Guild Wars 2

Thanks for the responses Evan I appreciate it!

thanks for the responses, but if u don’t mind, can you give a time frame when the change is going to go live? 1-2,4weeks?

Evan Lesh.3295:

Amante.8109:

Is the fix to this bug going to be immediate or will we have to wait a while?

Next release will have the updates.

So i’d imagine it’ll be the next patch update o_o

Just be happy that this problem is being identified and dealt with. It is the exact problem that I have been documenting and reporting for quite some time now.

Woo, updating match making…

When are we getting Random Arenas from GW1? Play with a random team till you lose. That kitten was fun! Hotjoin is 99% 1 sided 4v5’s and unranked/ranked takes FOREVER to get started.

well this explains the post i made…..

Evan Lesh.3295:

we’ve found a very sneaky bug in the way we distribute players across the two teams. The goal of the system is to ping-pong back and forth between red and blue team to evenly distribute player skill. In reality, the bug caused the highest rated player between the two teams to usually be teamed with mostly lower skilled players. For matches with high skill spread, this could result in one player feeling like they were matched with a surprisingly low-skilled team. This was hard to notice because matchmaking was always picking ten similarly skilled players to start with, so the problem only really presents in the high and low MMR brackets where there are fewer players to choose from.

I’m on a PvP team and often duo/trio queue when we cant get 5, most of us are ranked in the top 200. You don’t know how many times we get the most clueless team mate. Thank you so much for this change.

If this means the end of skewed match-making, then I’m entirely looking forward to it.

Evan Lesh.3295:

Magical stuff

I have been critical about match making and the PvP infrastructure in the past, so this is a good opportunity to be equally vocal in praise. Thank you for continuing to identify and fix issues with the match making. Thank you for listening to the empirical feedback from your players and investigating issues that might not be immediately visible in your data. Keep up the good work.

Is MMR only in ranked or is it in unranked too? I’ve played 300 games on Thief, 150 games on Warrior, and about 100 on Necro and mostly play unranked but I still get stuck with players that don’t know what they are doing and rage quit or log off right after they click on “accept” when matchmaking pops. Why the hell would you click “accept” if you’re not going to play? That really annoys me. They need a more reliable report system for leavers like that like the one in League of Legends.

Wilco.1458:

EARL.8312:

what about the pug unfairly matched vs premade guild?

Sounds like working as intended. That’s why they removed solo queue, right?

I was wondering the same, can we go back to solo players vs solo players instead of solo players vs organized premade teams so we don’t have to get slaughtered every match. That’s the most unfair thing about pvp really

This is really good news!

Unfortunately you found the bug 1 month before HOT and the change of the whole pvp system. )

So, maybe we can have some good matches now until HOT is realeased.

EARL.8312:

what about the pug unfairly matched vs premade guild?

Cuchulainn.7421:

This is really good news!

Unfortunately you found the bug 1 month before HOT and the change of the whole pvp system. )

So, maybe we can have some good matches now until HOT is realeased.

You really think they will throw the matchmaking system in the trashcan because of leagues ? LeL

Thanks Anet!

But honestly, the only time I get frustrated with the matchmaking is when as a 100% soloer in the unranked queue I get matched against a premade there to do nothing but grief people.

Guess you cannot account for everything…collect 500 and pass go…

Your matchmaking still sucks and is putty low skill players with higher skill players

Evan Lesh.3295:

Mewling Kittens.7965:

Noob question… what’s MMR?

It stands for Match Making Rating. We use the Glicko 2 algorithm to assign a number to each player that represents skill level in order to match similarly-skilled players together.

Details here:
https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm
https://en.wikipedia.org/wiki/Glicko_rating_system

So, just a quick question……. Is anyone looking into the Glicko rating system for WvW? Could we have the same bugs on our end, will this fix you did affect us, too?

Kitiara.2706:

Evan Lesh.3295:

Mewling Kittens.7965:

Noob question… what’s MMR?

It stands for Match Making Rating. We use the Glicko 2 algorithm to assign a number to each player that represents skill level in order to match similarly-skilled players together.

Details here:
https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm
https://en.wikipedia.org/wiki/Glicko_rating_system

So, just a quick question……. Is anyone looking into the Glicko rating system for WvW? Could we have the same bugs on our end, will this fix you did affect us, too?

This wouldn’t affect World vs World. It’s not a problem with Glicko, it’s a problem with the way the system uses Glicko to create teams.

So I usually attribute complaints about match making to just general whining about team mates, but it looks like there were some legitimate problems here.

My question is, how do we know when we are getting unlucky with match making versus when there is a real problem? I guess what I’m asking is…. how do we give realistic feedback that’s meaningful on this issue, as opposed to complaining about the MMR system each time we get a bad team? Seems tricky.

Evan Lesh.3295:

In reality, the bug caused the highest rated player between the two teams to usually be teamed with mostly lower skilled players.

Ouch… That’s what I felt but it’s nice to have it confirmed.

Hey, what if the MMRs of each team (without player names) were put in Game Messages after games? So we can help debug? Something like…

Red Team – 1500 Average – (1600) (1500 1500) (1450 1450)
Blue Team – 1580 Average – (1650) (1600) (1550) (1800 1300)
Expected Result: Blue 500 – 350 Red

This way we can see the actual teams, MMRs, etc and know for sure it was a garbage match?

Oh wow, this is probably first positive news in months…. I actually thought it was intended algorithm that punished you for having higher win rate than 50% and would put you on absurdly bad teams vs really good team so you lose enough to go back to 50%.
This will be huge improvement for solo players with high MMR. Thank you (if it actually works).

Hopefully no more defending 2 points by myself while entire team is being perma farmed at spawn point; no more 3 ppl on my team chasing 1 mesmer on beach and getting wiped; on the other hand i won’t be able to solo 2 medi guards on thief (lol) while defending mid point anymore (yes, that happend) xD

I got to admit tho, perma farm players at there base is kinda funny when it happens. :P

how about this thing from closed beta where you could rate how good content was
so at end of pvp match we rate from 1 to 5 for how we felt the matchup was and the individual players of course only for anet internal usage as addition to their automatic mmr ?

This proves that this Forum is a valuable resource for the devs to keep improving their product.

Improved product = happy customers

The biggest challenge for them is knowing what to filter out. There are legit complaints on here but the majority of them are learn to play issues….because it’s easier to whine than it is to learn.

Glad Evan listened to the gut instincts of the right people!!!

Wow, this sounds amazing.

Okay I’m excited now. I feel like this could go a long way toward giving better matches that make a lot more sense to be in. Crossing fingers!

Vapour.7348:

This proves that this Forum is a valuable resource for the devs to keep improving their product.

Improved product = happy customers

The biggest challenge for them is knowing what to filter out. There are legit complaints on here but the majority of them are learn to play issues….because it’s easier to whine than it is to learn.

Glad Evan listened to the gut instincts of the right people!!!

Am I a right people? (sorry for my english, it’s my primary language and only language)
Since I’m modest, I will say I’m not the right people even though my comments are always spot on coz Im the best. Too bad Im so humble.

Lee.5120:

Is MMR only in ranked or is it in unranked too? I’ve played 300 games on Thief, 150 games on Warrior, and about 100 on Necro and mostly play unranked but I still get stuck with players that don’t know what they are doing and rage quit or log off right after they click on “accept” when matchmaking pops. Why the hell would you click “accept” if you’re not going to play? That really annoys me. They need a more reliable report system for leavers like that like the one in League of Legends.

Ranked and unranked each have their own MMR according to a recent post from a dev.

The problem is that youre overcomplicating things Evan.

You created an MMR system with 20 variables while only 2 or 3 are needed.

Youre like the guy that tries to act like a really smart guy while hooking up with girls… but who ends up going home alone at the end of the night since he made himself look like a major kittenhead.

Stop being arrogant and trying to reinvent the wheel. All we need is a simple and transparent MMR system similar to what for example world of warcraft has. AND split team and soloq.

I see were still losing game after game to premades.

While the devs are still trying to figure out how their formulas work, would they please consider letting us mortals queue outside of this heart of the mists?

Your matchmaking is still pathetically awful. You fixed nothing.

I wanna say something and before u start to answer think.
I have 4 accounts that i play (Why? Is my problem) I play almost only PvP.
The matchmaking really scks. Some days my 2 accounts are so easy doing daily and fast that i can’t believe ( in pvp always)
The other 2 to get 1 win takes sometimes 10-12 games. This thing make’s me feel that account’s random targeting and the kitten them up in matchmaking . I do not not know the reason but it happens. I pay attetion this thing the last 5 months everyday the same. Sometimes is better to leave a game and log another account ( this i try for science is not something that i like ). Why this happen?
Why?
Please answer something logical and not only for paserby my question.
And really i don’t care if is difficult or not to find a solution ! 3 years i support u with gems buy accounts etc . I want an honest answer.

Grimegg.1297:

Your matchmaking is still pathetically awful. You fixed nothing.

Fix is not live.

Maybe tuesday

Locuz.2651:

The problem is that youre overcomplicating things Evan.

You created an MMR system with 20 variables while only 2 or 3 are needed.

Youre like the guy that tries to act like a really smart guy while hooking up with girls… but who ends up going home alone at the end of the night since he made himself look like a major kittenhead.

Stop being arrogant and trying to reinvent the wheel. All we need is a simple and transparent MMR system similar to what for example world of warcraft has. AND split team and soloq.

It is an MMR system, it’s ran by Glicko2.
You’re wanting high win rate = skill level and that does not equate to skill. That’s only if it’s a premade team vs a premade team – it doesn’t help the solo, duo, trio queue players at all where winrate is always around 50%.

The old system had huge issues with it…. I don’t want to go back.

Ajax of Telamona.6974:

I wanna say something and before u start to answer think.
I have 4 accounts that i play (Why? Is my problem) I play almost only PvP.
The matchmaking really scks. Some days my 2 accounts are so easy doing daily and fast that i can’t believe ( in pvp always)
The other 2 to get 1 win takes sometimes 10-12 games. This thing make’s me feel that account’s random targeting and the kitten them up in matchmaking . I do not not know the reason but it happens. I pay attetion this thing the last 5 months everyday the same. Sometimes is better to leave a game and log another account ( this i try for science is not something that i like ). Why this happen?
Why?
Please answer something logical and not only for paserby my question.
And really i don’t care if is difficult or not to find a solution ! 3 years i support u with gems buy accounts etc . I want an honest answer.

There are several factors as to why.
Matchmaking definitely doesn’t show favoritism nor is it random.

Assuming you’re running the exact same class with the exact same build across all 4 accounts – matchmaking is ultimately based on mmr. The longer you play a class on that acount, the more “true” your mmr is going to be for that class.

Judging by “it takes 10 games to get a win”, that account has a high mmr setup for you. The only way to truly know is if we can see our mmr (make it happen anet!)

From what I understand, mmr isn’t seperate on classes, it’s seperated by game modes. If you play Mesmer for the first time in Rank PvP and do terrible, it’s going to lower your mmr on your main class that you usually do good on.

(It’s like this because people would choose a new, low, mmr class to get a newbie matchmaking then switch to a high mmr class for the win. They haven’t set an angorithm to fix it, from what I understand)

So unless you play the exact same number of games with the exact same build across all 4 acounts, there’s going to be differences with your matchmaking because all 4 accounts have differences in mmr.

Saiyan.1704:

It is an MMR system, it’s ran by Glicko2.
You’re wanting high win rate = skill level and that does not equate to skill. That’s only if it’s a premade team vs a premade team – it doesn’t help the solo, duo, trio queue players at all where winrate is always around 50%.

The old system had huge issues with it…. I don’t want to go back.

NO higher win ratio shouldnt equal higher leaderboardrank/MMR. I never said that… Gaining/losing MMR should be based on who u lose or win against.

Soloq should be seperated from teamq anyway…

And dont come with the “the community isnt big enough to support that” arguement. It was big enough before december 2014 and according to a a post a few days ago from Anet the size of the community is bigger than ever.

What issues?

Pre december patch queueing solo would mean ud face other soloq players.

Pre december patch queueing as higher rated player meant you would face ppl at the top of the ladder 90% of the time. Which meant you would know a big percentage of the players youd face. The games where actually fun back then….and not as absurdly frustrating as they are now at times. Its barely playable atm…

No duo/trio or quadruple capping empty points.
No truckloads of ppl rushing bosses early.
Not seeing 3-4 teammates chasing 1 single thief.
Ppl dodging (lol).
Not seeing teammates die in 3v1s.
Etc etc

Saiyan.1704:

Locuz.2651:

The problem is that youre overcomplicating things Evan.

You created an MMR system with 20 variables while only 2 or 3 are needed.

Youre like the guy that tries to act like a really smart guy while hooking up with girls… but who ends up going home alone at the end of the night since he made himself look like a major kittenhead.

Stop being arrogant and trying to reinvent the wheel. All we need is a simple and transparent MMR system similar to what for example world of warcraft has. AND split team and soloq.

It is an MMR system, it’s ran by Glicko2.
You’re wanting high win rate = skill level and that does not equate to skill. That’s only if it’s a premade team vs a premade team – it doesn’t help the solo, duo, trio queue players at all where winrate is always around 50%.

The old system had huge issues with it…. I don’t want to go back.

Ajax of Telamona.6974:

I wanna say something and before u start to answer think.
I have 4 accounts that i play (Why? Is my problem) I play almost only PvP.
The matchmaking really scks. Some days my 2 accounts are so easy doing daily and fast that i can’t believe ( in pvp always)
The other 2 to get 1 win takes sometimes 10-12 games. This thing make’s me feel that account’s random targeting and the kitten them up in matchmaking . I do not not know the reason but it happens. I pay attetion this thing the last 5 months everyday the same. Sometimes is better to leave a game and log another account ( this i try for science is not something that i like ). Why this happen?
Why?
Please answer something logical and not only for paserby my question.
And really i don’t care if is difficult or not to find a solution ! 3 years i support u with gems buy accounts etc . I want an honest answer.

There are several factors as to why.
Matchmaking definitely doesn’t show favoritism nor is it random.

Assuming you’re running the exact same class with the exact same build across all 4 accounts – matchmaking is ultimately based on mmr. The longer you play a class on that acount, the more “true” your mmr is going to be for that class.

Judging by “it takes 10 games to get a win”, that account has a high mmr setup for you. The only way to truly know is if we can see our mmr (make it happen anet!)

From what I understand, mmr isn’t seperate on classes, it’s seperated by game modes. If you play Mesmer for the first time in Rank PvP and do terrible, it’s going to lower your mmr on your main class that you usually do good on.

(It’s like this because people would choose a new, low, mmr class to get a newbie matchmaking then switch to a high mmr class for the win. They haven’t set an angorithm to fix it, from what I understand)

So unless you play the exact same number of games with the exact same build across all 4 acounts, there’s going to be differences with your matchmaking because all 4 accounts have differences in mmr.

Wrong! Every day is different account! and all accounts are Dragon!

Locuz.2651:

Saiyan.1704:

It is an MMR system, it’s ran by Glicko2.
You’re wanting high win rate = skill level and that does not equate to skill. That’s only if it’s a premade team vs a premade team – it doesn’t help the solo, duo, trio queue players at all where winrate is always around 50%.

The old system had huge issues with it…. I don’t want to go back.

NO higher win ratio shouldnt equal higher leaderboardrank/MMR. I never said that… Gaining/losing MMR should be based on who u lose or win against.

It’s kind of already based on that.
http://dulfy.net/wp-content/uploads/2014/11/gw2-pvp-ladder.jpg

Locuz.2651:

Soloq should be seperated from teamq anyway…

I agree, soloQ would help the casual player base who prefer to play against other solo players. Hotjoins should be 100% solo players only. If you have a friend who wants to queue, or on a full team, it should all be played on Unranked PvP.

The previous SoloQ leaderboard was the most frustrating, ridiculous system ever. It was not based on skill level, it was based on pure luck – any player who won 10 games in a row would be on the top of the leaderboards.

Ajax of Telamona.6974:

Saiyan.1704:

Locuz.2651:

The problem is that youre overcomplicating things Evan.

You created an MMR system with 20 variables while only 2 or 3 are needed.

Youre like the guy that tries to act like a really smart guy while hooking up with girls… but who ends up going home alone at the end of the night since he made himself look like a major kittenhead.

Stop being arrogant and trying to reinvent the wheel. All we need is a simple and transparent MMR system similar to what for example world of warcraft has. AND split team and soloq.

It is an MMR system, it’s ran by Glicko2.
You’re wanting high win rate = skill level and that does not equate to skill. That’s only if it’s a premade team vs a premade team – it doesn’t help the solo, duo, trio queue players at all where winrate is always around 50%.

The old system had huge issues with it…. I don’t want to go back.

Ajax of Telamona.6974:

I wanna say something and before u start to answer think.
I have 4 accounts that i play (Why? Is my problem) I play almost only PvP.
The matchmaking really scks. Some days my 2 accounts are so easy doing daily and fast that i can’t believe ( in pvp always)
The other 2 to get 1 win takes sometimes 10-12 games. This thing make’s me feel that account’s random targeting and the kitten them up in matchmaking . I do not not know the reason but it happens. I pay attetion this thing the last 5 months everyday the same. Sometimes is better to leave a game and log another account ( this i try for science is not something that i like ). Why this happen?
Why?
Please answer something logical and not only for paserby my question.
And really i don’t care if is difficult or not to find a solution ! 3 years i support u with gems buy accounts etc . I want an honest answer.

There are several factors as to why.
Matchmaking definitely doesn’t show favoritism nor is it random.

Assuming you’re running the exact same class with the exact same build across all 4 accounts – matchmaking is ultimately based on mmr. The longer you play a class on that acount, the more “true” your mmr is going to be for that class.

Judging by “it takes 10 games to get a win”, that account has a high mmr setup for you. The only way to truly know is if we can see our mmr (make it happen anet!)

From what I understand, mmr isn’t seperate on classes, it’s seperated by game modes. If you play Mesmer for the first time in Rank PvP and do terrible, it’s going to lower your mmr on your main class that you usually do good on.

(It’s like this because people would choose a new, low, mmr class to get a newbie matchmaking then switch to a high mmr class for the win. They haven’t set an angorithm to fix it, from what I understand)

So unless you play the exact same number of games with the exact same build across all 4 acounts, there’s going to be differences with your matchmaking because all 4 accounts have differences in mmr.

Wrong! Every day is different account! and all accounts are Dragon!

/facepalm

Saiyan.1704:

The previous SoloQ leaderboard was the most frustrating, ridiculous system ever. It was not based on skill level, it was based on pure luck – any player who won 10 games in a row would be on the top of the leaderboards.

Nope.

The old system was great besides one thing:

New accounts would have MMR that would fluctuate to much during the first 10ish games. So only extremely new accounts would get this effect that u mentioned; where a few lucky wins (due to 4v5s for example) would propel them to top leaderboard ranks.

Older accounts needed to play and win a lot vs top 200 oppponents to even get from top 100 to top 10. Iirc i would gain/lose 10-15ish ranks on average while in the top 100. The lower the team average the higher the risk and the lower the reward ofc.

Evan Lesh.3295:

I
…In reality, the bug caused the highest rated player between the two teams to usually be teamed with mostly lower skilled players. For matches with high skill spread, this could result in one player feeling like they were matched with a surprisingly low-skilled team….

I KNEW IT

I couldn’t have put it into words or nail it down like that but yea. That’s what I’ve been feeling.

…hopefully I’m in the higher tier MMR and not the low end xD

hmmm…I do not read any word of apologize to all the palyers that are beliving in a faithful way in your talent of programming and saying every day: No, they do a really good job, THERE IS NOTHING WRONG IN THIS MATHMAKING.

Is this your REALLY serioues statement to players that do 6000 matches in PVP and allways playing with 4 others, that are noobs, cause of your bug?

The difference between games is absurd.

Of the last 3 games my first 2 where good. Games with what seemed to be evenly matched players. Probably the best two games i played this week.

Than last game happened. I saw in the starting area that my entire team was filled with people with pve titles and what seemed to be pve/wvw specs. We had two thieves with 2 daggers for example. Minutes in to the game i saw the deathblossom spam.

But we actually won since the opposing team was equally bad. Filled with rifle warriors, thieves that just ran to my engi without stealth or openers or w/e. Its hilarious, but wtf why do i have to play with and vs these guys?

My starting MMR shouldve been high enough and it seems that i win 80% of my games as of late so the MMR shouldnt drop that much, should it?

was this in todays patch ? because i dont see it mentioned

I’m also curious as to if this was in todays patch.

I played some SPVP tonight. FWIW, it seems to be working at least somewhat. Out of 7 games I got to play, 2 were really close and 2 were still competitive even if they weren’t super close. Buried the other side on one and got blown out on the last two, but this is much better than before. Sample size too small to confirm but for whatever it’s worth even the blowouts were much less frustrating than the usual blowouts.

Didn’t get a lot of time to play tonight. Tomorrow I’ll run some matches and record the data. Maybe I’ll do it for the next 7 days.

I seem to be getting better teams now, less silent clueless people that do nothing all game.

I did lose quite a bit though that was mostly due to being countered h_rd on* a profession basis, or trying out things with patch changes.

P.S.

Pistol / Pistol thief and rifle warriors are still trash. I really wish my f1 was as strong as skillshroud =x

Would also be nice if the game disallowed more than 2 of one profession on the same team. It’s really frustrating to play vs 3 necro / 2 mesmer.

*(lol, those 2 words got kittened XD)