Guild Wars 2

Chris Barrett.3280:

• 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.…..

not really. vote was every month not every 2 months with no indication some sort of averaging would take place… if that had held, people would be less likely to transfer every re-link and really bad links ala t4 would only last a month.

It would have made more sense to re-link frequently until you had the the re-link algorithms “perfected”. Then dial it back to 2 months re-links…

Just for the know, are ALL servers going to have their glicko manually adjusted, or just the servers having the issues?

Chris Barrett.3280:

• 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.

The public poll did suggest 1 month i believe but was changed on a “whim”?
Also, when people vote, the assumption is as follows:
If servers are properly and fairly linked, how often would you want to re-link?
one month
two months
three months
etc…
Unfortunately the linking was grossly imbalanced by sheer lack of effort in determining population and player activity by the people making the polls and using them as a scapegoat to not fix their mistakes.

while this is no single individuals fault, and the communication in the forums is appreciated (long overdue but still appreciated), be realistic and fix the mistakes without delay and excuses….and all is forgiven. Then, and only then, should you go back to playing with whatever numbers you chose to play with.

Unfortunately the linking was grossly imbalanced by sheer lack of effort in determining population and player activity by the people making the polls and using them as a scapegoat to not fix their mistakes.

Question:

What if, at the time of linking, and the last 2 months statistics said that the current match-up would work very well.

Then once the linking actually happened, all the fair weathers wake up on some servers, goes dormant on others, and a bunch of guilds and players that moved around to get on large server of preference, moves around to somewhere else again.

How are they supposed to ever get accurate numbers and statistics to create balanced links/servers/tiers ?

---

In this case, would there be any difference in re-linking ? Say they changed it this week, and gave Kain to SF group and ET to DH group.

Now say the new DH group "wins reset", and all their fair-weathers will be out in storm (DH got lots of fair-weathers), and suddenly all the fair-weathers in CD+BP will go dormant, and suddenly situation is reversed. (Toss in a few random transfers as well for good measure).

joneirikb.7506:

Unfortunately the linking was grossly imbalanced by sheer lack of effort in determining population and player activity by the people making the polls and using them as a scapegoat to not fix their mistakes.

Question:

What if, at the time of linking, and the last 2 months statistics said that the current match-up would work very well.

Then once the linking actually happened, all the fair weathers wake up on some servers, goes dormant on others, and a bunch of guilds and players that moved around to get on large server of preference, moves around to somewhere else again.

How are they supposed to ever get accurate numbers and statistics to create balanced links/servers/tiers ?

---

In this case, would there be any difference in re-linking ? Say they changed it this week, and gave Kain to SF group and ET to DH group.

Now say the new DH group “wins reset”, and all their fair-weathers will be out in storm (DH got lots of fair-weathers), and suddenly all the fair-weathers in CD+BP will go dormant, and suddenly situation is reversed. (Toss in a few random transfers as well for good measure).

Simple, they are using the wrong statistics.

Liston.9708:

Chris Barrett.3280:

• 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.…..

not really. vote was every month not every 2 months with no indication some sort of averaging would take place… if that had held, people would be less likely to transfer every re-link and really bad links ala t4 would only last a month.

It would have made more sense to re-link frequently until you had the the re-link algorithms “perfected”. Then dial it back to 2 months re-links…

You need to go read the poll thread again. The vote is to change linkings every 2 months.

Chris Barrett.3280:

• 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.

How is this different to now? In all but 1 tier.

If the linkages are done correctly, then this issue assumes less importance. If Eredon Terrace had large numbers after its link with BG then it shouldn’t have been in the 4 server link?

Ben K.6238:

We did have the 1-up 1-down system during one of the tournaments. The result was every single match in the mid-tier was broken. It is not a better system.

I wasn’t there, but this time with re-linkings every 2 months, it won’t be the same situation. In fact, as it currently stands, we’re seeing how broke glicko is in not reacting fast enough to the re-linkings.

joneirikb.7506:

Unfortunately the linking was grossly imbalanced by sheer lack of effort in determining population and player activity by the people making the polls and using them as a scapegoat to not fix their mistakes.

Question:

What if, at the time of linking, and the last 2 months statistics said that the current match-up would work very well.

Then once the linking actually happened, all the fair weathers wake up on some servers, goes dormant on others, and a bunch of guilds and players that moved around to get on large server of preference, moves around to somewhere else again.

How are they supposed to ever get accurate numbers and statistics to create balanced links/servers/tiers ?

---

In this case, would there be any difference in re-linking ? Say they changed it this week, and gave Kain to SF group and ET to DH group.

Now say the new DH group “wins reset”, and all their fair-weathers will be out in storm (DH got lots of fair-weathers), and suddenly all the fair-weathers in CD+BP will go dormant, and suddenly situation is reversed. (Toss in a few random transfers as well for good measure).

that’s part of the reason for a shorter link – will people really change servers every month? will servers continue to buy/rent guilds for 1 month? (they often send them to guest because it is cheaper)

Balthazzarr.1349:

Liston.9708:

Chris Barrett.3280:

• 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.…..

not really. vote was every month not every 2 months with no indication some sort of averaging would take place… if that had held, people would be less likely to transfer every re-link and really bad links ala t4 would only last a month.

It would have made more sense to re-link frequently until you had the the re-link algorithms “perfected”. Then dial it back to 2 months re-links…

You need to go read the poll thread again. The vote is to change linkings every 2 months.

No the vote was for every month, but not by a large enough margin to leave as is. Anet picked 2 months as some sort of averaging of the results of the poll.

Glicko preventing CD moving up to tier 3 seems like less of an issue than the metrics that were used to formulae the CD lead quad-server.

Isn’t the fundamental issue here that the quad-server even exists in the first place? What kind of metrics are available to you developers as you determine world-linking, and how are they so flawed that something as unbalanced as this 4v3v2 mess was deemed as a viable competitive environment?

Addressing Glicko is less important than addressing how we got the quad-server romp in tier 4 in the first place.

Julius Seizure.4985:

Glicko preventing CD moving up to tier 3 seems like less of an issue than the metrics that were used to formulae the CD lead quad-server.

Both are issues but I agree that the linkage calculations are being hidden to a certain extent by the glicko issue which is why for one I keep raising the question of how the linkage was arrived at.

At this state of the game, I would rather like to see random linking (excluding 2 T1 servers linked) and half the matchup time. Most of these matches are stagnant and boring; last two months and matches are determined before the weekend is over.

Making random linked servers every 3.5 days (Wednesday/Saturday) leads:
1. Interesting matchups on a shorter period.
2. Harder to stack “winning” servers.
3. Meet more of the WvW community members from different servers.

Granted this will make things harder to manage (ie: TS) along with a larger possibility of server trolling (however time frame will only be 3.5 days). At least matchups should generally feel different and more interesting compared to the 2 month cycle and the misleading Glicko ratings.

System hasn’t worked once they allowed transfers. There isn’t a system in place that can address wall clock coverage anyway. We saw this back in DaOC with 2 AM relic raids.

Unless you are going to cap the match to a smaller window and players on the map, this is going to be an issue for a long long time.

Reset the rating each time linking occurs. Keep the order but allow the servers to move up/down quickly.

This shouldn’t be a difficult problem to resolve.

Straegen.2938:

Reset the rating each time linking occurs. Keep the order but allow the servers to move up/down quickly.

This shouldn’t be a difficult problem to resolve.

We are into this now what, 2 or 3 years?

not even sure why i login anymore. oh yea to do dailies. maybe you’ll get the next linking right???

Great… With YB feeding glicko to the T3 servers the dream of getting rid of CD at 29th is now out of hand.

Jeknar.6184:

Great… With YB feeding glicko to the T3 servers the dream of getting rid of CD at 29th is now out of hand.

Yep, Anet just needs to go with 1U1D for some portion of the linking period. I was against 1U1D forever – but then I thought, why not just institute a 1U1D sometimes and Glicko rankings sometimes. That would seem to be the best solution.

However, I am liking this current T3 match. So I am hesitant to see it change. YB/SoS/SBI is a good match so far.

The real question is, how are the DH and SF linkings so much less populous than the rest.

Johje Holan.4607:

The real question is, how are the DH and SF linkings so much less populous than the rest.

Honestly, they weren’t… In the first week of the links we had 2-3maps queued on reset and we would had EBG queued constantly during pirmetime. But as the match goes on, people leave out of boredon. Fighting CD omniblobs at every corner gets old quick. Hell I bet even CD is bored at this point since there is barely anyone to fight anymore. Sometimes I just log in wvw for a few minutes to do my dailies and if there isn’t enough guildies to form a good party, I don’t even bother in keep playing it. I’ve made a lot of PvE progress thanks to that.

Jeknar.6184:

Honestly, they weren’t… In the first week of the links we had 2-3maps queued on reset and we would had EBG queued constantly during pirmetime. But as the match goes on, people leave out of boredon. Fighting CD omniblobs at every corner gets old quick. Hell I bet even CD is bored at this point since there is barely anyone to fight anymore. Sometimes I just log in wvw for a few minutes to do my dailies and if there isn’t enough guildies to form a good party, I don’t even bother in keep playing it. I’ve made a lot of PvE progress thanks to that.

I see it a bit differently, but, the result is the same…even the diehards have all but given up now. Few of us wvw regularly anymore. No longer is skill a factor, it’s just blob vs. blob…except DH doesn’t have much of a blob anymore.

It’s not enjoyable to play 7 servers vs 2. It’s no fun to be spawn camped. It’s ridiculous to have servers that greatly outnumber us using siege to defend camps in the bl. 15 vs 4 shouldn’t need siege. Nobody wants to just die over and over.

I don’t know what the answer is, but, the pairing has definitely not worked out for DH or CD. Perhaps it’s time to rethink the whole process.

How about ditching the outnumbered buff and having the BL’s queue for the team who is that much of a percentage higher instead? Then people would have to fight based on skill, not blobability.

briggah.7910:

not even sure why i login anymore. oh yea to do dailies. maybe you’ll get the next linking right???

2 servers vs. 7 servers. Seems fair (NOT!)

" It’s no fun to be spawn camped."- Especially when the opposing servers can spawn camp you on every BL at every exit. Remember, we’re talking SEVEN, 7, FREAKING SEVEN, servers against TWO (who have lost their largest WvW guilds and it hasn’t been reflected in the glicko or standings because it’s so messed up). How many other ways can I say 7v2??!?!?!?!!?!??!?!!!!!!!!!???!!!!!!!!!!!!!!!!!????!!!!!!!!!!!!!!!!!!????!?!?!?!?!?!?!?!?!!!!!?!!??!?!!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!!?!???!!!!!!!!!!!!!???!!!!!!!!!!!!!!!!!!?????!?!?!?!?!?!?!?!?!?!!!!!????!!!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!!!!!!!!?!?!??!!!!!?!!!!?!!!!!!!??!?!?!!?!

CD doesn’t belong in T3. After witnessing 15 on average DH/Ebay players wipe CD zergs twice or more than their numbers constantly you guys would get eaten alive by Mag.

You are fine where you are skill wise.

Mags is no longer in t3, so they’ll probably get eaten alive by acs instead.

How does Anet take fairweathers into account? PVE fairweathers and Gift of Battle farmers are swarming WvW, even to the point of covering off hours just to PvD. I’m pretty sure they’ll be gone when they actually get proper retaliation from other servers.

you people need to stop kittening about “time zone coverage”. thinking that your time zone is the important one is whats killing the game.

The CD/SF/DH matchup has been the same for weeks and weeks now and it’s not even remotely competitive.

I’m usually playing later in the PST evening hours and haven’t seen a queue on SF in the last week, even on reset night. Interest in this matchup is definitely tailing off.

Kyon.9735:

How does Anet take fairweathers into account? PVE fairweathers and Gift of Battle farmers are swarming WvW, even to the point of covering off hours just to PvD. I’m pretty sure they’ll be gone when they actually get proper retaliation from other servers.

bell shaped curve

Who would replace CD T4?
SBI?
YB? (seems to be the least populated)
SoS?

Haven’t played YB enough to say what kind of population they have, but, the other two would probably dominate that tier as well.

Obviously this is a question for those servers facing CD, not for the CD pop that want to get out of dodge.

Crapgame.6519:

We are into this now what, 2 or 3 years?

The score system was poor before linking. Post linking nobody has any idea how strong or weak a set of linked servers will perform. Using prior scores that have no relevance to a current match is stupid.

It’s interesting YB was T1 with no partner server and without one it has slid to T3. At the rate they’re shedding Glicko points they are likely to stay in T3 and lock Maguuma in T2.

Which of course does nothing to help resolve the stale matchup in T4.

Hesacon.8735:

It’s interesting YB was T1 with no partner server and without one it has slid to T3. At the rate they’re shedding Glicko points they are likely to stay in T3 and lock Maguuma in T2.

Which of course does nothing to help resolve the stale matchup in T4.

How true this is I’m not sure but, it’s rumored that YB either lost a bunch of people or they purposely lost to move tiers. I’m thinking they lost people because their population doesn’t seem T1ish.

Nuzt.7894:

Who would replace CD T4?
SBI?
YB? (seems to be the least populated)
SoS?

Haven’t played YB enough to say what kind of population they have, but, the other two would probably dominate that tier as well.

Obviously this is a question for those servers facing CD, not for the CD pop that want to get out of dodge.

Won’t know until it happens, if it will even happen in this rotation of server pairings. Obviously, it would have been more balanced if T4 was 3v3v3 instead of 4v3v2. We probably would have been okay with the same matchup week after week if it wasn’t so lopsided.

Nuzt.7894:

Hesacon.8735:

It’s interesting YB was T1 with no partner server and without one it has slid to T3. At the rate they’re shedding Glicko points they are likely to stay in T3 and lock Maguuma in T2.

Which of course does nothing to help resolve the stale matchup in T4.

How true this is I’m not sure but, it’s rumored that YB either lost a bunch of people or they purposely lost to move tiers. I’m thinking they lost people because their population doesn’t seem T1ish.

Both.

Links were very bad (or the timing of them) for YB as we were considered #1 for purposes of the first link and we clearly were not #1 based on more recent data/trends at that time. Specifically Hibergate returned to being Blackgate just prior to the big patch. TC got a bunch of guilds with some of them coming from YB.

Second link occurs and were still #3 in glicko for linking purposes, but again recent trends had us much lower than that. That is, glicko had not caught up yet. At this point, some/many were OK with dropping to T2 to be more competitive.

We roll T2 and 2 more guilds left. 1 wanted back in T1 (so cant really blame them) and the another went to Mag. Included in that was a popular pugmander.

Last part of the puzzle is another pugmander with a massive friends list has had connect issues for some time that are not easily resolved. 1 person should never be that important, but well it is obvious without him. He would message his entire friends list that was online when something was going down (or huge offensive was pending).

All that said, the last few weeks have been fun for the group I play with….

Too little too late by Anet, so many people I know have gotten bored and quit playing again including myself, it’s probably worst on DH/SF servers. Even with these proposed changes by Anet there is no guarantee we move up unless they do it manually. If CD were to get into T3 I see us stomping it with superior numbers as well. SBI and SoS actually have fewer players than DH and SF, it’s a complete joke this glicko garbage.

Straegen.2938:

Crapgame.6519:

We are into this now what, 2 or 3 years?

The score system was poor before linking. Post linking nobody has any idea how strong or weak a set of linked servers will perform. Using prior scores that have no relevance to a current match is stupid.

I wasn’t looking at it from a scoring system. I was looking at from 20,000 feet in which case the format of play. The fundamental problem is that we are seeing this has a “server” ranking and ego thing rather than bringing it down to the lower levels of Guilds, Players, and maybe future, Alliances.

Remove server ranks and place the responsibility back on the players. But this can’t happen until we get a web portal from Arena Net to include player and guild specific statistics and a true “active” player ranking system. By class, race, top 100, by week, month, and include guilds. Claiming, holding of camps, towers, keeps, and of course castles.

If Arena Net doesn’t want to invest the R & D dollar, that is fine. Then copy the data to a API data source off their production and open it to open source. We’ll create our own…

That is the only way this is going to be fixed. Not wallowing in a system that was built for something else and taking the lazy way out.

Can Anet confirm if the t4 matchup will change this reset? if not, i need to adjust my guild’s raid times for pokemon go.

shiri.4257:

Can Anet confirm if the t4 matchup will change this reset? if not, i need to adjust my guild’s raid times for pokemon go.

They are boosting CD’s glicko by 185 points, but with what their predicted glicko will be at the end of this week, they’d still be 70 points or so under the lowest T3 server. There is a possibility CD could end up in T3 next reset, but there is at least as much likelihood they won’t until they win a few more matches against the other two server pairs and boost their score a bit closer to T3. The 3 server pairs in T3 will be very close to each other in their predicted scores for the week (all 3 within 40-45 points), which makes it less likely for a server that is trailing by 70 points to move into that tier. If one of the T3 servers performed badly next cycle, that could help, too.

shiri.4257:

Can Anet confirm if the t4 matchup will change this reset? if not, i need to adjust my guild’s raid times for pokemon go.

Unfortunely, it’s just a maybe, not a certain.

Glad stuff like this is a priority for Anet “North Keep: One of the village residents will now flee if their home is destroyed.” Now I think I know why the glicko system is so bad and they couldn’t predict this problem in week 1, they have morons running WvW development.

Aitadis.8269:

Glad stuff like this is a priority for Anet “North Keep: One of the village residents will now flee if their home is destroyed.” Now I think I know why the glicko system is so bad and they couldn’t predict this problem in week 1, they have morons running WvW development.

To be fair that NPC live in WvW 24/7 so technically they are catering to the most hardcore WvW player in the game.

#villageresidentsmatter

A part of the glicko wall problem is the strength of victory calculation. I describe the problem here:

https://forum-en.gw2archive.eu/forum/game/wuv/Error-in-WvW-Glicko-Scoring/

The poor strength of victory modeling (for the actual score) results in a tier wall.

TorquedSoul.8097:

A part of the glicko wall problem is the strength of victory calculation. I describe the problem here:

https://forum-en.gw2archive.eu/forum/game/wuv/Error-in-WvW-Glicko-Scoring/

The poor strength of victory modeling (for the actual score) results in a tier wall.

Thats a really fantastic description of the problem with their glicko. The problem is obvious so why can’t they change? Do they not have anyone with the appropriate expertise working on WvW? If so get someone in temporarily to revise it. They’re talking about temporarily adding to a servers rating, well how do they calculate that?