how does the matchmaking work with duos?

[Khalisto.5780]

a 1200+1500 duo

 

does the matchmaking try to balance out this with 2 1350 players or it tries to find a 1200 and a 1500 player to match against the duo?

[Bazsi.2734]

The duo counts as the higher rated player, but twice. So in this case as two barely platinum players.

 EDIT: yeah apparently not, I assumed too much about a-net. Not going to make that mistake again. You can wintrade by throwing matches on alts and grouping up with an account you want to boost, it's totally legal. Best game.

Edited August 19, 2022 by Bazsi.2734

[Xentera.4560]

Pretty sure it takes the average rating. Just like it takes the average team rating. And then targets a win rate of 50%.

Pretty kitten up system if you ask me.

[Trevor Boyer.6524]

1 hour ago, Bazsi.2734 said:

The duo counts as the higher rated player, but twice. So in this case as two barely platinum players.

This is not true and we've gone over this dozens and dozens of times in this forum. We have even had Arenanet devs make quotes about this in years past.

A duo uses the average of the two players ratings. If his 1500 and 1200 queue together, it places them as a 1350 duo. This is why if they win the game, the 1500 gets +8 but the 1200 gets +12, or if they lose the 1500 gets -17 but the 1200 only gets -10. This is because they were placed as 1350 so the other players they are against are usually all around that 1350, resulting in the 1500 getting bad gains and heavy losses, but the 1200 getting high gains and minimal losses.

This is what the system tries to do, but with low population of course being queued as a 1350 duo could still likely result in going against Team USA. It is what it is in 2022.

@Khalisto.5780

Edited August 16, 2022 by Trevor Boyer.6524

[Multicolorhipster.9751]

It uses an average.

It doesn't use the higher-rated player's rank for matchmaking purposes. That's an old lie propagated by the wintrading cartel because DuoQ(as it is right now) is instrumental to their bidness. 

 

In all likelihood there's probably nothing, if not very little that separates a DuoQ from two individual SoloQs when they queue together. Anything else would take effort and some changes to the matchmaker; all specifically for PvP, which hasn't happened in several years now.

There's also nothing about it in the publicly viewable matchmaking algorithm on the Gw2 wiki.

 

[Bazsi.2734]

28 minutes ago, Trevor Boyer.6524 said:

This is not true and we've gone over this dozens and dozens of times in this forum. We have even had Arenanet devs make quotes about this in years past.

A duo uses the average of the two players ratings. If his 1500 and 1200 queue together, it places them as a 1350 duo. This is why if they win the game, the 1500 gets +8 but the 1200 gets +12, or if they lose the 1500 gets -17 but the 1200 only gets -10. This is because they were placed as 1350 so the other players they are against are usually all around that 1350, resulting in the 1500 getting bad gains and heavy losses, but the 1200 getting high gains and minimal losses.

This is what the system tries to do, but with low population of course being queued as a 1350 duo could still likely result in going against Team USA. It is what it is in 2022.

@Khalisto.5780

Oh is that so? Could you cite some of it? I legit remember(can't cite sources though) it just flat out using the higher rating. I don't want to spread misinformation but I cannot find what you're saying here either.

Edited August 16, 2022 by Bazsi.2734

[Khalisto.5780]

2 hours ago, Multicolorhipster.9751 said:

It uses an average.

It doesn't use the higher-rated player's rank for matchmaking purposes. That's an old lie propagated by the wintrading cartel because DuoQ(as it is right now) is instrumental to their bidness. 

 

In all likelihood there's probably nothing, if not very little that separates a DuoQ from two individual SoloQs when they queue together. Anything else would take effort and some changes to the matchmaker; all specifically for PvP, which hasn't happened in several years now.

There's also nothing about it in the publicly viewable matchmaking algorithm on the Gw2 wiki.

 

I thought so, since it is not very hard to stumble into a known plat player duo with a 800 ap free account with p2 gameplay level offhours

[Trevor Boyer.6524]

4 hours ago, Bazsi.2734 said:

Oh is that so? Could you cite some of it? I legit remember(can't cite sources though) it just flat out using the higher rating. I don't want to spread misinformation but I cannot find what you're saying here either.

I've already done this several times in this subforum, and actually I'm pretty sure I went through this with you specifically one other time already. Normally I'd go dig through and find a quote but not with this one, not again. This one is too hard to find, requiring surfing through hundreds of pages of "match making threads", which there are hundreds and hundreds of. I'm the only one who ever takes the time to find/post this quote and I'm getting tired of doing it in all honesty. It's quite a choir to do.

It is seriously faster to see what the algorithm is doing for yourself by organically queueing with some person who is on a low rated alt, and just pay attention to your gains/losses vs. the low rated duo gains losses. 100% guarantee you this is what will happen:

1. If a 1550 queues with 1400, the 1550 will notice his gains on wins drops from when he is solo queue and when compared to his partner, the partner is usually gaining a point or two higher per win than the 1550. The 1550 will also notice the games are a bit easier.
2. If the 1550 queues with a 1300, the effects are stronger. 1550 gets even less rating on wins, 1300 is getting even more on wins. Now we are talking a difference of like +9 vs +13, and the 1550 will notice the games are quite easier.
3. If 1550 queues with 1200 or lower, the effects are wildly noticeable. 1550 will get like +7 on wins and the 1200 like +14s, and usually the duo will go on something like a 5+ game win streak pretty easily. The lower the rating the 1550 queues with, the easier games will be because the duo average is getting lower and lower.
4. Here's the kicker, if the algorithm was only using the higher rated duo's rating like you suggested, the 1550 would experience no change in his rating gains, even if he were to queue with a bronze 1. But this is not the case. It is definitely using the average of the duo to pull in opponents to face, which is why the higher rated experiences massive gain loss the lower and lower rated his duo is, because when the party average lowers and lowers, it draws in weaker and weaker opponents.
5. All of this ^ is why alt smurfing is still definitely a thing. Players go on throw phases with alts to lower their alt's rating way down while pulling other players down with them during the throws. Then when they do a main account boost phase, the low rated alts come in and are suddenly played by p2+ players actually trying hard and it makes the matches a lot easier to win. This is all classic leaderboard manipulation that has been going on for a decade now.

You may think "This is why it should count only the highest rated player's rating for match making" and they went through a phase a very long time ago where they did do a patching that counted the duo as only the highest rated, but in the long haul this caused more problems than it helped. It was good vs. smurfing but it punished normal players who weren't cheating and people stopped wanting to queue together unless they were quite exactly the same rating, otherwise the higher rated guy gets punished for attempting to carry his buddy, and his low buddy gets a much harder match than what his rating should ever place him in. It was punishing for both the high rated and the low rated in the duo, just equating to much greater loss rates, making everyone just want to solo queue and not duo queue.. Arenanet eventually had a make a decision, and they chose to put it back to counts as duo's average, for that reason.

Edited August 17, 2022 by Trevor Boyer.6524

[Multicolorhipster.9751]

21 hours ago, Khalisto.5780 said:

I thought so, since it is not very hard to stumble into a known plat player duo with a 800 ap free account with p2 gameplay level offhours

I feel you there. I think we play on different servers, but I couldn't tell you how many variations of [name]godx i've seen, all duoQ'd with some upper-echelon wintrader, playing exactly as the 'pros' do.

 

And even if this didn't happen all the time, I don't know how anyone who's been here longer than a year could possibly believe the alternative. 

Even more surprising that anyone who believes that could go ahead and ask for proof when their greatest bit of evidence is "a dev said in a stream one time."

[Bazsi.2734]

On 8/17/2022 at 4:03 AM, Trevor Boyer.6524 said:

I've already done this several times in this subforum, and actually I'm pretty sure I went through this with you specifically one other time already. Normally I'd go dig through and find a quote but not with this one, not again. This one is too hard to find, requiring surfing through hundreds of pages of "match making threads", which there are hundreds and hundreds of. I'm the only one who ever takes the time to find/post this quote and I'm getting tired of doing it in all honesty. It's quite a choir to do.

It is seriously faster to see what the algorithm is doing for yourself by organically queueing with some person who is on a low rated alt, and just pay attention to your gains/losses vs. the low rated duo gains losses. 100% guarantee you this is what will happen:

1. If a 1550 queues with 1400, the 1550 will notice his gains on wins drops from when he is solo queue and when compared to his partner, the partner is usually gaining a point or two higher per win than the 1550. The 1550 will also notice the games are a bit easier.
2. If the 1550 queues with a 1300, the effects are stronger. 1550 gets even less rating on wins, 1300 is getting even more on wins. Now we are talking a difference of like +9 vs +13, and the 1550 will notice the games are quite easier.
3. If 1550 queues with 1200 or lower, the effects are wildly noticeable. 1550 will get like +7 on wins and the 1200 like +14s, and usually the duo will go on something like a 5+ game win streak pretty easily. The lower the rating the 1550 queues with, the easier games will be because the duo average is getting lower and lower.
4. Here's the kicker, if the algorithm was only using the higher rated duo's rating like you suggested, the 1550 would experience no change in his rating gains, even if he were to queue with a bronze 1. But this is not the case. It is definitely using the average of the duo to pull in opponents to face, which is why the higher rated experiences massive gain loss the lower and lower rated his duo is, because when the party average lowers and lowers, it draws in weaker and weaker opponents.
5. All of this ^ is why alt smurfing is still definitely a thing. Players go on throw phases with alts to lower their alt's rating way down while pulling other players down with them during the throws. Then when they do a main account boost phase, the low rated alts come in and are suddenly played by p2+ players actually trying hard and it makes the matches a lot easier to win. This is all classic leaderboard manipulation that has been going on for a decade now.

You may think "This is why it should count only the highest rated player's rating for match making" and they went through a phase a very long time ago where they did do a patching that counted the duo as only the highest rated, but in the long haul this caused more problems than it helped. It was good vs. smurfing but it punished normal players who weren't cheating and people stopped wanting to queue together unless they were quite exactly the same rating, otherwise the higher rated guy gets punished for attempting to carry his buddy, and his low buddy gets a much harder match than what his rating should ever place him in. It was punishing for both the high rated and the low rated in the duo, just equating to much greater loss rates, making everyone just want to solo queue and not duo queue.. Arenanet eventually had a make a decision, and they chose to put it back to counts as duo's average, for that reason.

So I just spent (way too much) time pushing the dev tracker to its limits and I found nothing related to this. Therefore I cannot tell which one of us is right. (Also now I'm angry. Time is precious.!)
If you can pair up with a lower rated account to make matches easier for you, then DuoQ-ing is by definition wintrade. I'm sure the game wouldn't allow such an obvious exploiting of the system, and this where I got my "information" from. Basicly it's just faith that whoever designed this PvP system isn't completely <REDACTED>.
 

I'm still quite sure I'm the right one on this, but... How come I cannot prove it? Why is there no dev statement on how this works???

[Trevor Boyer.6524]

28 minutes ago, Bazsi.2734 said:

So I just spent (way too much) time pushing the dev tracker to its limits and I found nothing related to this. Therefore I cannot tell which one of us is right.

Actually, if you searched my posts, you'd probably find it faster.

Somewhere in there, probably within the past 12 months, I had posted that quote somewhere for like the 5th time.

I mean, I'm not lying to you my dude. I have no reason to do that. There are like 2 or 3 direct quotes from devs, including one from the guy who actually wrote the algorithm, stating that it uses the average of a duo.

It USED TO use the higher rating of the duo, many years ago when seasons were first released, but they changed it average of duo for the reasons I mentioned in that previous post.

[Multicolorhipster.9751]

4 hours ago, Bazsi.2734 said:

So I just spent (way too much) time pushing the dev tracker to its limits and I found nothing related to this. Therefore I cannot tell which one of us is right. (Also now I'm angry. Time is precious.!)
If you can pair up with a lower rated account to make matches easier for you, then DuoQ-ing is by definition wintrade. I'm sure the game wouldn't allow such an obvious exploiting of the system, and this where I got my "information" from. Basicly it's just faith that whoever designed this PvP system isn't completely <REDACTED>.

When the Wintrader Cartel becomes the topic of the day then what you can see and hear is for the most part redundant anyway.

What you should really be asking yourself is what you're not seeing.

 

For example, i've never seen Arenanet put actual effort into PvP development since Ranked season 13(funnily enough the same season where DuoQ was readded)

And very few of the lot of us have ever heard them even mention DuoQ, let alone how it works for the sake of matchmaking.

4 hours ago, Bazsi.2734 said:

I'm still quite sure I'm the right one on this, but... How come I cannot prove it? Why is there no dev statement on how this works???

If it were me, I would hide all the evidence too. To escape the shame and mean forum words that come as a consequence of having aided an illegal cartel in breaking our own game's TOS.

This is a double-edged sword however, because while nobody can prove DuoQ works exactly as they think it does, the other side also can't prove their own little made up story about how DuoQ magically uses the highest rated player's ranking, despite the algorithm being publicly available and never once mentioning any of it, or DuoQ at all.

 I wish you luck in the endeavor though. When I tried, all I was able to find between here and the old forums was the oldest mention of the idea that "DuoQ uses the highest-rated player's rating for Ranked MMR." Which was by a player; Sindrener, and not a dev. Y'know, I cannot for a second imagine why a top player that DuoQs for every single match would say such a thing. Very bizarre. 🤯

[Crab Fear.8623]

Duos, teams, ect....don't effect the matchmaker at all.

Only each individual player rating matters.

That's it, and you can see it in the code and how it is explained by the wiki.

https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm#Configuration_2

 

A simple way to describe what is going on is that the matchmaker takes scores from each player in the roster and tries to find an appropriate match, which some has said is made possible by averaging, but this is not what is happening.

The matchmaker is only trying to make a mirror (or as close as possible) of both team, for each individual player.

 

SO YES, IF YOU QUEUE WITH A SMURF YOU CAN GET EASIER MATCHES.

 

"The first phase, called filtering, gathers players based on their current MMR. The primary purpose of this phase is to both reduce the number of players being considered for a match, and to ensure that the match is appropriate given each player's skill level. Over time, padding is added to your player rating. While this may decrease match quality, it helps ensure that outliers still receive matches. "

Edited August 19, 2022 by Crab Fear.8623

[Trevor Boyer.6524]

3 hours ago, Multicolorhipster.9751 said:

And very few of the lot of us have ever heard them even mention DuoQ, let alone how it works for the sake of matchmaking.

It was over half a decade ago now.

These quotes everyone is trying to find are in the old forum, not here - > Guild Wars 2 Forum - PvP (gw2archive.eu) This is all 2016 or older. Back then, the devs actually communicated often in those threads. People would just make threads saying "Can I get confirmation on this?" and people like Grouch, Ben, CMC, that guy who designed the algorithm, would actually respond.

When they made this new forum they made some kind of a new rule and just stopped communicating with us. I mean with the kind of salt that was thrown at them each and every time they said anything at all, I don't blame them.

Anyway, if anyone cares to find those quotes, this is where they'll be at. In those older forums you'll find a bunch of interesting stuff stated from Arenanet devs that you probably didn't know. Interesting read-through for those who didn't show up until about 2016-2017.

The threads with dev response are marked with big red arenanet signs, so they are easy to find. You'll notice that there are a lot of them. ~ Enjoy the hunt for those quotes if you dare, this was back when the game had 20x more population posting every day.

@Bazsi.2734

Edited August 19, 2022 by Trevor Boyer.6524

[Trevor Boyer.6524]

@Khalisto.5780 @Bazsi.2734 @Multicolorhipster.9751 

Last time I'm digging up this impossible to find quote for everyone.

Prepare yourselves for all to be answered.

-> GW2 - Does Duo Queue Use Highest Or Average Rating? - Twitch

Edited August 19, 2022 by Trevor Boyer.6524

[greedywholesome.9081]

Hello. I remember seeing a formula on this sub earlier this year about MMR and how duo's are calculated. From what I recall, it's either the average of the two and/or that they aren't even considered technically as duo's but calculated independently. To put in perspective, if this weren't the case, some people wouldn't be abusing this loophole to get easier matches

[Bazsi.2734]

1 hour ago, Trevor Boyer.6524 said:

@Khalisto.5780 @Bazsi.2734 @Multicolorhipster.9751 

Last time I'm digging up this impossible to find quote for everyone.

Prepare yourselves for all to be answered.

-> GW2 - Does Duo Queue Use Highest Or Average Rating? - Twitch

Well... if that info is accurate(2017 was 5 kittening years ago, but there are seemingly no changes since then so it totally could be), then... oh my god.
My strong opposition to having duoQ in ranked was with me thinking that it uses the higher rating of the duo, not an average. But with this... f*ck ranked. F*ck this pvp, I'm so done, jesus christ.

[Multicolorhipster.9751]

5 hours ago, Trevor Boyer.6524 said:

 

Last time I'm digging up this impossible to find quote for everyone.

Prepare yourselves for all to be answered.

-> GW2 - Does Duo Queue Use Highest Or Average Rating? - Twitch

How long did it take to synch it so perfectly to the music? 😲

[Trevor Boyer.6524]

10 hours ago, Bazsi.2734 said:

Well... if that info is accurate(2017 was 5 kittening years ago, but there are seemingly no changes since then so it totally could be), then... oh my god.
My strong opposition to having duoQ in ranked was with me thinking that it uses the higher rating of the duo, not an average. But with this... f*ck ranked. F*ck this pvp, I'm so done, jesus christ.

@Multicolorhipster.9751 told you man

That's why they throw on us even without supporting a main, because it's IN PREPARATION to go duo with a main to support it. When the alts get too high rating, they do phases where they just enter games and lose on purpose to lower the rating back down real far again, so they can queue with the main and help the main again.

And then of course when that 1000 rated alt joins the main, make sure an actual p2+ is smurfing hard on it so the matches are ridiculously easier than they should be.

It's just garbage. It's way too easy to exploit. This is why the match quality is so randomly volatile dice throw bad. If you aren't in on the throw wars like this, playing to even 1550 and maintaining it, is like Herculeic struggle in 2022.

This is why they make so many new alts all the time, because alts with really low total games played get easier matches. I mean pretty much anyone can snag a top 25 title if you play at around your skill level as example, and a have a couple buddies at the same level who want to help you do this. All you need to do is make sure to queue dodge other win trade crews.

6 hours ago, Multicolorhipster.9751 said:

How long did it take to synch it so perfectly to the music? 😲

lol had to practice it like 5x or 6x before making the stream go live.

Edited August 19, 2022 by Trevor Boyer.6524

[Kuma.1503]

Poorly 

 

/thread

[Multicolorhipster.9751]

6 hours ago, Trevor Boyer.6524 said:

@Multicolorhipster.9751 told you man

What? Me? I've been trying to tell people this exact same thing for 4 years!

After months spent in Skillshare(I followed the link on a random YT video to get 15% my first month) taking basic hour-long courses on Criminal Psychology, I arrived at the conclusion that the Wintraders are all clout-chasing egomaniacs. 

The last thing some clout-chasing egomaniac is going to do is queue up with their silver buddy to help them get to gold or plat. If a known wintrader is putting their precious rating at risk by queuing with someone rated way lower than them, then that's all the evidence one needs to make a very safe guess that something isn't right.

6 hours ago, Trevor Boyer.6524 said:

lol had to practice it like 5x or 6x before making the stream go live.

Well worth it though, that was pretty hype.

It will serve as good practice for the day when the cartel finally falls. With revelations like this, that's only a matter of time.

[Bast.7253]

29 minutes ago, Multicolorhipster.9751 said:

What? Me? I've been trying to tell people this exact same thing for 4 years!

After months spent in Skillshare(I followed the link on a random YT video to get 15% my first month) taking basic hour-long courses on Criminal Psychology, I arrived at the conclusion that the Wintraders are all clout-chasing egomaniacs. 

The last thing some clout-chasing egomaniac is going to do is queue up with their silver buddy to help them get to gold or plat. If a known wintrader is putting their precious rating at risk by queuing with someone rated way lower than them, then that's all the evidence one needs to make a very safe guess that something isn't right.

Well worth it though, that was pretty hype.

It will serve as good practice for the day when the cartel finally falls. With revelations like this, that's only a matter of time.

The cartel will never fall. lol 

The only thing I can see putting a dent in this is removing duo queue and creating a separate conquest mode for premades of all group sizes. 

I doubt they would do this as it's difficult enough to get a full queue for the current mode. But hey, frequent balance patches and not letting something on metabattle reign supreme for 3-4 years would probably help. Not doing bandaid fixes like slapping boons or nerfing and then reverting damage coefficients or condition and boon durations. Updating existing utility skills and taking a long hard look at traits that consistently remain choice options on popular sites like metabattle - or comparing existing skills and traits to those of other professions. Balance the design, not the statistics and simcrafts. 

I mean really,  to the people making these decisions, where do you see the state of these professions in another few years? Because I'm seeing every profession with access to every boon to the point that boons mean nothing anymore. They're already freely given in abundance to some professions whereas others have to sacrifice much more impactful traits just to gain the chance to have access to them. 

Here's another example - you give Virtuoso an entire traitline dedicated to conditions but then give dagger no conditions on their weapon skills. What do you give mechanist? Everything. What do you give harbinger and specter? Like 20 stacks of rapid application torment. You make Catalyst a melee bruiser spec - again - and then give it stab access as a trait - meanwhile engineers get pulsing stab just from auto attacking with a flamethrower?

There is no consistency or consideration taken into account when you start comparing professions with one another. I'm not suggesting homogenization, I'm suggesting research and consideration beyond what feels flavorful or what can be thrown together - or even previous specializations or class mechanics.

Again - I would like to you back to Harbinger who (isn't as punishing as it was) has a gravity well on skill 5 vs Chronomancer that has a ground targeted elite skill or weaver that has to cycle through 4 attunements and has a timer in which it can be used.

Or we can look at the signets on mechanist and compare them to some of the elementalist utilities - compare kits to conjured weapons, update core utilities that are rarely ever used or elite spec utilities that are rarely ever used. (Mirage advance which is basically just a lesser shadowstep?)

If they're going to give a bunch of abilities the same functionality, why give one profession a weaker version just for the sake of flavor? Why have abilities like the arcane ferocity boost for elementalists that as far as I know - has had next to no use from anyone at least since Heart of Thorns, if even before then. 

Update existing elite skills so there's more option for certain professions. Do some actual change, not just create the illusion of change. I understand it's not that simple given that there's balance passes required and higher thinking decisions but it doesn't feel as though they've put half the effort into it that they should have.

If you're going to revitalize core content and start improving the "new player experience" then perhaps dedicate more resources to things that impact the entirety of the game, not just making world bosses flashier with more health or changing dungeon currency that frankly should have been done shortly after they decided they were done making them. 

[Trevor Boyer.6524]

1 minute ago, Bast.7253 said:

I mean really,  to the people making these decisions, where do you see the state of these professions in another few years?

Oh man, in another few years, if they continue on this amazing power creep trend, we are going to see builds that can realistically solo raid bosses, given enough time to deal the DPS.

They really really really need to dial waaay back on the power creep.

[Bast.7253]

28 minutes ago, Trevor Boyer.6524 said:

Oh man, in another few years, if they continue on this amazing power creep trend, we are going to see builds that can realistically solo raid bosses, given enough time to deal the DPS.

They really really really need to dial waaay back on the power creep.

It's their obsession with boons, adding needless conditions for the sake of giving everything the "option" of running a condi variant despite half of them still not being actually viable for condition builds, and passive damage/procs on traits. Why make a balance patch lowering coefficients on nearly everything only to negate it by overloading traits or making weapon skills that deal multiple packets of splash damage? 

I mean I guess boons are fine but imo a profession should never run around with more than 10 stacks of self-given might, perma quickness, perma stab, or aegis spam. If stability is that necessary, reduce the number of crowd control effects in the game. These are also leading to insane power creep and it's often certain professions that have much easier access and stronger crowd control abilities than others. 

It's not fun being a new player entering a match and being stunlocked or tossed around the map by a mechanist, or trying to block an attack and having some Bladesworn's unblockable ability half their healthbar, or trying to get on a cap to stall it and winding up floating in space from some planted gravity well. 

It's like they have all these devices to enable counterplay and then just... forget they exist and create things to counter the counterplay. And it's skewed in favor of one profession over another.

[Khalisto.5780]

17 hours ago, Bazsi.2734 said:

Well... if that info is accurate(2017 was 5 kittening years ago, but there are seemingly no changes since then so it totally could be), then... oh my god.
My strong opposition to having duoQ in ranked was with me thinking that it uses the higher rating of the duo, not an average. But with this... f*ck ranked. F*ck this pvp, I'm so done, jesus christ.

i dont know which server you play or how much you play, but NA it's pretty obvious when you stumble into a known plat player duo with a 600 AP account playing 1 shot mesmer with plat 2 skills, to assume it's the average mmr taken to matchmaking