Guild Wars 2

For any server that has been in or watching the T8 scores you might easily notice how the bottom 2 servers are several hundred points behind the other servers. This was caused by the several month long matchup with Kaineng/Devona/Ferguson in which Devona had to absolutely obliterate the competition in order to move up 1 tier. Thus, while Devona gained little points every week, Kain and Fergs lost points every week. After Devona finally moved up, Kain stepped their game up (via transfers, SoS’s return, PvE recruitment etc) and additionally obliterated the competition. Well that’s just dandy for Eredon and Ferguson who are still several hundred points below the T7 scores. Also, if anyone believes that those scores are fair (I doubt anyone is stupid enough to think that, but), look at the scores that Eredon and Sorrow’s have put up immediately after falling to T8. They lost 200 points at least in their first week at T8, and Eredon lost up to 500 eventually I believe. Currently Sorrow’s Furance is losing (barely) to Fergusons, yet they will still be 120+ points ahead of Fergusons because of how insanely scewed the scores have been. You may think that in time this will auto adjust but its blatantly obvious that this will never happen. While Fergusons and Eredon may gain points from having a close match with Sorrow’s, the majority of the points are just being lost by Sorrows. This will eventually reach equilibrium in a couple weeks with all 3 servers probably having around 750 points (that’s 200 points below Henge of Denravi, bottom of T7, who are doing far worse than both ET and FC did against Kaineng). So that basically leaves T8 totally stuck as it is until some server in T7 does extremely bad or a server in T8 gets a flood of transfer, which I know everyone hates with a passion, and does extremely well.

I honestly don’t know a solution for this, but it definitely is not right. Perhaps ANet needs to put in some manual boost in points for Tier 8 so that its actually possible for a T8 server to catch a T7 server without needing a ridiculous matchup.

Paid transfers need to begin

It’s just a bad rating system that uses statistics even worse.

it’s working as intended, nothing wrong.

Ranking system is messed up, but right now T8 is quite balanced and I am enjoying it. Finally getting some fun out of WvW after 7-8 weeks of completely imbalanced fights like what is happening in T7 right now.

Catisa.6507:

it’s working as intended, nothing wrong.

It may be working as intended, but if you knew anything at all about the Glicko-2 system you’d recognize that it was never designed to be applied to tiered competition. It was designed for open field competitions like chess and tennis, and it was statistically invalid for ANet to have chosen it for WvW. Go look it up on WikiPedia if you don’t believe me. You don’t even need to follow the math … just read the description.

Well the Tier is very balanced, yes, but for how long? 2 servers already have gotten tons of people who switched from another server and are boosting up the rankings which messes up the tiers with low population. I heard that Eredon has gotten some guilds already. I don’t want to have to keep dealing with being outmanned, it’s super annoying and no fun to come across the huge mindless zergs and get whiped. Anet really needs to fix it, I have seen a lot of new players come in for the WvW feature and leave due to getting smashed. The money transfers really need to start.

Darek.1836:

I honestly don’t know a solution for this, but it definitely is not right. Perhaps ANet needs to put in some manual boost in points for Tier 8 so that its actually possible for a T8 server to catch a T7 server without needing a ridiculous matchup.

The main problem with the rating system with population migration is that Anet is currently not tracking WvW population via an individual player value.

If Anet gave each player a WvW rating value, then the system would be able to compensate faster to the sudden influx of a sizable PvP population.

So for now it’s stuck in a blinders mode rating system that really needs to do match-ups via the active population.

Realistically this won’t happen until the SPvP statistics system is done, which for those that played GW1 GvG will remember as pretty nifty to see a nice parse of all the skills used etc…after GvG matches in the ladders.

Cactus.2710:

Catisa.6507:

it’s working as intended, nothing wrong.

It may be working as intended, but if you knew anything at all about the Glicko-2 system you’d recognize that it was never designed to be applied to tiered competition. It was designed for open field competitions like chess and tennis, and it was statistically invalid for ANet to have chosen it for WvW. Go look it up on WikiPedia if you don’t believe me. You don’t even need to follow the math … just read the description.

wrong. its designed for larger populations than 24. chess and tennis still have finite populations with a conservation of rating points law… they just have rather large populations. when you drive the population size down to a number you can easily comprehend, you run into weird cases like current tier 8 where matches become repetitive enough amongst all players so the lack of diversity translates into a positive feedback loop or a trap instead of a measure of skill. glicko isnt designed for 24. its designed for 1 thousand. or 1 million.

Lukian Darkblade.8012:

2 servers already have gotten tons of people who switched from another server and are boosting up the rankings which messes up the tiers with low population.

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Krakah.3582:

So for now it’s stuck in a blinders mode rating system that really needs to do match-ups via the active population.

Define active? Is it active during your prime time? mine? someone else? An aggregate value across the entire 24 hours block? or the whole weeks match up?
In your system does someone who sending more time logged in gw2 then become worth more of these mythical point? Or does a new system have to track WvW log in time? If the latter what happens when people AFK in WvW just because they want to be ‘cool’ and brag how many points they are worth?
What happens if say a guild of 50-60 active wvw players decides to stand down for a week because they are feeling burnt out? How do you compensate for that the following week when they return?

They can modify the rating system to take in account population. That should inflate the bottom 3 servers relative ratings while keeping them in t8 so once one starts consistently winning they move to t7.

The biggest problem with the rating system is that 2 servers can grow equally and become better equally but be locked into the same tier until one or the other gives up.

Once that server escapes t8, its shoots up the chart like a rocket till it hits enough resistance, then sheds the fair weather players it built up over a month and probably takes a short tumble (or long one).

They will move up and be out of your desolated wvw maps very soon.

Here’s what I think, paid transfers to begin and get rid of Tier 7 and 8 by merging the four worlds into 2. GoM and HoD or whatever. Only problem with world merging is people will take it as “game dying,” which is not the case. That’s my 2 cents.

Catisa.6507:

Krakah.3582:

So for now it’s stuck in a blinders mode rating system that really needs to do match-ups via the active population.

Define active? Is it active during your prime time? mine? someone else? An aggregate value across the entire 24 hours block? or the whole weeks match up?
In your system does someone who sending more time logged in gw2 then become worth more of these mythical point? Or does a new system have to track WvW log in time? If the latter what happens when people AFK in WvW just because they want to be ‘cool’ and brag how many points they are worth?
What happens if say a guild of 50-60 active wvw players decides to stand down for a week because they are feeling burnt out? How do you compensate for that the following week when they return?

I use the term active to describe those that are participating in the PvP activities of WvW. Which points back to the whole missing statistics system which would be needed to even begin to kitten player rating.

The Glicko system has a confidence value for the condition your speaking of when players take a break for whatever time and come back.

insanemaniac.2456:

Cactus.2710:

Catisa.6507:

it’s working as intended, nothing wrong.

It may be working as intended, but if you knew anything at all about the Glicko-2 system you’d recognize that it was never designed to be applied to tiered competition. It was designed for open field competitions like chess and tennis, and it was statistically invalid for ANet to have chosen it for WvW. Go look it up on WikiPedia if you don’t believe me. You don’t even need to follow the math … just read the description.

wrong. its designed for larger populations than 24. chess and tennis still have finite populations with a conservation of rating points law… they just have rather large populations. when you drive the population size down to a number you can easily comprehend, you run into weird cases like current tier 8 where matches become repetitive enough amongst all players so the lack of diversity translates into a positive feedback loop or a trap instead of a measure of skill. glicko isnt designed for 24. its designed for 1 thousand. or 1 million.

I’m not sure why you’re saying I’m wrong since I think we’re saying the same thing. The Glicko system was designed for competitions (and it was indeed specifically designed for chess) where everyone plays everyone else on a more or less regular basis … it falls apart when applied to tiered systems where competitors (servers in this case) might spend a significant number of their matches playing against the same two enemies. It loses contact with the other tiers even though it purports to rate teams in one tier against those in another tier. GW2’s implementation is also messed up in that it treats AvsB, BvsC, and AvsC within the same tier as if they were independent matches, which they most certainly are not. Lastly, the score for one server in one tier, for example, is influenced by their history in other tiers even though it may have been a long time since they were in that previous tier and radical population shifts may have occurred in the meantime. ANet can say all they want that this was the best implementation they could come up with, but that doesn’t mean it isn’t heavily flawed.

The problems doesn’t happen only on lower tiers… Blackgate was stuck at T2 for 9 weeks after clearly dominate the tier (against SoR and IoJ/TC).

All we can do is play for the lulz and not for the victory, because we’ll get too much disapointment…

If you think that’s bad, look at EU t9 xD

MikeFerguson.8921:

Catisa.6507:

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

While glicko may not be the greatest thing to use for tiered competition, I think it’s actually worked fairly well to help ensure that a large portion of our matches end up being competitive. We haven’t been seeing nearly as many lopsided blowouts happening as we continue to have more matches being played out and then feeding those results into the formulas. The obvious exception to that being the worlds who either lose or gain large amounts of players in a short period of time and get caught in a matchup where they get dominated or are the dominating force.

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

And yes, we need to get paid server transfers in. We all agree with that statement, and we’re trying to make it happen as soon as we can, while still making sure the guesting system is as useful and easy to use as possible. We’re agonizingly close, so hopefully soon we can stop talking about it and just do it!

Do you guys really need to hire a rocket scientist to tell you to suspend server transfers until you put in place a paid transfer system that every other MMO in excistance already uses, if they allow transfers at all. How many players do you want to lose by inaction?

I love how Kaining getting transfers automatically gives these guys a big head. DR got stomped once they met an equal population, so will Kaining. You were in tier 8 because you belonged in tier 8, and now that you have had a massive number of transfers, you’ll move up until you face equally populated servers.

There is a reason T1 servers went from T1 to T8, and it isn’t because they sucked.

This is funny.

MikeFerguson.8921:

… while still making sure the guesting system is as useful and easy to use as possible. We’re agonizingly close, so hopefully soon we can stop talking about it and just do it!

I’ve heard there was going to be a guesting system.
Thanks for the update.

MikeFerguson.8921:

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

The problem is that it takes weeks, and psychologically is quite punishing to the poor sods being farmed.

Krakah.3582:

MikeFerguson.8921:

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

The problem is that it takes weeks, and psychologically is quite punishing to the poor sods being farmed.

Indeed.
While I don’t really mind the concept of guilds transferring servers, when many guilds do it at once it’s very harmful.

It benefits one server,
And harms the server that lost guilds for weeks, as they freefall tiers rarely getting more than +100 during their primetime.
And in the instance of guilds transfering to lower tiered servers (Like kaineng), is devastating to every server they face until they climb to the tier they belong in.

Every time these guild transfers happen it causes harm to far more people than it benefits.

I don’t mind the Glicko system much,
But I think the Dev’s should pay attention and maybe artificially adjust some server rankings when these things happen. The integrity of the glicko system is far less important than Anet’s playerbase having a good time when they go into wvw.

I am highly opposed to payed transfers. There is so much cheating in the tier I am in right now that if I want out of it I don’t want to pay money to move. The game advertised free transfers when I bought it. That is what I expect. Otherwise I want a refund.

Saweth You Him.9047:

I am highly opposed to payed transfers. There is so much cheating in the tier I am in right now that if I want out of it I don’t want to pay money to move. The game advertised free transfers when I bought it. That is what I expect. Otherwise I want a refund.

Might want to get a refund then,
They’ve been saying forever that free transfers were only temporary.

every account should get 3 free transfers (gives new players a chance to find a home server that suits them, as a new player you would be blindly choosing whatever was available without much hope of knowing what you are heading into let alone what you want to be looking for.) after the 3 are used up make it cost money, and not a token amount either.

Thanks for the responce MikeFerguson, good to hear you guys are having a look at this and that we hopefully soon get rid of free transfers.

About the scoring system. I agree with you that it does its job OK, but it ofcourse can’t compensate for large players moves between servers. Could it be an idea, to keep a bit track of this manually and than when a server one server is really outclassing the others in its rank, that you could speed the ranking up process? A.t.m. we have this problem in the EU regions with RoS, Vabbi and FoW where fow is totally outclassed by the 2 other servers that have an large influx of decicated WvW players. this moving left aside, if it lasts for a week, tops 2 weeks FoW can get through it as a server. However a.t.m. the score of vabbi is really far behind the scores of the next Trier, so that it would take them more than 2 weeks to move up. Being unable to do basically anything in WvW for that long will make lots of players lose interest in WvW or move to other servers, breaking ours just recovered WvW groups once again.

I’m sure the US server have similar problems as was indicated here. so perhaps some manual adjustment of the scores could help to move these overpowered wvw servers back up to where they belong faster. one week of total domination in WvW a server could outlast without to much problems, but longer than that is going to affect the dominated server in a very bad way, giving rise to more moving and more unbalance in the scoring system.

A concerned FoW player who likes the casual WvW that we had going there

and everyone in the same boat of other servers i feel with you, now we know what it feels like…..

MikeFerguson.8921:

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

Perhaps its better to just do direct match making by population? I agree that the biggest problem with the system right now is the players. You just can’t build any system that relies on lots of people to do the “right” thing – join servers with a mind to creating balanced matchups.

I think part of the problem is that the server ranking are effectively a leaderboard. Players feel a strong compulsion to go up the rankings and get to number 1, even though position on the leaderboard is correlated to population/timezone coverage more than anything else.

We are seeing this in the amount of low tier matches that are ruined by guilds or alliances of guilds mass transferring to low tier servers to “bring them to the top”.

They eventually get a rude awakening when population/timezone coverage advantage flips and there is a mass exodus as none of the players have experienced fighting heavily outmanned.

I think free server transfers are a good thing. I like going (by myself) to another server for a “holiday” so I can play with new people for a week, get new ideas and see different methods of playing WvW and share knowledge. People who play PvE rightly server transfer on a temporary basis to get around bugged events and so on. The problem is when 50 people in one guild all transfer together because that stacks the WvW deck so badly.

Random thought…. Maybe Anet could close transfers, but allow guild leaders to apply for a free transfer for all their guild members to a specific server. Of course, the guild master would have to give reasons why they’re transferring, but at least then guilds could leave the server for free for legit reasons (such as wanting to participate in higher tier competition, or fill a gap in another server’s coverage).

They’d also be able to limit how many guilds get to go to various servers for free, reducing the chances of creating new bandwagon servers, while allowing natural growth of others.

This would counter one of the main problems with closing free transfers, which is stopping already viable servers from increasing their ranking due to lack of numbers / coverage.

MikeFerguson.8921:

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

You guys could pay attention to your community and see when a zerg of 500 people switch to a tier 8 server that begins winning by 200,000 points. It’s not secret or mystery when this happens. When something like that happens you could manually move them up a couple of tiers. Or your CS team could decide that mass xfers to lower pop servers is unintended game play and ban those guilds for abusing the Tier. Who knows.

Seems like alot of Tiers have problems with either taking to long to climb up in tiers or taking to long to drop to an appropriate level. Why not simply have one team drop, one team go up and one team stay in the same tier each week.

To give an example right now. I am from CD/FA/IoJ server. CD has significant numbers this week and are dominating. So CD should move up as they will win the tier. Whoever scores last moves down. Whoever finished middle stays in the tier.

The system isn’t accomodating for the fluctuation of numbers as it can take several weeks just to go up/down one tier.

I would actually recommend retaining a very similar system to calculate ratings that are used now, but only make it account for the last 2-3 matches. Instead of the server’s entire history.

Hm in real world new foreigners aren’t put in the army as well and should they come originally from an opponent they sometimes even get be arrested.
Why don’t we have something like that in WvW:
- You are NOT allowed to enter WvW in a world where you are not resident since 14 days
- And your waiting-queue priority depends on the time of your residency on this server, i.e. being a loyal citizen on a world gives you a bonus over server-hoppers that clog your queue.

Dayra.7405:

Hm in real world new foreigners aren’t put in the army as well and should they come originally from an opponent they sometimes even get be arrested.
Why don’t we have something like that in WvW:
- You are NOT allowed to enter WvW in a world where you are not resident since 14 days
- And your waiting-queue priority depends on the time of your residency on this server, i.e. being a loyal citizen on a world gives you a bonus over server-hoppers that clog your queue.

Rule 1 will also penalize new players that just bought the game. What you probably want is for rule 1 to be enforced on transferees.

Rule 2 will also penalize new players who may never get the chance to play WvW if they keep getting pushed back in the queue for older players.

Why is there even a need for a point ranking system. 1v1v1 winner of the week moves up.Loser moves down. 2nd place stays in same tier.

Nothing complicated and keeps things fresh EVERY WEEK.

MikeFerguson.8921:

Catisa.6507:

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

I think this is a cop-out personally. It is your system and you allowed extensive free transfers far longer then was necessary. Sure it is easy to say no system can handle it, but you should have had a plan for when it did go bad.

Leaving servers like HoD and ET to an extremely slow decline to the bottom was not fair to them, and not fair to every match-up that had one of them in it. There is only so many tiers and servers, you have the stats, you should have manually intervened when there was an obvious need.

MikeFerguson.8921:

Catisa.6507:

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

While glicko may not be the greatest thing to use for tiered competition, I think it’s actually worked fairly well to help ensure that a large portion of our matches end up being competitive. We haven’t been seeing nearly as many lopsided blowouts happening as we continue to have more matches being played out and then feeding those results into the formulas. The obvious exception to that being the worlds who either lose or gain large amounts of players in a short period of time and get caught in a matchup where they get dominated or are the dominating force.

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

And yes, we need to get paid server transfers in. We all agree with that statement, and we’re trying to make it happen as soon as we can, while still making sure the guesting system is as useful and easy to use as possible. We’re agonizingly close, so hopefully soon we can stop talking about it and just do it!

When you say payed transfers, does that mean it will take irl money to transfer? OR in game money to transfer?

atomicjellybean.5389:

MikeFerguson.8921:

Catisa.6507:

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

While glicko may not be the greatest thing to use for tiered competition, I think it’s actually worked fairly well to help ensure that a large portion of our matches end up being competitive. We haven’t been seeing nearly as many lopsided blowouts happening as we continue to have more matches being played out and then feeding those results into the formulas. The obvious exception to that being the worlds who either lose or gain large amounts of players in a short period of time and get caught in a matchup where they get dominated or are the dominating force.

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

And yes, we need to get paid server transfers in. We all agree with that statement, and we’re trying to make it happen as soon as we can, while still making sure the guesting system is as useful and easy to use as possible. We’re agonizingly close, so hopefully soon we can stop talking about it and just do it!

When you say payed transfers, does that mean it will take irl money to transfer? OR in game money to transfer?

It’ll be a gem purchase, almost definitely.

So, both.

for some reason I think a curvilinear approach to scoring in the glicko system may be better. At times it feels like the system takes far too long to make needed changes that may seem small, like moving a top server in that tier up or another that is losing constantly by a marginal amount down.

endeavor.7530:

Here’s what I think, paid transfers to begin and get rid of Tier 7 and 8 by merging the four worlds into 2. GoM and HoD or whatever. Only problem with world merging is people will take it as “game dying,” which is not the case. That’s my 2 cents.

Are you even in t7 to say that?
T7 is currently the best match-up I’ve ever seen and played, we finally found the tier we belong right now and I’m loving the fight against HoD and GoM, the best fights, the best skirmishes, everything is just great and I actually am having too much fun, something that I couldn’t really say much in the past 4+ weeks.

I don’t really know about t8 status, so can’t say much on that, but t7 is “perfect”.

Kiexcolo.2416:

Dayra.7405:

Hm in real world new foreigners aren’t put in the army as well and should they come originally from an opponent they sometimes even get be arrested.
Why don’t we have something like that in WvW:
- You are NOT allowed to enter WvW in a world where you are not resident since 14 days
- And your waiting-queue priority depends on the time of your residency on this server, i.e. being a loyal citizen on a world gives you a bonus over server-hoppers that clog your queue.

Rule 1 will also penalize new players that just bought the game. What you probably want is for rule 1 to be enforced on transferees.

Rule 2 will also penalize new players who may never get the chance to play WvW if they keep getting pushed back in the queue for older players.

Beside the fact that a new player probably needs at least 14 days to be in a shape (know the game, have some lvl, have some equip) to enjoy WvW:
Do it as in real world: new (born) players that are choosing their first world get full residentship, i.e. they are considered to be resident since launch date, or set a cap: everyone that is at least two month on a world is handled equally.

There really are two separate problems: 1) if the matchup doesn’t change much within a tier, the ratings system doesn’t reflect when all three servers are getting better/worse together, and 2) when a tier becomes unbalanced, it can take too long to get rid of the offender. Both are exaggerated in tier 8.

To address the first, one relatively easy idea would be to occasionally swap the 3rd server in a tier with the 1st in the one below. This would help make sure the tiers are in the right order and that the numbers reflect the differences a little more accurately. Had Devona’s Rest gotten an earlier crack at the Tier 7 servers it would have saved us a month of lopsided matches. And even if they had gotten sent right back down, their rating would reflect how good a match they were for Tier 7, thus improving the accuracy of the tier 8 ratings. The other tiers already kind of work this way because the numbers are more bunched up — it happens organically. Another possibility would be to put a floor on any server’s rating. This would stop tier 8 ratings from spiraling down when we trade wins.

For the second problem, one idea would be to exaggerate the contribution of score differences to the ratings system as they increase. I think it’d be better to send an upstart server into a tier too high than too low.

All that said, the current matchup in tier 8 is pretty much perfect. But if one server suddenly gets stronger it’s going to take a long time to get out.

MikeFerguson.8921:

Catisa.6507:

There is no system in the world that’s ever going to compensate for server transfers, or guilds taking a break from wvw, or differences in primetime play hours.

Pretty much this. We use gliko to help make sure the matches end up being somewhat competitive, but no systems can really predict and react immediately to constantly changing populations.

While glicko may not be the greatest thing to use for tiered competition, I think it’s actually worked fairly well to help ensure that a large portion of our matches end up being competitive. We haven’t been seeing nearly as many lopsided blowouts happening as we continue to have more matches being played out and then feeding those results into the formulas. The obvious exception to that being the worlds who either lose or gain large amounts of players in a short period of time and get caught in a matchup where they get dominated or are the dominating force.

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

And yes, we need to get paid server transfers in. We all agree with that statement, and we’re trying to make it happen as soon as we can, while still making sure the guesting system is as useful and easy to use as possible. We’re agonizingly close, so hopefully soon we can stop talking about it and just do it!

I am from Devona and I approve everything original poster says. Game should encourage players to transfer to low populated servers and forbid to transfer high populated servers. Outmanned buff should have priority in Dev teams schedule to reward people playing outmanned and it has to have different levels. Harder you are outmanned better the buff gets. Legendary weapon requirements has to get revised to stop transfers for WvW map completion and badge farming. Jumping puzzle should be removed from WvW and work on its own. There should be a system like guild influence which will work as server influence encouraging players to stay and build it in one server. Guilds should be server only which means if you leave your server you can not be in that servers guild any more. Server transfers should be available every 3 months for each account for free and in all other times with a small fee and huge handicap in WvW after transfer for a certain amount of time. Glicko2 system will work as intended if all these changes are made I guarantee.

If players are changing servers every week how do you expect glicko2 system to define a servers rating correctly in a weekly match up system. That is the main question should be asked.

Random match ups as their should be no tier system to the casual side of PvP and their should be an option for guilds to seek a more organized and serious WvW experience in the future. Something along the lines of scheduling a WvW match up with your own options like player limit, team setup, time frame, etc nothing that would cause in balance(once the game starts no more editing, booting, or changing to the teams or set up you have created as well) but allow enough customization to make your own WvW match ups. Once you get all your settings made you put your server up their (private or public) and guilds will request or can be invited to the server either requesting a certain amount of server spots or being given a set amount by the server master. Then once the big day comes all the guilds will probably be ready. you can just add 2 new tabs to the WvW menu one for regular guildies for joining the WvW’s their gm have set up or joined(also keeping track of scores in those wvws) and a special tab for guild masters keeping track of invites and what servers they are doing WvW in.

Overall though Random Tiers would make Server transfer worthless as the WvW experience for each server would be changing each week. and matchmaking would make it so the organized guilds would have their experience without it being ruined by randoms they can’t coordinate with.

Bumping my own thread because we are now experiencing more fun in tied 8
Right now Sorrows Furnace is winning T8, and although it isn’t really a blow out it definitely isn’t as balanced as it could be. After next week, it is very apparent that Sorrow will win every week by around the same amount since there will be no transfers. Further, since they are not blowing out T8 like Devona and Kaineng did, they will NEVER (yes, literally never) move up to T7. Well that’s a problem, huh? Like I mentioned, it isn’t a blowout, however, despite Sorrow is very worthy of a T7 rank, they cannot and will not move up. That leaves T8 as basically the same matchup every week for AT LEAST the next 3 months. That’s right, Ferguson and Eredon probably won’t win a single match until summer because the Glicko system is so off balanced for T8 that it can’t match up the right teams.
Working as intended, right, ArenaNet?

MikeFerguson.8921

However, even with all the issues glicko presents, we have seen that while it takes a few weeks for the system to adjust to the population shifts, it DOES adjust, and eventually those worlds end up playing in competitive games again once they reach a tier that suits their current population.

Bumping an old thread once again. This quote given by Mr. Ferguson is no longer true and I would prefer if we could get another official response for this issue. Obviously since the time I posted this thread there’s been around 15 new threads JUST over this issue. Just because this issue only affects maybe 10% of WvW players does NOT mean it should be completely ignored.

A server who wins by as wide of a margin as SF has been winning for as long as they have been winning in tier 8 should have some avenue where they ‘get their shot’ at the next higher tier.

Not only is this system really plowing FC and ET into the ground, it is holding winning servers back and not giving them any real chance at moving up.

SF has shown their dominance in tier 8. Give them a shot at T7.

Glicko was not designed for a tiered system like this, and the reason is becoming obvious.
Now if we had random mathups where a T8 server may play a T4 server randomly, glicko would shine since the T8 server should indeed gain rating if they score any in game points at all, but we do not do that. We play the server based on tiers made from an ordered list from (ideally) the most powerful server to the least powerful server. Well if we are going to place the matchups in tiers based on an organized list sorted by power of that server, then why are we not simply using the classic winner moves up, loser moves down?
Somebody got it in their head that glicko would be cool, and we are suffering because of the over engineered WvW scoring system.

The answer has always been stupid-obvious: Keep tier colorization and glicko2, but calculate each server against nearest-neighbor, except for top and bottom server (which are each calculated in the tiered way, because they’re missing a neighbor.)

Ranks would be more volatile, but that’s not a bad thing in a game that never ends.

What breaks glicko in the current use case is the fact that historical rank actually prevents the servers at the edge of the tiers from fighting each other to determine which of those two deserves to gain or lose rank based upon their relative strengths.

Edit: This is because we’re calculating in threes instead of pairs of opponents. The mere fact of being pitted against a very weak opponent has the effect of stifling a powerful server’s rank growth. When this is done in matches of three, the effect is amplified, and because the next match is more likely than not to be the same opponent again, the stifling becomes a trend.

You can’t accomplish their decisive battle without actually letting them fight. To cause glicko to allow them to fight in cases like Henge of Denravi and Sorrow’s Furnace right now, you have to calculate glicko against nearest neighbors, and not restrict it to your fancy tiers.

PS: The current ratings data is poisoned by the tier system, and the above would not fix it as-is.