Guild Wars 2

9/30/2015

Decided to run solo tonight for the sake of testing the fix to the MMR algorithm. Eight matches were ran, beginning around 7:30 pm (eastern) and ended around 2:30 am (eastern). Each match was ran in ranked on my main which is a ranger. Here are the results with short commentary:

Match 1
The que took about 5 minutes
My team losses 343 to 500
The match felt balanced in terms of combat and the understanding of conquest on both teams. My team simply had a poor team composition. There were too many squishy stealthers and we could not hold points.

Match 2
The que took about 3 minutes
My team wins 500 to 40
_There were no DCs, no AFK’ers, no leavers, no ragers, the enemy team tried hard the entire match. Our team was just stacked with strong players and their team was stacked with novice players at the best. We were all solo ques on each side of the match so something was still very wrong here with why the algorithm would do this.

Match 3
The que took about 3 minutes
My team losses 224 to 500
We were a team of three solos that joined some duo que team. The opposing team was a five man premade that was trying hard. My team members were losing all 1v1 engagements and eventually losing every team fight, in the end. This leads me to believe that the MMRs of my solo and duo teammates were plainly lower than that of the five man premade team.

Match 4
The que took about 5 minutes
My team wins 551 to 278
Combat was fairly balanced between both teams but the enemy team had no one playing peel. We had their home point neutralized the entire match.

Match 5
The que took about 5 minutes
My team wins 500 to 330
The match started out terrible, we had bunkers trying to push far and peelers trying to bunk mid. Though it was a rare solo group that actually listened to good advice in the /t chat. They cleaned up their positions and we came back for the win. The match was balanced in terms of combat. We won because of mid-match communication similar to a TS3 team.

Match 6
The que took about 5 minutes
My team losses 350 to 500
Mid was the only place that any of my teammates wanted to go the entire match. I actually suspect that two of them were bot accounts following a real player as they did the same exact motions the entire match and never left the particular player’s side. Regardless, it amounted down to a match where we had three players sitting at mid when the enemy team was clearly ignoring mid due to this and playing sides for the two cap win.

Match 7
The que took about 30 seconds
My team losses 147 to 501
Similar to the listed Match 2. My teammates were visibly much less experienced than the team we were against in terms of combat and understanding of conquest. Again, why would the server not just swap two of our players for two of their players? This wasn’t as bad of a stomp as the listed Match 2 but it was still an unreasonable match that shouldn’t be happening.

Match 8
The que took about 3 minutes
My team wins 501 to 356
The enemy team was on par with us as far as combat as concerned but they had bunkers pushing around for decaps an peelers trying to bunk their mid and their home. The enemy lost purely due to bad player positioning.

Overall the match making felt better tonight. It was the first time in a very long time that I had: “win, loss, win, loss” rather than: “win, win win, loss, loss, loss”. This is a good thing. I captured a 50/50 win/lose ratio without seeing too many problems in the match making thus far. Moreover I felt that ALL my team members tonight were pulling their weight. I wasn’t experiencing the 10/3/3/2/1 effect to balance against 4/4/4/4/4. Match 7 felt a bit off but nothing like I have experienced in the past.

I do have to comment further on Match 2 however. Something was very wrong with that match up. The enemy team was absolutely clueless to conquest, combat and even their builds in general, while my team was stacked with veterans. I’m not sure what exactly is allowing matches like that to happen but it needs to be looked in to.

I’d also like to add that this is the first night in a very very long time that I did not see a single DC, AFK’er, Leaver or Rager.

~ Seems like a good patch so far. I’ll give more feedback when I can.

It’s a lot better than it was but it’s still bothersome when the game puts you against something you just can’t beat, such as 2+ necro teams. All they have to do to win is cap 2 points and crap out wells. They win by attrition, the time it takes to recap those points (assuming the necro even dies) still means they win by scoring points for their team.

Stupid tanky ground control builds aside, I agree. The matchmaking is a lot more even overall. Didn’t have a lot of big comebacks yet (most games were decided on ~300 points) but at least there’s a struggle in what feels like 7/10 games that don’t have a stupid ground-nuke wombo combo.

So now the imbalance comes more from profession composition moreso than MMR. They should do something to fix that as well.

I agree on that point actually and I noticed it last night as well. The matches now feel less decided by “who is a good or bad player” and more by which comp is capable of handling the other comp. Don’t get me wrong, with a perfect algorithm where 5/5/5/5/5 goes against 5/5/5/5/5, a match SHOULD be decided mostly on team composition and how that comp plays it’s positions. This is certainly the direction we want.

The next step would be for the algorithm to better identify current meta and the class counters that exist. Example: LB/GS Power Rangers usually counter Necromancer builds so there is never a reason for two solo que teams to roll where one team has two Rangers and the other has two Necromancers. The potential match altering weight of the Rangers vs. the Necromancers as a match circumstance where the Necromancers are being hard countered, is much greater than worrying about balancing MMRs, considering these MMRs were somewhat equal. Let’s say the Rangers are MMRs of 3 & 4, then the Necromancers are both 5 MMRs “1 is low, 10 is high”. Even though one team has the lower MMRs of 3 and 4, both are Rangers that can easily counter the Necromancers and that circumstance in strength outweighs the circumstance of having a slightly lower MMR team. If you see what I mean by this..

There are two other aspects that the algorithm needs to be strengthened on:

• Identifying how conquest is actually played and won so that it helps form teams based on this identification.
• Understanding that lower MMRs in general are worth more match altering weight than higher MMRs.

My final post in this thread elaborates on this: https://forum-en.gw2archive.eu/forum/game/pvp/High-MMR-is-punished-for-solo-que/page/3#post5543352

~ Please do not bump the above link.

On main class (not certain if that matters) did ok 3 out of 4 games since matchmaking update. Lost game still came out close. Last game before update was a complete mismatch.

Erindriel.2351:

It’s a lot better than it was but it’s still bothersome when the game puts you against something you just can’t beat, such as 2+ necro teams. All they have to do to win is cap 2 points and crap out wells. They win by attrition, the time it takes to recap those points (assuming the necro even dies) still means they win by scoring points for their team.

I agree, necro is too OP.

Trevor Boyer.6524:

The next step would be for the algorithm to better identify current meta and the class counters that exist.

There’s no point in the matchmaking getting involved with team compositions as long as we’re allowed to swap toons during matches.

Erindriel.2351:

It’s a lot better than it was but it’s still bothersome when the game puts you against something you just can’t beat, such as 2+ necro teams. All they have to do to win is cap 2 points and crap out wells. They win by attrition, the time it takes to recap those points (assuming the necro even dies) still means they win by scoring points for their team.

Hey Hey.. That’s clearly a valid strategy

But ya. Match Making seems better. I only played one match last night that was a blow out. 500 to 106 in favor of my team. Although the people on the other team were not bad they just had a really really bad comp.

MMR is still account wide = really bad.

Clearly shows that the devs dont take thier own game serious. They know its really shallow or else they wouldve made splitting MMR a priority.

The only reason u wouldnt is if all classes are so shallow that u can learn them within a day.

Which is how the devs see this game sadly enough :>

Locuz.2651:

MMR is still account wide = really bad.

Clearly shows that the devs dont take thier own game serious. They know its really shallow or else they wouldve made splitting MMR a priority.

The only reason u wouldnt is if all classes are so shallow that u can learn them within a day.

Which is how the devs see this game sadly enough :>

And if we had character based MMR, how would they stop people from switching class mid-game? About 9ish months ago, people automatically lost if they switched classes and everyone moaned about not being able to switch if they had an unbalanced team comp.

Belial.9350:

And if we had character based MMR, how would they stop people from switching class mid-game? About 9ish months ago, people automatically lost if they switched classes and everyone moaned about not being able to switch if they had an unbalanced team comp.

You shouldnt be able to. In a game with depth there should be a night and day difference between your main class and classes that u rarely play.

Especially if you consider that team and soloq are merged. Its clear that arenanet doesnt care one bit about soloq ers and all their QoL issues like mid game class switching.

Locuz.2651:

Belial.9350:

And if we had character based MMR, how would they stop people from switching class mid-game? About 9ish months ago, people automatically lost if they switched classes and everyone moaned about not being able to switch if they had an unbalanced team comp.

You shouldnt be able to. In a game with depth there should be a night and day difference between your main class and classes that u rarely play.

Especially if you consider that team and soloq are merged. Its clear that arenanet doesnt care one bit about soloq ers and all their QoL issues like mid game class switching.

Did you read the second half of my post? They used to give you automatic loss if you switched character mid-game but changed it because everyone moaned about it. The reason why nothing gets done is because people moan about every single change.

Belial.9350:

Locuz.2651:

Belial.9350:

And if we had character based MMR, how would they stop people from switching class mid-game? About 9ish months ago, people automatically lost if they switched classes and everyone moaned about not being able to switch if they had an unbalanced team comp.

You shouldnt be able to. In a game with depth there should be a night and day difference between your main class and classes that u rarely play.

Especially if you consider that team and soloq are merged. Its clear that arenanet doesnt care one bit about soloq ers and all their QoL issues like mid game class switching.

Did you read the second half of my post? They used to give you automatic loss if you switched character mid-game but changed it because everyone moaned about it. The reason why nothing gets done is because people moan about every single change.

Not hard to deal with. Be able to switch during the 1.5 min “prep time,” and you get automatic loss when you logout/change characters when the match proper starts.

Reikou.7068:

Belial.9350:

Locuz.2651:

Belial.9350:

And if we had character based MMR, how would they stop people from switching class mid-game? About 9ish months ago, people automatically lost if they switched classes and everyone moaned about not being able to switch if they had an unbalanced team comp.

You shouldnt be able to. In a game with depth there should be a night and day difference between your main class and classes that u rarely play.

Especially if you consider that team and soloq are merged. Its clear that arenanet doesnt care one bit about soloq ers and all their QoL issues like mid game class switching.

Did you read the second half of my post? They used to give you automatic loss if you switched character mid-game but changed it because everyone moaned about it. The reason why nothing gets done is because people moan about every single change.

Not hard to deal with. Be able to switch during the 1.5 min “prep time,” and you get automatic loss when you logout/change characters when the match proper starts.

But then this reverts to the original problem where you could just constantly queue with a low mmr character and switch to a high mmr character during the prep time.

You should track party compositions.

I’m really sure most of the times party compositions has a huge impact on the final result, more than little differences in MMR.

I queued solo and your team’s comp will be at the discretion of your teammates since there are no penalties for swapping. The ones that swap out classes to improve comp are the best teammates you encounter queuing solo

10/2/2015

Decided to run Stronghold beta in Unranked tonight as a solo que on my Ranger main. Again, for the sake of testing the MMR algorithm as well as giving Stronghold feedback. I have not pre-purchased HoT so I was using the general LB/GS serk/marauder. I started around 7:30 pm (eastern) and ended around 10:20 pm (eastern).

Match 1
Que took about 8 minutes
My team losses 500 to 240ish “couldn’t remember exact score”
Our team just refused to defend, ever.

Match 2
Que took about 6 minutes
My team losses 550 to 30
It was a bad match. Our players were visibly novice level at beast, they tried hard but couldn’t win combats against the team of stacked veteran players. Again, why is the server still tossing blowout matches where one team will surely lose?

Match 3
Que took about 4 minutes
My team wins 490 to 160
Very balanced match actually. The enemy team made one critical error mid-match and overcommitted to trying to kill me in a 2v1 near their gate. This gave my team relief at the supply, they pushed a ton of bombers, opened last gate and we killed the lord fairly easily.

Match 4
Que took about 5 and a half minutes
My team losses 520 to 215
The match was actually balanced. The reason we lost is because there was an enemy Flame Thrower Engineer who was hiding behind his gate and abusing an exploit where the flame thrower was firing through the gate and downing Bombers before they could do anything. This needs to be patched immediately it is a gross error in the Stronghold programming.

Match 5
Que took about 4 and a half minutes
My team wins 600 to 240
Match felt balanced in terms of combat. I think the enemy team just had players who were not used to playing Stronghold.

Match 6
Que took about 4 minutes
My team wins 635 to 295
Again, felt fairly balanced. My team just knew more about Stronghold than the other team.

Had another 50/50 win/loss night. Seemed good other than Match 2 being all wrong. It was another instance where it was very visible that one team was stacked with very low MMRs and the other team was stacked with much stronger MMRs.

I would also like to add that Rangers that are not Druids are having serious issues keeping up with the HoT specializations. Normally I’m one of the people that always promotes the idea that “Rangers are stronger than you think they are” but this has finally reached a point where Ranger as a class “not spec to spec” simply will not be viable in low, mid or higher MMRs when HoT releases.

Thanks for sharing!
Queue time still too long though.

Can someone tell me this new site where I can review and post my match history?
I’ve had a lose streak tonight where every match was a complete lopsided blow out again and I’d like to review it.

My teams have been stacked with players who go and get absolutely churned in every combat they engage in all night long. Why is this bullkitten still happening? A very good question that will probably never really be answered.

The problem is the fact that MMR follows a character. If you create a new class right now you will be stuck in ELO hell as they call it with incompetent players.

Trevor Boyer.6524:

Can someone tell me this new site where I can review and post my match history?
I’ve had a lose streak tonight where every match was a complete lopsided blow out again and I’d like to review it.

My teams have been stacked with players who go and get absolutely churned in every combat they engage in all night long. Why is this bullkitten still happening? A very good question that will probably never really be answered.

http://gw2efficiency.com/account/pvp

get the api from your gw2 account page

Inflames.8026:

The problem is the fact that MMR follows a character. If you create a new class right now you will be stuck in ELO hell as they call it with incompetent players.

All these wrong informations…
Yes, every class gets calculated a own mmr, but it is not used for matchmaking. Devs decided the algorithm will use the account mmr as long as you are alowed to switch classes after matchmaking during preperation time. So this statement is wrong.

I dunno, had around 7-8 loss streak yesterday. The people i played with is just… weirdest builds, getting farmed in every team fight or 2v2 even when having initial strong advantage, even playing something like d/d ele, farming creatures 24/7 on forest, leaving thief to cap/defend a point etc., it’s just mind boggling. Not a single match where i felt competent people were fighting each other, every match is basically a struggle to work around kittens, around none stop idiocy, obviously no use. Rage inducing experience at it’s best, so i don’t know what they fixed there. I have no desire to play pvp after that at all, i value my nerves and time too much. Perhaps it’s a weekend staff, all the bright ones came out to play, but still, no thank you, one of the worst ways to spend your free time.

I’m lucky if I queue into a game before 10 minutes. And when I do finally get into a match sometimes it is with players who confess it is their 3rd time in a match.

So basically the system waits so long to find players with high mmr for me, then it gives up and gives me baddies? I don’t know, doesn’t really make sense to me.

I see premades still winning everything.

This matchmaking suck so much It’s a miracle to win a single kittening match, a miracle

Its the exact way around for me. I win 8 out of 10 matches since patch. Started playing different classes i am not good at now. Its a joke if you win 1v3s as a thief. And thats surely not because i improved the last few days or there are no opponents for me in this game. Its just the matchmaking that gives to easy matches. Hope it will adjust soon.

I never lose matches. I just win so hard that I break the score and come out last.

It is normal to have series in both directions, there is nothing to do about it. Just give matchmaking a chance. Over all matches most ppl are around the 50% winning rate.
Edit: There we go, a few matches later everything is back to normal

Gonna post it not to be annoying but for the sake of tracking. First two days felt good and balanced but all of a sudden last night it felt broken again.

I’d also like to add that when this happens, I never get wins to balance it back out. I have 48% win rate in ranked and unranked but 56% win rate in custom arenas. Total account wins are at 50%. This makes me need to ask the question: When the match making aims for 50% wins, does it read all my wins accrued concerning all current and previous match types including hotjoins and custom servers? If so, this would explain the lose streaks. Is the algorithm trying to keep my account win rate at 50% and I am being punished in ranked nowadays due to an old lingering good win rate from hotjoin/custom arenas?

I would appreciate feedback on the above question. If this is true, it would probably widely explain the lose streaks for all of us who have been experiencing it for quite some time.

10/4/2015
Tonight was a mix of solo, duo, trio, quad and five man. I ran 50/50 tonight in 10 matches. No reason to post the pic for a simple 50/50 log.

Recent matches on main account (Sins.4782) have been absolutely terrible. Like, people trying to solo beast and dying terrible.

Evan Lesh.3295:

Trevor Boyer.6524:

I’d also like to add that when this happens, I never get wins to balance it back out. I have 48% win rate in ranked and unranked but 56% win rate in custom arenas. Total account wins are at 50%. This makes me need to ask the question: When the match making aims for 50% wins, does it read all my wins accrued concerning all current and previous match types including hotjoins and custom servers? If so, this would explain the lose streaks. Is the algorithm trying to keep my account win rate at 50% and I am being punished in ranked nowadays due to an old lingering good win rate from hotjoin/custom arenas?

Matchmaking doesn’t look at existing win rates, it only lowers your matchmaking rating if you lose and raises it if you win so that over time you should balance to a 50/50 win rate. Custom arenas, unranked, and ranked all have no influence on each other.

So tank an unranked match, then play a ranked match for better odds at winning a “meaningful” match?

Evan Lesh.3295:

Trevor Boyer.6524:

I’d also like to add that when this happens, I never get wins to balance it back out. I have 48% win rate in ranked and unranked but 56% win rate in custom arenas. Total account wins are at 50%. This makes me need to ask the question: When the match making aims for 50% wins, does it read all my wins accrued concerning all current and previous match types including hotjoins and custom servers? If so, this would explain the lose streaks. Is the algorithm trying to keep my account win rate at 50% and I am being punished in ranked nowadays due to an old lingering good win rate from hotjoin/custom arenas?

Matchmaking doesn’t look at existing win rates, it only lowers your matchmaking rating if you lose and raises it if you win so that over time you should balance to a 50/50 win rate. Custom arenas, unranked, and ranked all have no influence on each other.

Sadly, it also seems to match you with the same people too when you lose. I know my MMR isn’t that high since I rarely pvp and just started doing it with my dailies, but I got stuck in two matches in a row with the same rangers who basically did nothing but autoattack the whole match, so double loss for me. I had loss earlier, and basically just took a break hoping they would stop playing too

10/5/2015

Greatly appreciate the response, Evan.
~ Thank you

Played 8 ranked matches tonight. Another mix of solo/duo/trio/quad/five man. Won 5 and lost 3. The last loss was due to a DC though.

Ross Biddle.2367:

Evan Lesh.3295:

Trevor Boyer.6524:

I’d also like to add that when this happens, I never get wins to balance it back out. I have 48% win rate in ranked and unranked but 56% win rate in custom arenas. Total account wins are at 50%. This makes me need to ask the question: When the match making aims for 50% wins, does it read all my wins accrued concerning all current and previous match types including hotjoins and custom servers? If so, this would explain the lose streaks. Is the algorithm trying to keep my account win rate at 50% and I am being punished in ranked nowadays due to an old lingering good win rate from hotjoin/custom arenas?

Matchmaking doesn’t look at existing win rates, it only lowers your matchmaking rating if you lose and raises it if you win so that over time you should balance to a 50/50 win rate. Custom arenas, unranked, and ranked all have no influence on each other.

So tank an unranked match, then play a ranked match for better odds at winning a “meaningful” match?

How do you get that from what he said? He said they have no influence on each other. They’re separate mmr. Your win rate in one has no affect on the other.

Ran five man premade almost all day. It was some of the best match making I’ve ever experienced. Every match came down to 500 to 480ish or closer. No one was even paying attention to win rates because we were having so much fun. Maybe we just got lucky with our ques today but either way, this is exactly the type of match making that makes you want to play the game.

People don’t get frustrated at losing matches. They get frustrated at being placed in bad team setups vs. a superior team setup. As long as a match is balanced, people have fun despite if they are winning or losing.

Not a good patch to me. Still broken

Experience from today:

When I play with the classes I have most games on I regularly get people that are obviously completely new to pvp. I don’t get why I get paired with lvl 2 new players when I have 800 games (not that much I know).

On the other hand when I play my 3 game with ingi, just strted him, I have premade enemy teams with phenix or dragon rank.

I’m just confused if the MM is currently active or not.
I don’t have a problem with my win rate.. that is ~60% although I play mostly solo. I just don’t get the different people that often get paired together.

How has match making gotten so bad. I honestly just witnessed one of the worst games I’ve ever played…. and we were the victors. We had to explain the to point of conquest to the opposing team. Not one of them played pvp before.

Anet, we need to talk about this relationship we have. I like to soloq, and you loooovveee to dump on soloq players. I can’t help but ask everytime I soloq if I’ve got Stockholm syndrome.

PVP improvement suggestion:
Please keep completely dumping on soloq players and make the matches as terrible as possible until I queue with players so I can dump on other soloq players. I feel for you guys and its high time you take the stick to the tiny PvP community and relentlessly beat us into playing it your way.

The 501-53 win on Kyhlo was against 4 players as the 5th even load in.

Cheers,
One of your players, fans and fecalpheliac. Just keep on dumping on me Anet.

Appreciate the collaborative effort of sharing match results.

MMR update and BWE
all solo, 65% win rate

Pre MMR update
all solo, 48-50% win rate

Post BWE 3
all solo, 25% win rate

Pre MMR update
all solo, 48-50% win rate

yo guys, you need at least 30 matches recorded each to infer any sort of descriptive statistic -win likelihood, etc. Its a stats thing, trust me.

Was just in a match against a 5 man guild premade, while my team was all solo queue’s and even had a ranger who just got the game.

I was hoping for a revised matchmaking announcement as part of HoT. I wonder if ANET doesn’t realize that it is a problem or if they’ve just given up.