the Outnumbered buff or the queues.

trueanimus, I just checked our logs for the counts of each team, as output by the same area of code that calculates the need for and applies the Outnumbered buff, for the period of 2016-09-03T02:00:00+00:00 to 2016-09-03T05:30:00+00:00 in NA’s T1 matchup.
This is from the last pre-skirmish match’s opening reset in NA to 3.5 hours later. Other than Dragonbrand being Outnumbered starting at roughly 04:45Z (2.75 hours after reset), no team in that match should have had the Outnumbered buff for that period.

That said, right there I’m only looking at the logs of team counts and how the ratios play out. That doesn’t rule out the possibility of a bug with when/how the buff is being applied. We’ll keep an eye out around reset times in-game.

Please do let us know when you see something that seems wrong with the buff, with as much information as possible. Even if it’s something we can know by our logs, sometimes it can save some time if it’s listed, so it’s not an extra lookup of something on our part. At the least, I’m looking for: your team (any one or more of your color, your world, or your host world will do), time/date with timezone.

Regarding how Glicko will be updated at the end of a skirmish-based match. The same inputs and outputs will be used as usual, except that total Victory Points from the week’s match will be fed in as the score for each team, instead of the week’s total War Score.

This can have different effects on how ratings will change. If team A consistently beats team B— but only a little— their ratings will be split apart faster than before. If team A steamrolls team B consistently, their ratings won’t split apart as quickly as before.

Keeping scores closer together through the game’s mechanics keeps moment-to-moment gameplay more competitive and interesting. But, if teams facing one-another that really belong in different tiers have more similar scores than before at the end, their ratings updates won’t reflect that as quickly.

The effect Victory Points have had versus traditional War Score for matches prior to now are too mixed in the final relative results to say yet if this will definitely have too strong or weak an effect in any one direction on the week’s final results. Not without seeing how they affect people’s behavior, as well.

Regardless of how they affect Glicko ratings in their first Live run, if skirmishes improve the moment-to-moment feeling of being in WvW, we can work with updating how ratings (and more importantly, match-making) are determined.

[Edit: Removed a confusing negation.]

WvW is now available in build 66,577.

Thanks for the information. I’m looking into this. If you’re not in build 66,536, you’ll be having this issue.

difficultshame, Aileras, TingleTangleBob, which server(s) are you on? Are you getting the error when you open the WvW dialog, or when you try to join a map? If while joining, which map?

Dayra, I’m seeing:
Aurora Glade: 5 VP
Abaddon’s Mouth [DE]: 6
Seafarer’s Rest: 4

So each were awarded 3 VP from an initial setup skirmish that went off, in which they were all tied. Then Abaddon’s Mouth won the first “real” skirmish, and Aurora Glade took second.

The display for PotM bonuses in the WvW dialog’s second tab is incorrectly showing what bonus would be present with the War Score from only the current Skirmish.

The tooltip on the WvW button in the quicklaunch bar along the upper-left of your screen still shows the real PotM bonus you’re currently receiving behind the scenes.

Skirmish scoring will go into effect in about 15 minutes (near 18:00Z), with the EU reset.

NA players: while you’ll immediately see the new skirmish data in the WvW dialog, this week’s match will still have its end result scored in the old War Score-only style, ignoring Victory Points from skirmishes.

Sites that use our API for WvW will still display the “vanilla” War Score that accumulates over the week.

Jeknar.6184:

… why not reset the ratings when the links change since the population will be shuffled around?

Originally with world linking the thinking was that earned ratings themselves wouldn’t be directly altered, and an increase in Glicko deviation and volatility would take care of ratings needing to shift significantly when their composition of worlds/players changes. You can see from rating development over time that this is somewhat true: http://mos.millenium.org/na/matchups/histories

Unfortunately that doesn’t solve (and even makes worse) the problem of worlds drifting too far away to be shuffled up/down tiers from the bottom/top ranks.

We intend to have a poll in the future asking if players would prefer that Glicko ratings be completely reset with each relinking of worlds.

• Will Glicko ratings be adjusted before or after a match?
  Glicko ratings are temporarily adjusted at exactly the time of matchmaking. The adjustment will never be visible. We’re not actually modifying any world’s rating, just preventing matchmaking from letting worlds drift away from the others.

• Why not just change the links again?
  World links change on a schedule to maintain some stability. Among other reasons, those who maintain voice chat servers don’t have to redo permissions every week. We also wouldn’t want to start shifting world links around on a subjective whim. The frequency of the schedule was previously determined by a public poll.

• Why have Glicko ratings?
  Strictly on the functional side of reasoning: matchmaking. No doubt within the space of all possible ways to calculate matchups in WvW (both pre-existing and unwritten) there exists one or more better solutions. Matchmaking in WvW is a very different scenario from what any matchmaking system I’m aware of aims to solve.

Other matchmaking systems, such as 1-up-1-down, have been discussed in the past. We’ve not previously been convinced that they’d be significant enough improvements to prioritize them over other work. One of the bigger problems with 1-up-1-down is that at natural barriers, two teams would commonly swap week after week, leaving three teams (the one that went up, and the two that remained where it was) getting steamrolled every other week.

Since alternate matchmaking systems require some of the same peoples’ time as the scoring improvements we’re currently working on, Glicko alternatives or adjustments may be options in future polls once the scoring improvement work is done.

Currently some worlds, such as Crystal Desert, are stuck in what’s come to be referred to as “Glicko hell”. This is where a few worlds’ Glicko ratings drift off from the rest of the group, leaving a wide enough gap that those drifting away won’t be matched up against the rest.

To help prevent stale matchups, we’ll soon be able to artificially adjust Glicko ratings solely for the purpose of matchmaking.

For example, if we use this for NA’s lowest tier of worlds (T4), bumping them up to be just below T3’s ratings, they’ll have a decent chance of being shuffled into T3 during matchmaking. Given the ratings today, an adjustment we might make for July 29th could look something like:

T3:
- Maguuma: 1,838
- Sea of Sorrows: 1,777
- Stormbluff Isle: 1,767
T4:
- Crystal Desert: 1,512 (plus-185) = 1,697
- Sorrow’s Furnace: 1,398 (plus-130) = 1,528
- Darkhaven: 1,366 (plus-100) = 1,466

However, note that this artificial adjustment is invisible to anything but matchmaking. So if a T3 world doesn’t beat a T4 world significantly, the T4 world’s actual Glicko rating will increase more than usual, while the T3 world’s actual rating will decrease more than usual. Both worlds will also have increased deviation and volatility after. This is all because the T4 world’s lower rating means to Glicko “this world is expected to lose significantly to that T3 world”. So in an upset where this doesn’t happen, Glicko both makes larger adjustments to rating, since it was shown to need change, and increases deviation/volatility, to tell itself on the next round that these worlds haven’t settled into their most accurate rating yet.

EU’s lowest-rated world is still within its deviation of the lowest-rated world in the next tier up, making EU much less likely to receive artificial matchmaking adjustment on July 29th.

smashie.3074:

Thank you Tyler and Chris for the additional information.
I have to say, even with the changes that are undesirable to ME, I welcome the communication and effort of the WvW team I hope this continues to go on.

Even with the additional information you’ve given us, I still have a problem that I’d like to see addressed. Unfortunately, it’s quite a brutal point in my opinion but a point that needs to be raised.

1) Toxicity and the effects on the community

If many of the WvW posters in the forum are anything to go by, there are many individuals here who constantly berate people of their idea of WvW, their idea of “blobbing”, what WvW should be, or how “server pride” or whatever reason you’d like to include.

The very same people are also telling people to quit the game or WvW.

Speaking for my girlfriend and I, we simply do not want to play with these players.
And I’m fairly certain that many other people would feel the same way.

Problem with the server link, because it’s going to be updated quarterly or whatever the timescale is, there’s a danger of being linked with server with THOSE people. If the toxicity of each server gets to server wide instead of just individual spats, the teamwork that’s required to play WvW is completely eroded. The match-up for that duration will stale. Only the server with those individuals will continue playing.

That’s why, it’s still a resounding no for me

At the previous state of WvW, had we chosen a server with that kind of toxicity, we simply would have upped and left, if we had the gems.
Now, with the updates, we can’t even stay in one server and hop to another one, just in-case, the Glicko decides that we should be linked with the server we desired to leave behind.

I ask many of you to consider this scenario.
Perhaps it’s a very pessimistic view, but at the moment, it’s happening right now in my server.

Thanks for your measured response not claiming to represent an unidentified majority.

Your particular linked pair of worlds may be having a rougher time of things than most. By world, you’re among one or ones that have a higher than average percentage of “No” votes. So I’m hopeful it’s something more to do with your particular pairing, rather than something with linking itself, or with any specific players. Time and a change of linked worlds will help tell.

As a reminder, please do report anyone engaging in toxic behavior. There’s little immediate feedback beyond that player’s chat being blocked for you, but reports do help.

sionnach.5391:

Whatever happens, please address the ‘guest’ server identity in any ways which are feasible, and the more the better – in fact do away with the host/guest concept and treat both servers as partners. If we flip a keep, messages should show the server majority who flipped it – if that’s not feasible, then list both servers. Sounds petty but it really isn’t; this would go a fair way towards addressing morale problems in T7-8.

Not to disagree with the idea that server identity is important, but something that’s been interesting to see is that “guest” worlds are tending to have higher ratios of “Yes” votes in favor of keeping world linking. Which I’m definitely seeing as more of a “this improves my day-to-day gameplay”, and not as a “I don’t care about my world identity”.

One problem with showing both world names is that the strings of text can get very long. When/if more than two worlds are linked together, it gets very long. Much too long for the UI.

I do like the idea of showing, for example, “Devona’s Rest captured Durios Gulch”, if Devona’s Rest had more members involved in capturing the objective. Though for the moment this isn’t among our top priorities. Such a change requires programming support, and right now we’re mostly focused on the large task of scoring improvements previously voted on.

CrimeMaker.8612:

But if you make it an hour early with the daylight savings it would 8pm est ;-; more time for us est players to play. So basically with the current time the people who can’t play rn will only be able to play reset every 6 months?

Which is 5pm PST. Too early for those living on the West Coast.

Unless our internal R&D project is successful— flattening the Earth so there’s only one sunrise/sunset for everyone— there is no time which exists that will please everyone.

Reminder: WvW deliberately doesn’t obey Daylight Savings Time changes.

Reset times in NA are currently 7pm/10pm PDT/EDT.

On November 6th, when US time zones (those that obey DST) resume their “Standard” times, it will look like 6pm/9pm PST/EST.

Or, more tersely, it is always 2am UTC year-round.

Extra note: I had to review this post about three times to get it all right. For being simple offsets, it’s amazingly easy to mess up timezone math.

WvW is available again with the alpine borderlands.

WvW will re-enable and spin up Alpine maps not long after the new build is available.

Entering downtime momentarily.

This WvW maintenance period will be beginning shortly; likely within the hour.

dank.3680:

I would guess they call it linking because its not a solid merge, it’s a dynamic merge which could be changed on the fly.

Yep, this is exactly why “linking” is a good term for this feature. On the tech side, we could change world links every week with the WvW reset. Different links, more links, no links, anything. Players are still always on their original home worlds.

We don’t currently intend to change world links that frequently because of the logistical issues it would cause for players. Voice communication setup, commanders new to one-another working together, and so on.

So while it may by some perspective feel more like a destructive merge right now, when world links change you’ll continue to be with your closest allies on your home world.

FogLeg.9354:

They all show host server names. I was fighting against UW players today and they all had FSP tags now. I knew them only from their guild tags so, nothing will show which smaller server they from.

Thanks for answering, FogLeg. That’s exactly correct. The biggest motivator for showing only the “host” world in enemy nameplates was to help identify enemies quicker. With only two other world names showing up in a given week, it’s easier to quickly figure out which color team they belong to. Since world linking supports any number of “guest” worlds linked to a single host, it would potentially get pretty confusing to see a big zerg of nameplates with any number of world names, and figure out the composition of one team versus another.

As for not seeing an allied player’s guest (actual) world name anywhere, we wanted to be sure world linking wouldn’t create any avoidable friction between players. One of the major goals we have with GW2 is always making sure encounters with other players are positive experiences (my paraphrasing). So anything we could do to let guests blend in and only stand out if they felt like calling themselves out as such is likely to be a good thing toward that goal.
Plus, there actually just aren’t that many places in the UI where we show allied players’ world names, so this was an easy goal to keep for this side of things.

Helping maintain guest world pride is a little tougher. We could show each player their world’s name in more areas of the UI, regardless of if they’re on a guest world or not, but this may cause communication issues, with some maps and areas of maps being named after worlds.

I’m open to hearing any feedback people have on little ways we could improve the feeling of world linking, while maintaining unambiguous communication for players and avoiding any negative “tribalism.”

Virtute.8251:

Okay. How are Glicko ratings for individual servers being calculated during linked matches? Some of us are wondering what the trick is to separate the linked worlds during Glicko rating calc, so that new (different) pairings can occur at each (for now 3 month) pairing interval.

Linked “host” worlds have their Glicko data updated in the usual way. Linked “guest” world Glicko data isn’t affected. It’s not yet determined what approach we’ll use for Glicko data for guest worlds when world links are changed, in the case that a guest either becomes a host or goes solo.

Dream In A Dream.7213:

What about the glicko hell of low/top tiers? Will you prevent tier walls with this system?

As Glicko “figures out” an appropriate rating for a given world over time, both its deviation and volatility decrease. This is Glicko’s way of determining how well it “knows” where a world belongs, and settling it into place so it neither gets stomped by a higher-tier opponent, nor effortlessly dominates a lower-tier one. However, this intended purpose of Glicko is also a major factor in causing matchup stagnation.

For this first beta we’re raising Glicko’s deviation and volatility for each world, but leaving rating alone. This will cause a greater variety in matchups as Glicko now “thinks” it has less of an accurate rating, and will willingly shuffle worlds around more as it tries to sort them out again. EU will have particularly high variation in matchups since ratings aren’t spread quite as far between top and bottom tier worlds, as they are in NA.

When creating each week’s matchup we also do a little bit of random shuffling after Glicko sorts things by rating, with the shuffle being based on deviation and volatility. So matchups will again be more variant in the coming weeks. Though EU will have higher matchup variation than NA, because it won’t take quite as strong a random roll for a world to be matched against others in different tiers, due to EU having less of a spread in Glicko rating from top to bottom tier.

“Glicko hell” occurs when a few worlds have enough distance from the rest for these random rolls to never match those up against the rest. The difficulty of climbing back out of this comes about because of how Glicko determines how far a world’s rating should change after a round ends. If a higher-rated world beats a lower-rated one, this is as expected, so the winner’s rating doesn’t change very much; it’s already where it belongs relative to the opponents it just played against.

Due to the increased deviation and volatility combined with fewer total worlds, there shouldn’t be a “Glicko hell” again for a little while. We’ll continue to watch how Glicko rating, deviation, and volatility develop to see how well world linking— and our approach to it this first time with Glicko— plays out in the Live game. It’s a delicate balance to hit where there’s enough matchup variation to not be stagnant, but little enough to not create unfair matchups.

Status updates for NA reset time change.

3/25/2016 7:08pm PDT – Disabled; updating.
3/25/2016 7:29pm PDT – Enabling.
3/25/2016 7:33pm PDT – WvW is enabled again. Enjoy!

Status updates for EU reset time change.

2016-03-25 17:35 UTC – Preparing.
2016-03-25 17:59 UTC – Disabling and updating.
2016-03-25 18:35 UTC – Enabling.
2016-03-25 18:48 UTC – One-time hiccup with the time change. ETA pending.
2016-03-25 19:33 UTC – WvW is enabled again. Enjoy!

Here’s what the Heart of Thorns release will look like for WvW:

1. On the morning of Thursday, October 22nd, WvW will be terminating without completing or updating Glicko ratings as normal. This will be the end of the classic (“Alpine”) Borderlands maps for the foreseeable future.
2. On entering the build with the new Heart of Thorns content, all of your spent WvW Ability points will be refunded. Recall these can be spent differently per character, and spending them on one character doesn’t spend them on any others.
3. WvW’s reset time will be changed to be 24 hours later in NA and EU. It will now occur on Saturdays instead of Fridays.
4. At the moment Heart of Thorns content becomes available, WvW will prepare its maps. Including the new Desert Borderlands map. There’s no guarantee the matchups from earlier in the week will be the same as the new ones (or that they’ll be different).
5. The new round will run until the new reset times on Saturday, then end and update Glicko ratings. After it comes up again, WvW will be running as “normal” except for the changed reset times and the new Desert Borderlands.

The issue with siege bundles in-hand being randomly destroyed has been fixed. Siege bundles are now working as described in this post: https://forum-en.gw2archive.eu/forum/game/wuv/Siege-bundles-no-longer-drop-to-the-ground/5448736

The issue with siege bundles in-hand being randomly destroyed has been fixed. Siege bundles are now working as described in the original post in this thread.

We’re aware of and investigating the return of an old issue with siege bundles disappearing while being carried. Thanks for the details provided on where and how this has been encountered. Especially your thoroughness with providing a video.

We’re aware of and investigating the return of an old issue with siege bundles disappearing while being carried. Thanks for the details provided on where and how this has been encountered.

Previously, weapon-swapping, using a non-combat transform, or getting another bundle item would drop your in-hand siege blueprint bundle to the ground. Now, it will instead be returned to your inventory.*

*Exception! Guild siege golem blueprints are still lost instead of being returned to inventory or dropped on the ground. Whether by pack-up skill ‘2’ or any other means. This behavior specific to guild siege golems hasn’t changed. Feedback on this is welcome.

Reasons for making this change:

1. Play-for-free accounts otherwise had a means to skirt our restrictions on sending wealth to other accounts by buying siege, dropping it on the ground, and having other accounts collect it and sell it. Restrictions on outgoing wealth from play-for-free accounts helps make the accounts less appealing as gold seller/bot accounts, so we can’t have holes like this around.
2. That same exploit would work with Beta Weekend Event accounts. Beta characters/banks would be copied and forked from the normal data, siege blueprints could be dropped, then collected by someone else; essentially duplicating the items. This was previously blocked by temporarily kill-switching the ability to pack up siege blueprints.
3. Dropping siege on the ground when you suddenly, reflexively do something like trying to swap weapons felt very clumsy.
4. The ability to drop siege bundles on the ground was originally added before siege could be put on the trading post or mailed between players. Since better means now exist to transfer siege between players, this old functionality isn’t as important.

The cause of the “NPC coup” bug (enemy guards at some objectives) has been identified. We’re working on testing and moving the fix towards the Live game.

We’re investigating the ongoing issue with NPC guards spawning for the wrong team. It appears to begin affecting objectives after some period of time (or some event) after a map restarts (such as after a game update). If you see any objectives that aren’t experiencing the NPC guard spawning bug, or have any other information, please feel free to share it here.

A longer-term fix has been applied to the “World peace” problem. Players should now always be on their proper teams, asura gates and waypoints in WvW maps will be working again, etc.

The fix was applied at about 9:15am PDT/12:15am EDT/4:15pm GMT (May 20th, 2015). If you see any other problems, especially things like NPC guards or dolyaks misbehaving, spawning too often or not enough, etc. please post here with the map (Red Borderlands, Eternal Battlegrounds, etc.) and objective (Garrison, Bluelake, Godslost, etc.).

WvW matches have resumed. We’ve successfully ended World peace for the time being. Though there’s still a chance it will come back until we can apply more complete fixes to prevent it.

We’ve identified causes of most of the issues, and are working on moving fixes towards the Live game. I’ll update again when things are a little more settled. Patch notes should also call these things out if I’m behind those.

Thanks for your patience.

Additional issues (also added to my first post here):

• NPC guards spawning at wrong times and/or on wrong teams within objectives.
  • Is this only happening when the objective flips?
• Score reduction when some maps in a match reset/rollback.

NPC workers getting stuck is an issue I’ll take note of for later, but probably isn’t related to the immediate issues with WvW cross-server data we’re investigating now. We’ll have to take care of the World (vs. World) being on fire first.

Thanks for the extra reports and information.

Thanks for the information, all. We’re investigating WvW issues now.

Issues we’re aware of:

• Objective supply caps getting set to un-upgraded levels.
• “World peace” (everyone on a map is friendly to one-another). Happening in some maps. Eternal Battlegrounds EU matches with Gunnar’s Hold and Abaddon’s Mouth are known instances.
• WvW API not updating/giving data.
• NPC guards spawning at wrong times and/or on wrong teams within objectives.
• Score reduction when some maps in a match reset.

Issues for later:

• NPC workers getting stuck.

Please post here if you find anything else, have more information you’d like to add, or see more “World peace” maps where all players are friendly (be sure to include at least one involved World’s name, and which map it’s happening in).

I’m also investigating and may not post on this topic frequently.

Thanks Tarnished Coast for putting up an amazing fight toward the end of Round 1. We were only 1.1k behind Jade Quarry at reset, and managed to close in to less than 700 behind at one point. Those were some great fights.