Guild Wars 2

well over all i like the how you dont know who your going to fight next, but there are tiers (T3,T5,T7) that are uneven. although it has been little over 12 hours since reset, we shall see how the rest of the week goes.

#2, the current scheme with RNG element with two tweaks:

1) Tighten up the variance of the RNG element so it’ll be more of a 1-tier shift and only rarely a 2-tier shift but no more than that.

2) Add another variable that makes it so servers that have played each other the past week(s) already will be less likely matched up against each other the following week.

2
I know we are gonna get creamed since we are playing a tier 1 server but i enjoy the fights. Plus its no fun fighting the same servers and wiping them for months.

I think it’s hilarious that so many people seem to think the ranking system is at fault here, no matter what it is. The Glicko-2 system is totally misapplied to GW2 (especially with randomization to subvert it), but the fundamental problem is that WvW will always be flawed as long as there is so little control over the number of players matched up against each other. It beats the kitten out of me that ANet can’t see that either.

I would guess #2 will be the most popular. It’s not like hand holding zerglets, which make up the vast majority of the player base, are overly impacted by a blow out match. They still get to zerg around in a giant ball regardless if they go up against anyone or not.

Its the folks interested in challenging fights that get the shaft in the new system.

Here’s my idea which is simple and should be fairer than the new system.
Taking the list in ratings order. Take rank 1 and randomly match them against the next 3 available ranks. So 1 will always matchup with 2, 3 or 4. For matchup 2, take the highest available rank and match them randomly with the next three available ranks.
Rinse and repeat.
For example:
Match 1: 1 v (2, 3 or 4) = 1 v 3 v 4
Match 2: 2 v (5, 6 or 7) = 2 v 5 v 7
Match 3: 6 v (8, 9 or 10) = 6 v 9 v 10
Match 4: 8 v (11, 12 or 13) = 8 v 11 v 12
Match 5: 13 v (14, 15 or 16) = 13 v 14 v 15
Match 6: 16 v (17, 18 or 19) = 16 v 17 v 19
And so on….

The biggest difference in opponents in this system would be 5 ranks, so 8 v 10 v 20 can’t happen (yes I am on GH!). Match 1 will always be strong with only the top server of tier 2 involved. All matchups are likely to be varied with only the bottom tier having two set teams, which keeps it competitive for them.

Well that’s my idea, no system is going to keep everyone happy but this would ensure matches are more competitive in my opinion.

4. i dont know what it would call
i want Winner move up, last place move down, 2nd place stay in the same tier…hope it make sense x.x

example is in the picture

Current random system (2). I’m in T5 and am extremely disappointed at this weeks match, but it was a 6.4% chance based on the math snowreap did for our server. If we got unlucky, we got unlucky and most people in this tier are mad because we had actually done enough to get a new match without the system change. If it happens again next week then I might cry foul, until then I believe that giving everyone the prospect of new fights is an excellent idea. We can’t get accurate ratings without match variation.

I’m dead set opposed to WULD systems though. I hope they never implement that unless they can magically balance populations.

3 most likely, even if I am no expert on how this system actually works lol.

3)
https://forum-en.gw2archive.eu/forum/pvp/wuvwuv/Simplify-the-Matchup-System

2

3 is oversimplified, IMO. And yes, I’ll take a chance at a blowout against a much higher ranked server in order to see the wide variety of opponents available with option 2.

What’s wrong with 1 (besides the stagnation), is that it actually creates more inaccurate rankings than 2, not less. In any ranked sports system, such as tennis, the ranking a player maintains has to do with his performance against a variety of opponents, including lower and higher ranked players. There is always a random element to it, depending on who’s playing, who won wildcard placement, who has an injury, etc. Players are constantly facing opponents who are much higher or lower in the rankings than they are.

In order to work out what a servers true ranking is, it can’t just play the same 2 other servers all the time. They have to have a chance to demonstrate all their skills and weaknesses, not just the ones that tend to show versus the two other servers that they play the most. The new system, by adding back in the element of randomness, will allow a broader picture of any given server’s abilities to be reflected in their rankings.

What’s missing is a purpose for the rankings, besides kitten. Granted, that’s enough to drive a lot of people, but having a true tournament or sports season of some kind, with reset rankings at the end and a chance to start over in a new season, and accolades of some sort for the winning servers, and—ideally (although I know this would require some serious work by the developers), a chance for the winning servers to face off in a game-wide championship, NA vs EU, with a wildcard placement for the 3rd server.

2
Because diversity in the matchups is so much better than being stuck in same tier forever.

Somehow I doubt overwhelming support for the current (new system) was what the op expecting to see lol.

Have fun with option 2. I’ll check with you in a week and once this gets implemented you will be stuck with it for a month. Watch ppl quit then and zergy servers have nothing to fight.

Cactus.2710:

I think it’s hilarious that so many people seem to think the ranking system is at fault here, no matter what it is. The Glicko-2 system is totally misapplied to GW2 (especially with randomization to subvert it), but the fundamental problem is that WvW will always be flawed as long as there is so little control over the number of players matched up against each other. It beats the kitten out of me that ANet can’t see that either.

This is the real winner of this thread.

The scoring in the system doesn’t reflect how well your server does beyond WvW population unless the servers are evenly matched and very few servers are evenly matched, some are close, many are unbalanced.

None of the options are good..

and 2 is the worst..

Its relying on random glicko like this is Baseball…Only one team has 30 people in the outfield and the other team has 1…

CreativeAnarchy.6324:

Cactus.2710:

I think it’s hilarious that so many people seem to think the ranking system is at fault here, no matter what it is. The Glicko-2 system is totally misapplied to GW2 (especially with randomization to subvert it), but the fundamental problem is that WvW will always be flawed as long as there is so little control over the number of players matched up against each other. It beats the kitten out of me that ANet can’t see that either.

This is the real winner of this thread.

The scoring in the system doesn’t reflect how well your server does beyond WvW population unless the servers are evenly matched and very few servers are evenly matched, some are close, many are unbalanced.

Exactly. This is not how the glicko system should be applied. It’s like painting your house with a sharpie.

What people are currently failing to realize is that each week high ranked servers will bleed ratings points to lower ranked servers. This won’t lead to “stability” in the ratings, it will lead to massive, massive blowouts as the possibility of very high ranked servers facing very low ranked servers will increase each week.

I think they should, first of all, decrease the player cap for each WvW map to reduce skill lag. Then they should base server ratings on participation instead of 24/7 PPT. If JQ has 600 people playing every night they shouldn’t be matched against a server that has 200 people playing. Just time-weight the WvW participation (total hours played by population during the match/total hours of the match) and rank the servers based on the results. I honestly can’t believe no one’s been fired over this glicko fiasco.

Maladon.5760:

CreativeAnarchy.6324:

Cactus.2710:

I think it’s hilarious that so many people seem to think the ranking system is at fault here, no matter what it is. The Glicko-2 system is totally misapplied to GW2 (especially with randomization to subvert it), but the fundamental problem is that WvW will always be flawed as long as there is so little control over the number of players matched up against each other. It beats the kitten out of me that ANet can’t see that either.

This is the real winner of this thread.

The scoring in the system doesn’t reflect how well your server does beyond WvW population unless the servers are evenly matched and very few servers are evenly matched, some are close, many are unbalanced.

Exactly. This is not how the glicko system should be applied. It’s like painting your house with a sharpie.

What people are currently failing to realize is that each week high ranked servers will bleed ratings points to lower ranked servers. This won’t lead to “stability” in the ratings, it will lead to massive, massive blowouts as the possibility of very high ranked servers facing very low ranked servers will increase each week.

I think they should, first of all, decrease the player cap for each WvW map to reduce skill lag. Then they should base server ratings on participation instead of 24/7 PPT. If JQ has 600 people playing every night they shouldn’t be matched against a server that has 200 people playing. Just time-weight the WvW participation (total hours played by population during the match/total hours of the match) and rank the servers based on the results. I honestly can’t believe no one’s been fired over this glicko fiasco.

We had that system already, hence why the ranking system was nothing more then a population chart.

I’ve said it before..The Ranking system is not the fundamental problem with the WvW system.. and until they realize that…nothing will change..They can change it all they want but unless they fix the fact that Zerging is flat out the best method of WvW with zero down side, The Rank system no matter what they choose, will always be based on Steam Rolling the less populated side.

Xsorus.2507:

Maladon.5760:

CreativeAnarchy.6324:

Cactus.2710:

I think it’s hilarious that so many people seem to think the ranking system is at fault here, no matter what it is. The Glicko-2 system is totally misapplied to GW2 (especially with randomization to subvert it), but the fundamental problem is that WvW will always be flawed as long as there is so little control over the number of players matched up against each other. It beats the kitten out of me that ANet can’t see that either.

This is the real winner of this thread.

The scoring in the system doesn’t reflect how well your server does beyond WvW population unless the servers are evenly matched and very few servers are evenly matched, some are close, many are unbalanced.

Exactly. This is not how the glicko system should be applied. It’s like painting your house with a sharpie.

What people are currently failing to realize is that each week high ranked servers will bleed ratings points to lower ranked servers. This won’t lead to “stability” in the ratings, it will lead to massive, massive blowouts as the possibility of very high ranked servers facing very low ranked servers will increase each week.

I think they should, first of all, decrease the player cap for each WvW map to reduce skill lag. Then they should base server ratings on participation instead of 24/7 PPT. If JQ has 600 people playing every night they shouldn’t be matched against a server that has 200 people playing. Just time-weight the WvW participation (total hours played by population during the match/total hours of the match) and rank the servers based on the results. I honestly can’t believe no one’s been fired over this glicko fiasco.

We had that system already, hence why the ranking system was nothing more then a population chart.

I’ve said it before..The Ranking system is not the fundamental problem with the WvW system.. and until they realize that…nothing will change..They can change it all they want but unless they fix the fact that Zerging is flat out the best method of WvW with zero down side, The Rank system no matter what they choose, will always be based on Steam Rolling the less populated side.

No the system we had before was based on PPT to feed a rating. Time-weighted participation would mean servers who log approximately the same man-hours per week would be matched up. If one server has 100 players a day who play for 2 hours each day they’d be a relatively close match for a server that has 20 players who play 4 hours a day and another 40 who play 3 hours a day. It won’t necessarily matter which hours they play, just that the total hours played by each server is close.

Wondering if people voting 2 are serious or just trolling… still trying to decide.

2 and im from a tier 7 server fighting a tier 3&4 server and having a blast!

I’ll also be voting for 2, it keeps things interesting, and although the rankings will go a little everywhere for a while they will still give an accurately weighted overall ranking over time, and still keep matches interesting without them going stale.

A promotion demotion system wouldn’t be a bad thing either, but wouldn’t solve the issues relating to the top tiers and bottom tiers having somewhat stale matches, and could end up with the same groups of servers competing, just not limited to the 3 in the matchup.

dragon rank 5, storm rank 13, maguuma rank 7

Stormbluff Isle has 50k points less than Dragonbrand.
Stormbluff is gaining+26 rating.

Maguuma, a measley 5k more points than Stormbluff, is losing 66 rating.

Dragonbrand gains 45 rating.

Tell me how this was intended to be fair.

I suppose it doesn’t matter what rank you’re at with the high volatility of this system, though. Glicko whatever; this volatility needs to go, or it needs to be a hell of a lot smarter.

People voting 2… you playing gw 2 right? Have you checked if you are in the wrong forum or something? Or… you are just that casual player that logs in WvW 2 times a week?

i would prefer an all out random fight

Umm…. this might sound nutty but, how about a polling system? Where we actually vote on it every week? Where in the WvW overlay we get to pick (once per account) on who we want to fight the next reset.

Think about it, do we really need to be forced by a system? What if for one week we wanted a grudge match between another server from on another tier? It’s not likely to happen by any ranking system. Unless you purposefully came in 3rd week after week to drop down.

Yes it’s a blowout this week, so what next week might be better stagnation has you playing the same match everyweek so after a while you just give up this new system makes WvW exciting again.

2, but with less deviation IMO is the best solution atm. The only thing they have to prevent (for NA anyway) is servers from moving more than 1 tier up or down.

I really don’t see how you can avoid blow-outs with the current scoring system (especially if you have 4 servers with similar population, and the 2 servers below you have significantly less population)

It would be fantastic though, if server movements at least correlate with their prior week performance (e.g. if SBI was last in their previous matchup, it’s pretty silly to have them matched up against servers with a rating 200-300 higher)

However, more variety is good.
Big change for us SoS Oceanics from PvD to facing big zergs haha

Crossposting this from here https://forum-en.gw2archive.eu/forum/pvp/wuvwuv/Two-hours-after-reset-and-WvW-fulled/ but it’s still relevant to here imo…

4

Split the servers into three (probably fuzzy, as 3 doesn’t divide equally for NA servers) pools of servers, top, mid, low, and randomly dividing the servers up from there. Fuzziness might pull someone from top to mid, or mid to low and vice versa to overcome the problem with numbers. Randomising within a pool means servers with too high a variation for randomisation to cause effect (eg. EU Vab/FoW) will still get split more often than not, but servers will still be relatively similar strength in most cases.

The only reason people are saying 2 is because its only been out a short while. The old system was far better for making balanced matchups. If people want more different opponents, they should have simply implemented a “winner moves up, loser moves down, 2nd place stays put” system. This would allow for more variety without making total joke matchups and actually reward people for winning matchups. Its absurd that you can win your matchup and fall 1 or more tiers as your reward. The new system boneheaded.

I like the concept of the new randomization system (2), but it seems like the variance is too high. Maybe this will mellow out in a week or two.

I do think that there needs to be some system in place to keep servers from being “Tier locked”. There should be a way to do this that is more deterministic than the current system. E.g. Some modification of winner moves up, loser moves down, and the only ranks that matter are the “Tiers” (with colors assigned randomly at the start).

2 people dont seem to understand that things will balance better over time…yes there are going to be blowouts from time to time but one week you may lose next you may win, stacked servers will soon get bored and the population will spread out. Give it time and just keep fighting

I vote for #1

I hate RNG do #2 is out. I do like being able to just roflstomp lower servers but that only that’s for a couple hours before that too gets stale. Its like have GOD mode in Doom 2 turned on.

Option #3 much to everyone talking about it worse than #2. While 1 up 1 down seems legit when you think about it its actually like #1 except your stale matches happen every other week.

Option #4 is the best option but I haven’t solved the answer to that. There are some great suggestions I’ve read that would make sense but not any one post has nailed it yet. The key problem is coverage and being unable to defend against great numbers. The arrow cart miscalculation did help to fix this a little bit but then you got the flip side. All the zerg complaining that 5 people could hold off 30 quite easily if they had carts. I noticed new strategies coming out to counter the arrow carts but it still wasn’t enough.

ATM I think the world skills help a lot in making the fights more fair. I can’t wait to see them fleshed out and fine tuned more. I would also like to see a respec option for those skills. I also think the outmanned buff needs something a lil extra to help bridge the gap. The no durability lose is huge as it allowed people to just fling themselves at the enemy over and over to stall for time and whittle away their health.

Perhaps a stacking buff (with limits) that works similar to outmanned but increases the 3rd place’s players stats to help them bridge the gap, or a stacking debuff on the 1st place team, or both. This would have to be done carefully though as is should be designed to make matches closer but not cripple any one team.

All in all.. bring back option 1 ANet and look for a new option because this current system isn’t going to be fair until population and coverage is nearly 100% equal.

2. While I take issue with some WvW balance, the same problems were present and unsolved already, and I’d prefer to address them in ways besides keeping a dull matchmaking system.

2 is okay if they control the rating gulfs that can potentially be jumped by a roll. I don’t want to fight servers 150-200 rating higher, since rating is effectively a measurement of population and coverage.

Basically, lower your volatility by a TON, and it might work.

Why do Anet love so much Random thing. We have RNG and now we have a tipe of RNG in WvW

2 but decrease the standard deviation of the randomization rolls. And this is coming from someone on a server that had a bad luck(SBI). It felt so good when we took back our entire borderlands from Dragonband. I knew we were going to loose but that alone made it a success in my eyes. Also I think we are doing a lot better then a lot of people think we would.

But then it is also too early to say if they need to change things a bit more.

Has nobody pointed out that (1) and (2) cited by OP are both identical rating systems? Just the spread of ratings among opponents are higher, on average.

You should be talking about matchup systems, not rating systems, I think.

My server is at T5 and we are fighting against the same 2 servers from last 10 weeks. So the new system I really have no comment XD. It make no difference for us but different colors, which I like it. :P

To me the ranking has no meaning due to imbalanced wvw populations. If Anet can find a different way to resolve the population/coverage issue, any ranking will work. Until then, take it easy and do your best wherever you are. :P

2

Last week’s T4 servers (SoS CD SBI) are moving up in rating by quite a lot compared to last weeks T3 servers (Kain Mag YB) even though we are in the same (or more challenging in the case of SoS) matches. This suggests the new system is working as it means that these six servers will more likely play against each other in future.

DB has increased rating so they are less likely to play against lower ranked servers (but they are more likely to cop SoR BG or JQ).

This new system needs to run for a few weeks – we need to have the T1 people face off preferably against the T2 servers, and poor old T5 broken up. You’ll see huge shifts in ratings when both of those things happen.

I’m for the new system… but for the love of corndogs, please tweak it so T1 can have some fun deviations as well. Not that I don’t… erm… love BG and SoR. That warm and tender love you feel when you’re holding someone under water. It’s just that we could all use some variety. Constant weeks of “Wow, same matchup. Maybe we’ll be BLUE this week, gais!” isn’t… you know, fun for anyone.

fishergrip.4082:

The only reason people are saying 2 is because its only been out a short while. The old system was far better for making balanced matchups.

Go look at the recent history of T3, T4, T5 and back a ways on T8, at the servers who have been locked up tight in inescapable unbalanced matchups, then say that again with a straight face.

Bridget Morrigan.1752:

2

3 is oversimplified, IMO. And yes, I’ll take a chance at a blowout against a much higher ranked server in order to see the wide variety of opponents available with option 2.

What’s wrong with 1 (besides the stagnation), is that it actually creates more inaccurate rankings than 2, not less. In any ranked sports system, such as tennis, the ranking a player maintains has to do with his performance against a variety of opponents, including lower and higher ranked players. There is always a random element to it, depending on who’s playing, who won wildcard placement, who has an injury, etc. Players are constantly facing opponents who are much higher or lower in the rankings than they are.

In order to work out what a servers true ranking is, it can’t just play the same 2 other servers all the time. They have to have a chance to demonstrate all their skills and weaknesses, not just the ones that tend to show versus the two other servers that they play the most. The new system, by adding back in the element of randomness, will allow a broader picture of any given server’s abilities to be reflected in their rankings.

What’s missing is a purpose for the rankings, besides kitten. Granted, that’s enough to drive a lot of people, but having a true tournament or sports season of some kind, with reset rankings at the end and a chance to start over in a new season, and accolades of some sort for the winning servers, and—ideally (although I know this would require some serious work by the developers), a chance for the winning servers to face off in a game-wide championship, NA vs EU, with a wildcard placement for the 3rd server.

What you are describing as an ideal system is not #2…
You are actually considering that rankings mean something. In tennis for instance, yes they meet players of all levels, but in a structured way. If you notice, they always go from easy matches against low ranked opponents and work their way through tougher matches ending up one against the other at some point.

In perspective to #2, even if you win your match against an opponent, you can still lose you ranked position to another server that does not deserve overtaking you.

I’m convinced that, just like myself, its not necessarily facing new opponents that bothers us, its losing the competitive aspect of wvw and striving to overtake those ranked above us and defending our rank from those that want to overtake us.

A win shoud be a win, 2 servers that win their matchup should not exchange ranks just because one of them faced opponents that prevent them from getting a similar variation in rating as the other server… its not a competition anymore… ranks dont mean kitten anymore.

1. is better than #1 as it allows servers to not be stuck in their tiers, but it breaks the competitiveness, makes ranks too volatile.

Losing when winning by a blowout doesnt make sense. Use the ratings and standard deviation + rand to generate matchups, no problem. But use something else, more stable to determine actual ranks, maybe treat it like a sports season. Or reward the winners of the matchup with something.

3, which I understand to be “winner up, loser down” (WULD).

The basic problem with the purely Glicko-based matchups was that there could be ratings gaps between tiers that were never tested anymore: if no server ever crossed the gap, there was no way to ever find out whether the ratings gap was actally justified (one might also call this “incestuous match-ups”).

WULD makes it impossible for such an untested gap to exist, because by definition, each tier boundary is crossed every week, so that ratings from the different tiers are tested against each other.

More pragmatically, here are more reasons for WULD:

• More variety than the old system, but more balanced pairings than the current system (could also be achieved by reducing the random factor in the current system),

• The next match-up is “earned” rather than determined by RNG; in many cases this means a tense fight for the result until the very end of a match (this could also be achieved by publishing the RNG values for the next matchup early(!) during the preceding match-up),

• The fight against your current opponents becomes more meaningful, as you are racing against the servers you are currently fighting, rather than server(s) elsewhere whose expected rating gets close to yours,

• No complex mathematics involved for the next match-up (the Glicko rating can continue to be tracked for server ranking and prestige).

2

Just wait for it to stabilize jeez.

IMO a point reset was in order, but not reset all to 0. Tiers sould have been left in place, but everyone’s numbers should have been reset with a difference of 100.

Tiers got way out of whack by free transfers. The scores were never adjusted after the free transfers ended so there is still a big difference in ratings due to player hopping and bandwagon servers. Reset the scores so last server has 100 and go up from there at 100 intervals. The normal Glicko would work as intended without the player hopping interference. Servers would actually be able to move tiers without having to overcome large rating differences.

Just my 10chars+.

We want change, we want change….ummm…no…wait!…no change, no change!!

All of this is eerily familiar….

Ping.5739:

2

Just wait for it to stabilize jeez.

You’ve missed the point completely…
For instance its not fair for TC lose rank to DB just beacause db got a matchup where they can hold all maps for most of the day while TC has to fight round the clock and is losing rating even though its winning.

THAT is what sucks with this system, and it will keep happening. And the way this system works with its randomness, it will never “stabilize”

2

NA T8 server saying our part. Yah, we are going to lose, what else is new, but at least it not the same match up week in and week out, and next week it will be different.