First the disclaimer, I’ve benefited from the changes to MM in S2. I reached Legendaryx2 in S1, and I’ve played 6000 matches, and this has lead to two things being in my favor for season 2: my MMR is relatively high and my MMR is very stable. Both of these are huge advantages in Season 2. A high MMR will put better players on my team, and a stable MMR will protect me from a losing streak artificially lowering my MMR and setting me on a downward spiral. And… what do you know, getting to Legendaryx2 was very easy for me in Season 2 (I reached Legendaryx1 in 170 games played, and Legendaryx2 in another 30 to 40 games).
So why am I writing this post? I have been one of the people who are skeptical of the complaints about MMR hell. I still believe that in some cases the people who are complaining about MMR hell are in fact just hitting their current skill ceiling. But, that said, I also believe the current MM is flawed, and flawed in some ways that people haven’t yet mentioned.
Flaws in Season 2 Matchmaking:
1) The “average MMR” for a particular division and tier is NOT constant. It varies by:
1a) Time since start of season. Amber is very hard on day 1, but by the 2nd or 3rd week has probably settled down.
1b) Time of day. There are fewer people playing during off hours. And Ive noticed that the level of competition can vary by time of day, perhaps certain people play in certain time ranges. Things are most competitive on weeknights from server reset to reset+5 hours. People who hit their appropriate tier playing weekday mornings, might go on a massive losing streak if they play on a weeknight or weekend when competition is greater.
2) People can pair up with other players (or a premade team) and get carried to a division/tier that is beyond their own personal skill level. Then when they try to soloq they can go on massive losing streaks and be stuck in MMR hell because they cannot drop tiers or divisions to reach their appropriate skill level.
3) People can get into a MMR negative feedback loop. If they have fewer matches played, their MMR can vary more. If they get unlucky and lose matches that the MM thinks they should have won their MMR will drop quickly, and will cause them to get poorer team mates, and set up a viscious cycle. If this happens enough times, and especially if they have already hit the division appropriate to their skill, they will likely keep losing matches. And worse still, even if they get better, that increase in skill may not cause an increase in MMR (as it ideally should) if they keep getting paired with team mates who just can’t win matches at that division level.
All in all, I think season 2 was an improvement over season 1 in terms of allowing people to reach a more appropriate division quickly. But it fails in terms of creating matches that are competitive. Even in Legendary I see more blowouts than I see close matches.
Some things they could do to improve the system:
1) Pre-sort players into divisions at the start of the season either by MMR or by their division in the previous season.
2) or Hold a pre-season qualifying period to allow people to get a higher placement at the start of the season.
3) Assuming the implement either suggestion #1 or #2 to presort people into divisions, then they could stop using MMR to determine matches. Instead they could just use division and tier to randomly draw both team mates and opponents. This would eliminate MMRhell from the equation. Who you would play with and against would purely be a function of your current division/tier, and in that case assuming your allies and opponents are drawn from the same pool on average the main variable would be your own skill relative to the other nine people, allowing players better than their div/tier to be more likely to win and players worse than their div/tier to be more likely to lose.
My suggestions do not create a perfect system. In fact there would still be blowouts. But it might be an incremental step forward from what we have in Season 2.
Discuss.
@LordOtto.2650
What criteria are you using to evaluate my build? Because I can tell you that I made Legendaryx2 a few nights ago, and that build was doing very well, even against ESL level players. And I soloqed through almost all of Diamond and Legendary, and did a mix of solo/duoq for everything else.
So you mean to tell me 2 warriors a thief, a ranger and a DH and dear a team of 2 ele’s and 3 reapers if all the players are of equal skill?
No of course not, but that team of crappy classes/builds has the same ability as anyone else to swap to an appropriate class comp before the match begins. The fact that they don’t is one contributing factor to their difficulty progressing past their current division.
Elia, I am skeptical that MM is the problem. You seem to have some knowledge of good rotations/deployments, yet you are stuck in Sapphire. If you don’t mind my asking, what build(s) do you run? What role do you typically play with those builds? What is your strategy for carrying your team, and how to do you execute that strategy?
To answer the questions I posed to you:
1) For this season I main a hybrid reaper built to sustain/pressure in team fights. Build: http://gw2skills.net/editor/?vRAQNBmWD7kZTocTszGwbTgeTs4NYUdgQQha6iFAiVxJwI4FE-TJxFABCt/AAXEg9LDkwTAAA
2) Role is team fighter, though the build can win 1v1s and can sustain 1v2 until help arrives.
3) I always try to head to the mid fight initially and help my team win that. I pressure the focus target and am always prepared to pop shroud and cleave on downed. I leverage my 2 health bars to survive focus fire, and I spread chill to as many enemies as possible. I have 2 condi transfers (anti necro tactics), as well as 2 additional single condi clears (for emergencies). I also have boon corruption with scepter AA which I use on eles and engis (and other classes with huge boon stacks) to pressure them.
Anyone can play my build, and anyone who masters it and uses it will carry up to Diamond at least. Class stacking is only an issue for people who refuse to adapt to the meta.
Once I got to legendary, I needed to get 10 wins on 2 other classes, so I also played Engi and DH, and was able to win 10 matches with those classes as well with a 60% win rate. So it doesn’t just require an OP FOTM build to win in Legendary.
The OP needs to focus on picking a meta build, learning the mechanics of that build (using the proper skills at the proper time), learning the builds role and making sure he’s at the right place and the right time. If he does that I’m pretty sure he’ll rise quickly out of Emerald. This really is a L2P issue here.
Playing a Staff Ele without auras is just plain bad. Even with good mechanical skills and good rotations (which OP seems to lack from other people’s review of the video) it’s going to lead to extra losses. Start with a good build and go from there. Don’t handicap yourself playing a terrible build…
Hey ArrowSlanger, congrats on legendary via solo queueing.
While I didn’t soloq the entire way from amber to legendary, I did most of diamond as soloq and 4 tiers through legendary soloq. It can be done. Just to be clear, any match where I didn’t soloq I instead duo queued with a friend of similar skill and MMR.
(edited by Soul.9280)
- - make a change to address the problems with huge MMR disparity between teams in ranked matches at the start of each pvp season that occur until the players finally reach an appropriate division. These hugely unbalanced matches (and long losing streaks for “less experienced or weaker players”) hurts the pvp community and causes some people to quit (which none of us want). Please do NOT revert to season 1, as that also wasn’t fair. Three possible solutions:
#1a – implement a pre-season that gives people a chance to sort out their starting division
#1b – or just determine starting divison placement based on previous season division: sapphire (if legendary/diamond), emerald (if sapphire/ruby), amber (if amber/emerald)
#1c – place people into divisions based on MMR (not ideal, but could work if you don’t want to implement #1a or #1b). - - do a better job of balancing professions prior to each season. One huge patch isn’t enough, you guys often over nerf or don’t take into account how your nerfs may help other professions. A beta test server and a small balance patch just before each season (in addition to the major patch just after each season ends) would help a lot with this.
- stop forcing soloq people to go up against essencially full premade teams, the change you made for this season doesn’t really fix this, as currently “2 + 1 + 1 + 1 vs 5” and “1 + 1 + 1 + 1 + 1 vs 4 + 1” happen and frustrate the teams with the majority of the +1s.
- don’t give starting players in ranked pvp “average mmr”, instead start them with the lowest mmr and move them up quickly if they win a lot.
- show us more than just the top 5 players during the season. I really liked the original pvp leaderboards because I could see where I stood and could see upward movement if I got better. Could you re-enable the pvp leaderboards and have it only show us the results for the current season (ie reset it at the start of each season)?
(edited by Soul.9280)
Now this is the right way to deal with a problem! Rather than complain, or blame others, or ask the devs to make changes, he’s taking the initiative to organize other players to take charge of their own destiny. My hat is off to you sir!
I see a lot of people complaining on the forums about how they are good players but now caught in a possible MMR spiral. Like many I am concerned about the health of pvp in this game. I value having a large community of people who enjoy pvp.
But I wonder, how many of the people who are claiming to be good but stuck in emerald and in an mmr spiral are the people who cheesed season 1 by tanking their mmr at the start and then riding a wave of easy matches to legendary.
That seemed to be a thing in season 1 for a small minority. There was a lot of complaining on the forums about how cheesy that was and many people were angry that it could even be a thing. So now, they fix things so that having a high MMR is an asset rather than a liability, which is a good change. Presumably some people are now paying the price for their poor decision to tank their MMR. I wonder if some of the people who are complaining have considered that?
1) Day 1.
2) Diamond division, tier 1
3) No, probably about 70% wins this season
4) Still advancing. Won 10 out of 12 three days ago, lost 3 out of 4 last night, won 7 in a row last night and then split the next 8 matches.
5) No
6) Yes. I’m aiming for legendary division and the new title. I’m also trying to finish out the fourth/final of the ascension acheivements.
(edited by Soul.9280)
Only registered and stable teams should participate.
Really? Do you realize that soloq people account for the vast majority of people playing ranked? So if Anet were to follow your advice it would be GG because the size of the population left playing ranked would be too small to be viable.
Individual skill does matter in this system, and part of the skill being measured is your ability to figure out how best to fit in with the team mates that you have for that match. People who are good at adapting will win more games on average than those who aren’t, and hence will have a higher MMR.
Just “out necro” them, if they have 3 necros bring 4 of your own. ;-)
Ok, in all seriousness, mobility does work well against necros, and coordinated power burst with cc chains also works well especially at the start of a match. Scrappers and Power Revs are good for this.
Eater of Peeps, I feel for you, it can be very demoralizing to lose many matches in a row and have a low win percentage.
If you don’t mind my asking, what class/build are you playing? Perhaps there is more you can do to help you team by making better choices about what role you play to carry your team.
Not specific to Eater of Peeps, many people on the forums complain that “while I was capping home, the other guys on my team rushed mid and wiped”, well, then why were you capping home? If you want to carry you need to play a build that does well in team fights and can help your team win the team fight. Winning the team fight typically means winning the match. So if you want to carry bad team mates and escape ELO hell and a low division, you need to play a strong team fighter and help your team. I play a sustain oriented reaper build with good survivability and good damage, and I anchor team fights, it’s one reason why I’m in Ruby now.
I guess what I’m trying to say is, look beyond mechanical skill and rotations. While those are important, that’s not all that’s important.
Matches that end 500-500 still have a winner and a loser. Winner is determined by adding the personal scores of everyone on each team, the winning team is the one with the highest total of the sum of personal scores. So the pip exploit farming that Pandratrice is suggesting in the above comment isn’t possible.
How about instead of giving people pips for losing a close match, we instead just don’t take away pips for losing a close match. That way, people don’t get artificially boosted for losing, but they don’t get penalized for being competitive and completing a very tight match (which suggests that they are in the right division).
It would be very nice if the game would show us our MMR level… and nicer still if it would show us the MMR level of the other players in the match.
Also, will the PVP leaderboard be reset and reactivated for Season 2? Please?
I’d be interested in more statistics about season 1 as well, but I can tell you from my own experience in Legendary division for the final 2.5 weeks of the season that there actually weren’t that many 5-person premades in the matches I played. I speculate that the reason is that the top players quickly realized that the matchmaker unduly penalized them (giving them an inflated MMR) for being in a 5 person group, and that they would win more matches as a 3 person group (or even solo queue) than as a 5 person group. I do remember going up against 2 and 3 person groups of ESL level players from the same ESL team, and those were very difficult matches for me to win in a 1-1-1-1-1 pug or even a 2-1-1-1 pug.
Now, in season 2, with MMR not a factor in terms of finding opponents, there won’t be a disincentive to group, in fact the opposite is true, it will be a bigger advantage to group up with (and use teamspeak) other people of your MMR level (or higher!) as long as you all are in the same division.
(edited by Soul.9280)
Queue watching will also be a huge deal (even more so than in Season 1), as anyone who’s in a high level division will be working very hard to avoid getting matched up against the ESL teams.
I’ll offer a prediction for season 2, the ESL (and other high level) teams will queue in ranked as 4 players rather than 5 to get around this change, and due to the imbalance inherent in 4 people who have synergistic builds and are in team speak going up against 5 PUGS will continue to have an unfair advantage. I’m not worried about the lower divisions due to the positive changes you guys have announced… I’m talking about dealing with this at Diamond and Legendary.
Some other combos that are inherently not fair:
5 vs 2-1-1-1
4-1 vs 1-1-1-1-1
4-1 vs 2-1-1-1
The matchmaking in unranked is so bad right now, last night in one of my matches one of the people assigned to my side disconnected during the 90 seconds before the match started and never came back, and in spite of a 4v5 my team won (though it was a very close match). The other team was full of bad players, whereas I got to Legendaryx2 in season 1 (and while not ESL level am clearly not “new” or “bad”), and at least one person on my team is a stronger player than I am (I’ve grouped with him many times). I appreciate that we won, and I enjoyed the challenge, but that match-up was in no way shape or form “fair” or even close to fair if we were able to win without a player (even if he likely was our worst player).
(edited by Soul.9280)
I was hoping for a small patch halfway in between the Jan 26 patch, and the start of League season 2. Still hoping…
My bad, I missed the “power” qualifier. I’m not sure however that that qualifier is important when defining a bruiser.
Reaper would have to be included in this list, even if they are mostly ranged.
Evan, the last night of League play, I saw lots of people swapping classes during the 90 seconds wait period before the match started, and I was playing in the Legendary division against many of the top players. And I did see a surprising number of people who queued as thief or warrior and then switched to something meta.
(edited by Soul.9280)
They have the infrastructure in place to provide most of this info, as they already have a Leaderboard for ranked pvp, all they need to do is reset it at the start of each League season and the data is there for all to see. I’m asking for them to do this retroactively for Season 1, some work possibly, but not much to ask.
I’d like to see which division everyone who participated in Season 1 ended up, how many games they played, wins/losses, and favorite class.
Are there any plans to make this available?
I completed the Primordial Legend achievement and gained access to the title about 10 days ago when I reached Legendaryx1 (ie Division 6, the first of the Legendary divisions).
I received the Primordial Legend title when I made Legendaryx1 (Division 6), this happened about 7 days ago. I made it to Legendaryx2 last night. I did alot of duoq to get through Sapphire/Ruby, but most of Diamond/Legendary was soloq as bunker mesmer.
Evan, will there be some type of “Season 1 results” board that shows the number of games played, won, lost, % win, and final division for all participants?
Could you comment on any of these well thought out requests for improvements?
- Better visibility into upcoming balance changes.
- A Public Test Server where the community can experiment with proposed changes and provide early and informed feedback
- Corrective minor balance patches in the days/weeks after the quarterly big balance patches to make things better going into each PVP Season.
- Clarity of vision for the various classes (what their role might be in pvp) and what balance might mean for each class
- Mechanism to fix a 4v5 (due to dc or quit) in the 90 seconds prior to match start
Lastly, it seems like some of the announced changes for the quarterly balance patch are knee jerk and over the top, likely killing Mesmer in PVP for the next season. Can you comment on how you guys determine the overall impact of many different but interlocking changes (soldier/sentinel amulet removal, removing quickness from rez/stomp, well of precog nerf, alacrity nerf, and significant buffs to thieves) on a class like mesmer? Because many of us who are in the top tiers of pvp think that instead of simply bringing mesmer back into balance you’re destroying it outright with this balance patch. Do you remember the days before June 23 when a mesmer was essentially a free kill for a thief assuming similar skill level?
What makes you think this is the case? What data do you have to support this?
There are many things to love about pvp in GW2.
That said, the dev team (and their managers) have repeatedly demonstrated a lack of … polish?
- Visibility into changes is poor.
- There is no Test Server.
- Feedback from the most informed from the player community is either ignored or glossed over.
- Balance patches are too infrequent, and the current model of huge changes every 3 months followed by almost no corrections is a flawed model, though admittedly better than what we used to have (which was balance patches every 6 to 9 months).
- There is no clarity of vision for the various classes and what balance might mean
- They leave certain classes at the top of the food chain for too long, while others are non-viable (also for too long).
- They often show a profound lack of understanding about how their classes work and what previous metas have showed us.
- There is minimal interaction between the devs and the player community. Evan and Josh occasionally post here (thank you guys!) but mostly we’re ignored.
- There are still problems with matchmaking, leavers/afkers, and pip gain/loss that remain unaddressed (like people disconnecting during the 90 second wait period before the match starting, and the game letting the match proceed in spite of this).
On the community side of things:
- We have too many players who have vendettas against certain classes and spout non-sense at every turn trying to get their class buffed and their hated classes nerfed.
- People get frustrated by things, and have knee jerk reactions to over nerf rather than considering the repercussions of the changes that are being proposed. Sometimes there is a fine line between a class/build being OP and it being non-viable.
Just some thoughts from someone who’s played GW2 for most of the past 3.5 years.
To be more clear. Yes this current bunker meta isn’t very good. Yes the dev team needs to make some corrections to bring it into balance. The problem is that instead of getting shaves and corrections we’re getting wholesale hacks and knee jerk reactions. And if the past is any indication they will release their huge balance patch with limited or no testing (at least not by the playerbase). And then we’ll be stuck with a new but equally broken meta for the next 4 months. It won’t be a bunker meta, but it will be no less broken. A broken meta is one where a few classes/builds dominate and win all matchups, making many classes completely unviable.
The solution is more visibility into proposed changes, not just in terms of documentation, but also a Test Server. And a small correction patch (or two) a few weeks after the huge patch, and before the next season starts.
The old adage goes that if you put enough monkeys in front of type writers eventually one of them will write a work of Shakespeare. Sadly, Anet is following a similar course for MMO class balance design, only they don’t have nearly a large enough sample size, and appearantly no oversight to distinguish between the “good” and the “bad”. I’d rather pay a monthly subscription and get a game that is good, than get this free to play garbage. Devs, you guys aren’t taking very good care of your assets… your playerbase. Please do something to correct this before it’s too late.
Impact, did you fail to mention Revenant intentionally (because they will be trash after the patch) or due to oversight? You also failed to mention Druid/Ranger… and given the lack of nerfs to Druid, I’m thinking Druid will still be in the meta.
If you’re going to get rid of Soldier, Sentinel and Celestial, then giving us these amulets as a replacement is the right approach to plug the gap. I do worry that such big changes may force many classes out of the meta, but we’ll have to see how it shakes out. Giving us a +healing amulet that also has +vitality and +power is a big improvement, as it gives Elementalists something that is viable other than Celestial. Many people will probably find the Paladin amulet more fun to play with (and against) than Soldier or Sentinels amulet, so I applaud that change.
I reached Legendary last night. I felt it was harder to get through Ruby from Dec 15 to Dec 23 than it was to get through Diamond from Jan 4 to Jan 14, perhaps some of that was that they changed match-making a few days ago, and perhaps Ruby was harder earlier in the season because many of the better people were still working their way through it. I noticed a big change in average pip gain per 10 games after the match making changes.
(edited by Soul.9280)
Any idea how many people have already reached Legendary division?
Just wanted to say thank you to the devs for being active in the pvp forums today. Hopefully this is a trend that is sustained. I for one have a much higher tolerance for imperfection in a game if I feel like the developers of that game are listening when the player base is providing feedback.
Thank you!
Agreed!
Additionally it’s very pathetic that throwing matches to lower MMR to then progress in leagues is not only viable but likely the optimal approach. And Anet as usual refuses to dialogue with their player base.
The pvp leaderboards seem to reflect results that have occurred prior to the start of this pvp season. I see the top people with many hundreds of games played. I thought that the leaderboards would reset with the start of a new league.
I’ve also been crashing intermittently (1 to 3 times per hour) since Tuesdays patch, prior to that almost never. I’m running the 64-bit client on Windows 7 and I have an nVidia Titan graphics card with the latest drivers (from about a week ago).
This happens to me more frequently than I’d care to admit. Typically it occurs the day after I’ve had a good run of luck. I think they only reset MMR once per day (at night), so if you lose alot on a given day you don’t get credit for those losses (and get easier matches) until the next day. Anyone else see this type of behavior?
Could more people who care about this issue reply to this thread? Anet will have a harder time ignoring a thread with hundreds of replies…
I second this…
If they don’t do this, then those of us who need soloq and don’t get it, will likely move on to other things (hint: other games), as no one enjoys being farmed repeatedly by pre-made teams that have the advantages of good team comp, a strategy, and can also discuss tactics/deployment in real time over team speak…. versus PUGS who have none of those.
The past few days most matches have been the same, my team is a bunch of PUGs, the other is a premade. My team loses. It’s getting to the point where if you can’t bring your own team you might as well stop playing.
Please bring back soloq. If not, you’ll drive away those people who don’t have a team,
(edited by Soul.9280)
Client unable to connect to login server [merged]
in Account & Technical Support
Posted by: Soul.9280
I’m having the same problem as the OP. I can get to character select, but as soon as I try to load any of my characters the game pauses while loading and eventually error’s out with the message the OP reported. The game was working just fine prior to the new build…
Still looking for a team in NA for league play, also available for tournaments.