After hundreds and thousands of matches, it is very visible that your MMR system functions as this:
Ten players que at the same time
High MMRs are 10, low MMRs are 1
The ten players qued look like this: 10, 6, 5, 4, 4, 3, 3, 2, 2, 1
System tries to balance two teams from these MMRs
Result looks like this -> RED: 10, 4, 3, 2, 1 --- BLUE: 6, 5, 4, 3, 2
Looks good in theory but it’s bad in conquest application. Even if the elite 10 MMR on RED holds a point the entire match and farms opponents back to back without ever dying, it still won’t be enough point value to make up for the other two points where BLUE is slaughtering RED’s lower MMRs. Why are high MMRs punished for solo queing?
I normally run 5 man pre-mades within our guild “which kind of avoids the above problem” but when you are a team captain and run 5 man premades, sometimes you need a break to run solo or maybe duo while listening to mp3s for some casual play. But this is no longer a realistic option for those who have relatively higher MMR due to the visible MMR algorithm.
As priority you need to eliminate impractical and unreasonable match ups
It is one thing to lose 500 to 400 or 500 to even 250 but losing 500 to 100 or lower means that your team never had a chance in hell of winning to begin with and these sort of matches need to be eliminated from ever happening within ques.
Your MMR system is propagating all of the problems that players complain about:
4v5s ~ People leave when they realize a match is impossible to win
5 solos placed against 5 man pre-mades ~ No one has fun and people stop playing
r80s being placed with or against r1s ~ Why is this even allowed to happen?
ect.. ect.. the list goes on
The above problems are directly responsible for diminished player basis.
I am sure that you do all that you can already.
But please fix this calamity so that high MMR solos can enjoy the competitive scene.
man I agree with you
i run into this problem constantly
it’s just that there are not enough players for simple match making system to work
anet has came up with this obscure system to try and compensate
i think maybe they need to put less emphasizes on players with high mmr
it’s very hard to carry when the rest of your team is getting rekt
Getting matched as a solo vs a premade is a problem at all MMRs. I’m pretty sure I’m not very high but this constantly happens to me both ways and it’s not fun.
I know this isn’t an easy job ANet, but this seriously needs to stop happening.
And look at it from the other point of view. Low MMR’s are punished, too. This player with MMR 1 could be new and trying to get better, but he’s going to get destroyed by other players and gain no pvp experience. Do you think such player will want to stick around or just give up?
Before this new algorithm, the problem was having no middle-ground intermediate community. It was all elite MMRs going against each other and rabbits/deer having fun being placed on teams together. As soon as they hit around Dolyak, they would quickly become discouraged from being more and more frequently qued again Shark, Phoenix and Dragon play. They would receive a swift boot out from the world of Rabbits & Deer, only to find that there were no intermediate players playing, only older veteran players and this makes them ask themselves “why try? these guys have +3 years of experience on me” and that is exactly why there were no intermediate level ranks/MMRs out there.
Of course with this new algorithm, it rewards the intermediate ranks and MMRs as they are often placed together. More often than not you’ll see an enemy team that consists of 5 players with titles as such: Marauder, Reaper, Ransacker, Champion Slayer and they are usually able to easily beat the teams that have that 1x Legendary Champion paired with the 4x PvE titles.
Was this algorithm put in to place on purpose to help farm up a new and avid intermediate community? In the last month or so, I have seen more intermediate level players than ever before due to how rewarding it now is to be intermediate within the community. Even the leaderboards now simply reward grind time rather than skill and that is something any intermediate level player can take part in. Do not get me wrong, this is “in a certain fashion” exactly what the community needed but the method used was more of relief than it was a cure and has delivered salty repercussions within the higher MMR community and even the low MMR beginners.
I don’t feel that this thread is exactly the right place to list 101 suggestions so we’ll avoid that but the point being is:
Even players from great teams enjoy solo/duo roaming smash & bash from time to time and it would excellent if solo/duo ques could once again be made practical.
Its all levels of players that are being penalized not just high MMR players. They are probably prone to being a little more screwed then other levels when solo but its there for everyone. All it takes is one really bad player in the match (compared to everyone else) to ruin it. They need to tighten up the allowable range of MMRs in a match and if it hurts queue times so be it. You could have a match like this (assume 5 is average) 6,5,5,4,1 verus 5,4,4,4,4. Nobody high MMR but imo the second group is going to win because they don’t have the 1.
I completly agree with you, but this can hardly be done since it’s a random system overall, there will always be subjective issues. And actually, i don’t believe that would solve the problem, since it’s most likely builds rather than players skill..
Give 5 #100 players 2 mesmers 2 necro 1 thief, and put them against 5 #400 players with engis eles and warriors, we can predict the result of 3 out of 5 matchs probably…
It’s the same issue with pugs vs premades : most of the times i did complain about it, it was because we had to deal with OP bunkering setup on legacy of foefire, having 5 squishy armors..
Using your above player-skill distribution, there really isn’t any way to create a “fair” matchup. If you balance the MMRs on each team, then we run into the problem you described. But if we give the high MMR player more of the better teammates, then it becomes unfair for the other team.
I think the only way to fix this would be to reduce the maximum MMR variance allowed in games, such that a “10” would never be in the same match as a “1,” but would only queue with 7+’s. But that leads to longer queue times unless we increase the PvP population.
I don’t know about anyone else, but in the old queue I waited about 6 minutes, now I wait about 5.
I’d rather wait another minute or two for a better chance at a good match, then spend all this time to get rekt by a premade with new players on my team.
The current system is more likely to produce a result like this…
Red: 10, 5, 4, 3, 2
Blue: 6, 4, 3, 2, 1
Which could be argued is less fair, but really because MMR isn’t the only factor this is an oversimplification.
I think the bottom line though is that since players aren’t being compensated for high MMR then it is undisputed that the MMR system serves to hurt them. Particularly because if MMR was eliminated as a factor then players with high MMR would theoretically do significantly better with respect to points gained on the leaderboard (which is relevant for players who want to win the prizes).
That being said, unless there is compensation for being a high MMR it is just a liability which leads to:
1) more difficult matches
2) more risk towards losing more points
3) more harm for when someone on your team dc’s (since low MMR players typically don’t lose as many points per loss as high MMR players)
4) less potential gain of points per win since theoretically less teams will be above you, and even less -significantly- above you to warrant increased rewards points for the win.
So MMR only comes with negative factors unless you consider the match-making to be the benefit. In which case we have custom arenas and we can scrimmage teams we find equally skilled there. I don’t think anyone has been going into ranked Q thinking “oh boy, can’t wait to find a random team evenly matched with me.” While in theory it would be nice, in practice it turns into either:
1) get stomped by a team like abjured that’s been waiting in Q 15 minutes
2) wait 15 minutes and stomp randoms
Of course just focusing on MMR is an over-simplification but the reality of it being a factor is still relevant and doesn’t diminish the points being made.
My recommendation is to either:
1) Eliminate MMR as a match-making factor
OR
2) Reward players and incentivize them to attain a high MMR
I just joined my first competitive team this week, and we’ve been testing comps and learning each others playstyles in ranked. Don’t quote me on this, but this week I believe we’ve only face 2 premades running the same guild tag, a few premade pugs (lol), and the rest have been pug groups.
I actually just posted on the forums looking for teams to scrim, because with such inconsistent matchmaking, we have no way to judge or skill level and improve. We don’t want to fight pugs. We’re queued kitten men on voice com, and we would like to fight another 5 man on voice com without having to spend hours on the forums :’c
(Please don’t shoot meh down, Justin. Don’t tell everybody how noobish our team is and how perfect your algorithm is.)
I just joined my first competitive team this week, and we’ve been testing comps and learning each others playstyles in ranked. Don’t quote me on this, but this week I believe we’ve only face 2 premades running the same guild tag, a few premade pugs (lol), and the rest have been pug groups.
I actually just posted on the forums looking for teams to scrim, because with such inconsistent matchmaking, we have no way to judge or skill level and improve. We don’t want to fight pugs. We’re queued kitten men on voice com, and we would like to fight another 5 man on voice com without having to spend hours on the forums :’c
(Please don’t shoot meh down, Justin. Don’t tell everybody how noobish our team is and how perfect your algorithm is.)
Sorry, but I’ve got to shot yeh down, Boggs. Since we can’t detect voice coms, I can’t include it as a factor in matchmaking.
We do have some knobs that we can turn to increase competition for teams, but we have to be careful because a premade team does not always imply more skillful play.
Server Programmer (sPvP)
Isle of Janthir: Flux, Latch, Aegir
This is where you re-introduce solo queue and everyone likes you again.
- Solo ranked queue
- Team ranked queue
- Anything goes unranked queue
Thank you.
/agreed.
Why Josh and Evan are being stubborn about this, after so long and so many complaints, is beyond me.
I know they are not real PvPers, nor should they have to be in order to do a good job concerning PvP. That being said, they aren’t doing a good job and not because they aren’t real PvPers but because they are being stubborn and trying to re-invent the wheel.
There’s a reason why other games – Football/Soccer, Basketall, Rugby, [insert any team sports game here] – do not allow for random teams to fight organized teams (aka: “premades”)
Some very competitive games such as Counter-Strike, Magic the Gathering, [insert another games here]…and also Guild Wars 1 – do not allow for random teams to fight organized teams neither.
What if Josh and Evan could try to send their precious algorithm to FIFA/NBA/Wizards of the Coast? Because the algorithm is so good and FIFA might be loosing some awesome balanced games.
And also send it to whoever is in charge of GW1 since GW1 did/does have a split between random and premades (random arena and team arenas)?
/sarcasm on
Maybe GW1 had it wrong?
Maybe GW1 PvP split of RA and TA was horrible but the players just didn’t know it?
And maybe GW2 sPvP is the second coming of Jesus and the algorithm is the Holy Father, but the player’s just don’t know it neither?
I appreciate the response Justin and I did read through the link that you posted about how the algorithm actually works but one thing still confuses me:
The algorithm on paper would seem to work solos in as a great balancing factor when considering all of the variables that the algorithm is using. It would seem on paper that solos should produce more balanced match ups than premade vs. premade but I have personally experienced quite the opposite. About 7 out of 10 matches that I run solo, I land on a team of 4 players who very visibly have no clue what they are doing and we are placed against a team of intermediate level MMRs who are visibly superior combatants with a far greater understanding of how to play conquest properly. I have spoken with and frequently run with many older veteran players who are experiencing the exact same thing. I’ll save you from the needless details of specific matches but I hope that you don’t take me as a buffoon when I say that the information I have listed is no over-exaggeration.
The above experiences are why I assumed the algorithm worked in the way listed in my initial post “from raw match history it would seem to be true” but in contrast to your post and link posted, I have to ask:
Why are most of the intermediate skilled players on 50% to 55% win streaks?
Why are better players struggling between 45% to 50% lose streaks?
Are some accounts flagged for losses through the system or are we all just unlucky?
Keep in mind that I am referring to solo ques, not team ques.
Team ques seem rather balanced in my opinion.
I did not receive any response for the above post so I decided to document a single night of solo que and post something that was louder than the previous post. Please watch this video and support this thread if you are experiencing the same kind of problems with the current MMR algorithm
@ the OP well known facts you have stated , but its falling on deaf ears , the queues need to be split and the MMR also reworked , showing the chances of winning losing etc . As for your comment of josh not being a real pvper , i loled hes a good engi actually and can prolly e sport
I still don’t get why people want yolo queue back. It was horrible.
Sure, feeling like a god was fun for some time, but honestly, yolo queue generally came down to which team had the best player, as nobody even WANTED to cooperate and everyone just brawled off-point. It was just that bad.
The only crime, turrets committed, is being good against the celestial meta.
The mob has spoken and the turrets shall be burnt at the stake.
It’s not that I am opposed to the return of solo que but this thread is not about the return of solo que. It is about trying to understand why some players are funneled in to a strange outlier bracket of lose streaks where they are always placed on an inept team, each and every time they que.
I still don’t get why people want yolo queue back. It was horrible.
Sure, feeling like a god was fun for some time, but honestly, yolo queue generally came down to which team had the best player, as nobody even WANTED to cooperate and everyone just brawled off-point. It was just that bad.
If you where top 500ish you would get in to quite some quality games at peak hours. Most games you wouyld know pretty much everyone. The top players played non stop, and played soloq when their teams wherent playing.
In the current system those players arent playing at all if their teammates arent playing.
In the current system those players arent playing at all if their teammates arent playing.
Perhaps this is the case in ranked, but in order to avoid medium MMR 5 man teams, the players you refer to (the true top 250-500ish) will often be found soloing in unranked. My gameplay experience has been excellent in unranked at all hours of the day, with fair and exciting matches throughout.
I agree with the problems mentioned above, don’t get me wrong, and I don’t touch ranked other than in a team, to avoid skewed MMR. However, it would be unfair to say that those same players are not still around, having great games. Solo Q unranked is rewarded just the same as ranked (other than in leaderboard rewards, which many of those players don’t care about anyway due to tournament rewards!) and it’s going strong.
If I lose with two players on my team, matchmaking should take that into account and track those players or classes. If the same targets are seen in the same game; on the same team, matchmaking switches these targets regardless if that further imbalances the teams. Reason is synergy; matchmaking is bad at it.
A whole lot easier said than done. To be honest, I don’t even know where to start with the pseudo code. Can a Kalman filter be used?
Pseudo Code Questions
Can “Rank Points Earned” be modified to the following?
Ladder default=“0” min=“0” max=“1000000” leaderboard-points=“1”
Matrix odds=“0.0”
Score min=“0” points=“-1”
Score min=“200” points=“0.0”
Score min=“205” points=“0.5”
Score min=“210” points=“0.1”
Score min=“215” points=“0.15”
Score min=“220” points=“0.2”
Score min=“225” points=“0.25”
Score min=“230” points=“0.3”
Score min=“235” points=“0.35”
Score min=“240” points=“0.4”
Score min=“245” points=“0.45”
Score min=“250” points=“0.5”
//what’s a better pseudo code for this? lol.
Score min=“255” points=“0.55”
Score min=“260” points=“0.6”
Score min=“265” points=“0.65”
Score min=“270” points=“0.7”
Score min=“275” points=“0.75”
Score min=“280” points=“0.8”
Score min=“285” points=“0.85”
Score min=“290” points=“0.9”
Score min=“295” points=“0.95”
Score min=“300” points=“1”
Score min=“305” points=“1.05”
Score min=“310” points=“1.1”
//Etc. Spread of 5 up till 400
Score min=“400” points=“2”
//Etc. Spread of 5 up till 500
Score min=“500” points=“3”
Matrix
Matrix odds=“0.2”
Score min=“0” points=“-1”
Score min=“300” points=“0”
//Etc. Spread of 5 up till 400
Score min=“400” points=“1”
//Etc. Spread of 5 up till 500
Score min=“500” points=“2”
Matrix
Matrix odds=“0.4”
Score min=“0” points=“-1”
//Etc. Spread of 5 up till 400
Score min=“400” points=“0”
//Etc. Spread of 5 up till 500
Score min=“500” points=“1”
Matrix
Matrix odds=“0.6”
Score min=“0” points=“-2”
core min=“300” points=“-1”
//Etc. Spread of 5 up till 500
Score min=“500” points=“1”
Matrix
Matrix odds=“0.8”
Score min=“0” points=“-3”
//Etc Etc Etc
Reasoning
Since a player can contribute at minimum of 5 points to his team’s score, the above detailed point system would reflect better leaderboard positioning for players of similar ranked positions as well as similar win rates. Have the leaderboard order these players based on these hidden values but obviously show Rank Points as rounded whole numbers.
Can we gain a “mini buff” if a player is late to join the season.
So the players who look at the leaderboard for the first time won’t get discouraged when they see players with 200 games played? The buff can take the current top leaderboard player’s stats and compares it to yours. It then averages a % of games played, leading up to that current time or rank points. Example:
Season started 7d ago.
Current Ranked #1 Player’s average games per day = “7” (Player A)
Total of 49 games since season started (Player A’s total games)
Player B just started his first ranked match.
Math Stuffs
Player A’s average games per day in last 7 days / 100 * Player B’s Missed Days / 3 * Points Earned At End of Match + Points Earned At End of Match //his final rank points
7 / 100 * 7 / 3 * pointsAwarded2 + pointsAwarded
//Need a new variable pointsAwarded2 so you can keep original pointsAwarded data intact. Or something like that…
This particular player gets .16 extra point per each whole point gained. If he was going to gain 3 rank points, he would have gained 3.49 points instead.
2 rank points = 2.32 points earned
1 rank point = 1.16 points earned
Effect ends in 7 days time or 49 games played or until _Target Player A’s points has been reached. Which ever of the three conditions comes first
This is not meant to push an inactive player directly to the top of the leaderboard, but give that player a slight jump start on what he missed. Additional factors could be added to moderate the buff’s effect such as less points earned, duration, etc.
aka FalseLights Rank: Top 250 since Season 2 #5 best gerdien in wurld
I have now a winrate ~62% in the old ladder I had something like 52%.
I am now > top 1000 because I do not play 10 games per day. I was in the old solo ladder Top 300.
My better winrate is not beacuse I have more skill, but I got worse teammades and opponents. I do not care about the algorithm but do you really think that you are going into the right direction with your approach?
In my point of view the system does not work very well. Additional I had twice games like this: Premade 5 vs 1+1+1+1+1 at 11 p.m. and the premade teams were 55Hp + SALT and once oRNG with a queue time less < 1 min.
I have many really good games. But in the old solo ladder I had more and better games than now. It is not bad, but it is not great at the moment.
I did not receive any response for the above post so I decided to document a single night of solo que and post something that was louder than the previous post. Please watch this video and support this thread if you are experiencing the same kind of problems with the current MMR algorithm
just because someone isn’t wearing a pvp title doesn’t mean that they aren’t high-ranked. I hate the pvp titles since the rank “overhaul” because I think they’re cheap and mostly worthless. I wear my HoM title because I’m actually proud of that. I don’t even use a finisher. To the random onlooker, I could be a rabbit for all they know.
The thief got nuked out in the quarry 2v2 because he’s a zerker thief. You’re a condi ranger. Of course you’re going to take a toll against 2 thieves, you’re bunky as hell and your condis wreck their 1-2 condi removal abilities
This is where you re-introduce solo queue and everyone likes you again.
- Solo ranked queue
- Team ranked queue
- Anything goes unranked queue
Thank you.
They combined the two cause the queues for team queue were long. Hopefully it was worth it to them cause i’m sure people have stopped playing cause of this
How to identify rabbits, deer and daily AP farmers:
no titles
smell of the fields of queensdale
wearing worn chain from queensdale
don’t cap home point on initial split
rushes lord on initial split
zergs svanir
not using any build relative to pvp what so ever
How to identify players who are here for the competitive scene:
wearing pvp titles, most do
visible understanding of conquest rotations and combat
using pvp builds
very rarely ever makes gross errors in rotations or tactics
It is not difficult to identify the separation between competitive players and casual AP farming players. It becomes apparent in a match on the initial split and their first combat engagement.
Actually, my understanding and experience is significantly different:
Novice players in PvP:
-When confronted with obviously bad players, these players rage at them in map chat
-Post on forums about how they get randomly paired with bad players
-Pretend like they’re competitive when they don’t compete in tournaments on a regular basis
-Rant when players don’t rotate according to how they think players should be rotating
-Ostracize people who don’t copy and paste their build off of metabattle
Seasoned Players in PvP:
-Realize that people of varying skill levels yolo Q and don’t rage at them.
-Realize that raging at newer players is pointless and usually say nothing
-Recognize that yolo Q’ing runs the risk of being paired with baddies
-Does not post about the bad Q because it’s pointless
-Realizes that players do not need to run meta builds to gain experience PvP and improve
How to identify rabbits, deer and daily AP farmers:
no titles
smell of the fields of queensdale
wearing worn chain from queensdale
don’t cap home point on initial split
rushes lord on initial split
zergs svanir
not using any build relative to pvp what so ever
How to identify players who are here for the competitive scene:
wearing pvp titles, most do
visible understanding of conquest rotations and combat
using pvp builds
very rarely ever makes gross errors in rotations or tactics
It is not difficult to identify the separation between competitive players and casual AP farming players. It becomes apparent in a match on the initial split and their first combat engagement.
I prefer using a PvE title and then pairing it with the dragon finisher. People don’t know what to expect. Much fun.
Sorry, but I’ve got to shot yeh down, Boggs. Since we can’t detect voice coms, I can’t include it as a factor in matchmaking.
We do have some knobs that we can turn to increase competition for teams, but we have to be careful because a premade team does not always imply more skillful play.
If you “can’t detect voice coms”, that’s understandable. However, it’s one of the most influential traits in determining the outcome of a game.
I’m sure you can provide anecdotal stories of “this one time in band camp…” where a bunch of new/novice players suddenly had lots of friends in this game…they were all online at the same time…formed a premade…all of them had TS…and lost to a bunch of soloqers with no voice com.
A team with voice com CAN occasionally lose to a team without voice com. (However, it would ironically indicate MORE negative things about your matchmaking algorithm if/when it does occur.)
The fact that I’m having to explain the value of voice com to developers of an MMO makes me feel like I’m in the twilight zone…who would argue this?
Yet, you guys at Anet must disagree.
Otherwise you would NEVER go through all the trouble to change from a 100% logical system where you had a separate team que and solo que (plus the OPTION of a solo player joining team que).
You ALREADY HAD the best option for dealing with the limitation of not detecting voice com in a matchmaking algorithm.
Who on earth convinced you guys that the current model makes more sense than previous model? That person should be in sales…not game design.
Actually, my understanding and experience is significantly different:
Novice players in PvP:
-When confronted with obviously bad players, these players rage at them in map chat
-Post on forums about how they get randomly paired with bad players
-Pretend like they’re competitive when they don’t compete in tournaments on a regular basis
-Rant when players don’t rotate according to how they think players should be rotating
-Ostracize people who don’t copy and paste their build off of metabattle
Seasoned Players in PvP:
-Realize that people of varying skill levels yolo Q and don’t rage at them.
-Realize that raging at newer players is pointless and usually say nothing
-Recognize that yolo Q’ing runs the risk of being paired with baddies
-Does not post about the bad Q because it’s pointless
-Realizes that players do not need to run meta builds to gain experience PvP and improve
Here is Justin (ODell) , the wizard who programs your bad loot & and matchmaking. He can also cast a curse on you which does +400 ping.
Hey now… you can blame me for matchmaking, but I had nothing to do with loot.
Maybe you could have a chat with the person that has something to do with loot and learn them something, b/c they don’t really seem to understand it very well.
The current system is more likely to produce a result like this…
Red: 10, 5, 4, 3, 2
Blue: 6, 4, 3, 2, 1
Which could be argued is less fair, but really because MMR isn’t the only factor this is an oversimplification.
so how does this work when solo q goes against a premade? if the premade is 10, 8, 8, 8, 7, its been my experience the solo q is not full 10s…. ive seen some pretty obvious lets call them newcomers on the solo q side and it is most definately not balanced at all
the real problem is that there is no real pvp community in this game – look at even the difference in score in the top 25 players on the ladder
the pvp community is non existent which means there isnt the numbers to pair you with same skill people – this makes it frustrating for high mmr players cause teams are full of learners
maybe its time to go back to dota where there is actually a decent size player base of pvpers
No one is asking to be paired with the same skill level of players or even against the same skill level of players. What is frustrating is what Justin said himself, in his initial response to my post:
The current system is more likely to produce a result like this… Red: 10, 5, 4, 3, 2 Blue: 6, 4, 3, 2, 1
Being placed on the team (B) “Blue”, most of the time you que, is a problem. Why does the algorithm allow for such clearly lopsided match-making to begin with?
It doesn’t really matter what MMRs are in play. What matters is that the algorithm understands how to balance the MMRs in play “considering the conquest dynamic” and eliminates players from becoming strange outlier MMRs that are funneled in to team (B) each and every time they que.
match making has gotten better in the last week or so. I still get streaks of "kitten" matches where I have no way of winning or get put in 4 v 5s but it’s happening a lot less. Maybe because I play a lot less matches now days , it does seem to work better.
It’s funny to see my leaderbord position move up some days when I haven’t played a single match at all lol. I just think how kittened the people who lost pts after playing many matches must be.
How to identify rabbits, deer and daily AP farmers:
no titles
smell of the fields of queensdale
wearing worn chain from queensdale
don’t cap home point on initial split
rushes lord on initial split
zergs svanir
not using any build relative to pvp what so ever
How to identify players who are here for the competitive scene:
wearing pvp titles, most do
visible understanding of conquest rotations and combat
using pvp builds
very rarely ever makes gross errors in rotations or tactics
It is not difficult to identify the separation between competitive players and casual AP farming players. It becomes apparent in a match on the initial split and their first combat engagement.
Actually, my understanding and experience is significantly different:
Novice players in PvP:
-When confronted with obviously bad players, these players rage at them in map chat
-Post on forums about how they get randomly paired with bad players
-Pretend like they’re competitive when they don’t compete in tournaments on a regular basis
-Rant when players don’t rotate according to how they think players should be rotating
-Ostracize people who don’t copy and paste their build off of metabattle
Seasoned Players in PvP:
-Realize that people of varying skill levels yolo Q and don’t rage at them.
-Realize that raging at newer players is pointless and usually say nothing
-Recognize that yolo Q’ing runs the risk of being paired with baddies
-Does not post about the bad Q because it’s pointless
-Realizes that players do not need to run meta builds to gain experience PvP and improve
LOL, the irony !!!
“Raids are like fortune cookies. You eat the cookie and then read the paper scraps.”
match making has gotten better in the last week or so. I still get streaks of kitten" matches where I have no way of winning or get put in 4 v 5s but it’s happening a lot less. Maybe because I play a lot less matches now days , it does seem to work better.
That was exactly my opinion, until this afternoon… I played 5 or 6 matchs :
- 1st was 3 + 1 + 1 : we get beginner engineer suiciding twice on far, then rage quitting
- 2nd and 3rd were against the same premade : not much to say concerning these
- one was 2 + 1 +1 +1 : beginner necro made us lose a ( probably) won match
- last one i was alone, we’re 4v5 after a few seconds, we can still win but we end up losing because we got no bunker against 2 eles + sword war + engi
Before this new algorithm, the problem was having no middle-ground intermediate community. It was all elite MMRs going against each other and rabbits/deer having fun being placed on teams together. As soon as they hit around Dolyak, they would quickly become discouraged from being more and more frequently qued again Shark, Phoenix and Dragon play. They would receive a swift boot out from the world of Rabbits & Deer, only to find that there were no intermediate players playing, only older veteran players and this makes them ask themselves “why try? these guys have +3 years of experience on me” and that is exactly why there were no intermediate level ranks/MMRs out there.
I’m not even sure it’s doing a great job of fixing this issue. Maybe it’s just me? I’m not precisely new, but I’m not very good. (To everyone who’s had me on their team: I’m terribly, terribly sorry.*) The thing is, I was doing pretty good in unranked! I wanted to improve, and I wanted more challenge, so I started joining ranked… and promptly got steamrolled. Repeatedly. Like I said, maybe it’s just me being a crummy player and I need more time messing around in unranked, but I feel like the existing system hasn’t done a very good job of handling the transition from beginner to intermediate.
That said, I completely agree that the situation high ranked players are dealing with is unfair. I think all of these issues with the matchmaking system stem from a combination of a) a low population of PvPers and b) a desire to keep queue times down. Personally, I’d rather wait longer if it meant fairer matches, though I think that in the end, the population is the bigger problem.
This is where you re-introduce solo queue and everyone likes you again.
- Solo ranked queue
- Team ranked queue
- Anything goes unranked queue
Thank you.
No.
Split solo and premade queues for unranked as well. Just because it’s unranked doesn’t mean there is any excuse for ruining the experience of the teams involved with rigged matches. Unranked is a place to learn and to get away from the stress of ranked – but this doesn’t exist when one can solo queue and get (FREQUENTLY) paired against 3-5 man premades or, conversely, premake and get thrown against a team that can’t even provide you with a coordinated challenge.
Actually having high mmr does punish people because matchmaking will take average MMR into consideration instead of individual. Basically if average MMR is 5, and you have people of 9s, 8s, 7s, 6,s etc, matchmaking will put lowest MMR players and highest MMR players against mid to average MMR players, forcing those players with high MMR to carry players of low mmr.
And since gw2 is conquest, which some people just don’t get, most players with low mmr do not appreciate the role of nodeholder/defender and likely to chase people off the node, get decapped, and die over and over. And there’s only so much players with high mmr can do. Even they cannot win 2v4, and if they do, the players with low mmr will go random places to die.
This results in very uneven match up for high level players and low level players do to sheer skill difference. From my experience I either steamroll the opposition or get steamrolled myself. Basically people need to know their mmr, and should get matched with people in their MMR zone rather than get balanced out by the matchmaking system.
