Guild Wars 2

So, I would like to know what exactly happened here. Before this horrible loss streak I had a positive win/loss ratio (12+ wins over loses) and 1646 rating points as far as I remember. I multiclass and play DH, Daredevil, Reaper and sometimes Scrapper depending on what’s best for the team. Also I know the classes as if they were my main (necromancer).

This is something that happened to me last season too, I was 1690+ with similar stats and bam! suddenly terrible loss streak. At that time, I wanted to see how far down the match making would take me so I didn’t mind it and ended 1420 mid season. And now, Season 8, the story seems to repeat and there’s a screenshot. I’m definitely not new to pvp, have more than 15k pvp matches and consider myself a above the average player.

But, I’m curious about what happened and I would appreciate if a dev look into this and give some feedback. Thanks.

I had the same thing multiple seasons in a row. Quite unexplainable. I also felt that I got matched up with teams below my skill level, which often caused that loss streak. Im also around the 1700 MMR normally. Like the algorithm determines that you have to carry bad teams in order to climb up the ladder. And no, Im not a guy who blames his team for every loss. I just observed it.

As much as I hate to be that person who says this, if you’re not the one who isn’t carrying their weight on your team, then someone on your team is causing drag vs someone on the enemy team carrying. As long as your dealing DPS and taking objectives, you should be okay for yourself. However, the downfall of Solo queue is the 4 random people you’re going to deal with every single game, and spoilers: These people are picked from the same pool of people you were playing with during this 2-3 hour period, so its highly likely you had someone on your team who just wasn’t doing enough and you were unfortunate to have that person on your team several times in a row. You’ll probably get a lot of “get good scrub!” hate cuz you’re complaining about losing, but sometimes, no matter how good you are, you can’t carry a sinking ship.

Simple Answer: Solo Queue has “those moments” – Especially when you’re pulling from the same queue of people over the 2-3 hour time frame you posted.

It’s been happening to me too: The worst parts are when my gains only give me +13 point score, but my losses give me like -20 point score, meaning despite me having an approximate 50% win ratio this season, I’m still losing points.

Why did you continue queuing into the loss streak?

Never queue after multiple losses you just get tilted and play worse which makes you lose more.

I always stop after two losses.

Crinn.7864:

Why did you continue queuing into the loss streak?

Never queue after multiple losses you just get tilted and play worse which makes you lose more.

I always stop after two losses.

You forgot to add that you’ll queue into the same pool of people, and will probably end up with the same teammates that caused you to loss.

that happens if you take a simple random numbers generator and call it matchmaker..

its like a pyramid scheme… you got a few ppl at top in legendary div and a endless pool of loser and noobs at copper div.. and since no one wants to wait.. the keep filling up the teams from the bottom up….

legendary gets filled up with platinum, if not enough platinum available then with gold player, then with silver and so on…

and since the copper swamp is a bottomless pit, they will fill up your teams eventually.
i even got ppl still doing their prematches into my team while ranked at gold.
during prime time, the win chances drop exponentially. playing ranked from 1pm to 11pm, my win chances are about 10-20%. on one win i get 5 to 10 loses. playing outside prime time the win chances are getting better and better, i played around 2am-5am and my win chance was about 80-90%

AlphaWolvesGamer.5790:

Crinn.7864:

Why did you continue queuing into the loss streak?

Never queue after multiple losses you just get tilted and play worse which makes you lose more.

I always stop after two losses.

You forgot to add that you’ll queue into the same pool of people, and will probably end up with the same teammates that caused you to loss.

See, what I truly don’t understand is why the system tends to put the same people on the same sides over and over again, instead of randomizing the teams. There are some players who I’ve played against for 10+ matches and have never had them on my team, and same goes for teammates who I’ve never fought against.

Welcome to the club, happens every season.

Out of 15,000 games played, you only played 4k ranked. I don’t know what you expect. Wtf were you doing in 10k + hot join games?

masterfan.8375:

Out of 15,000 games played, you only played 4k ranked. I don’t know what you expect. Wtf were you doing in 10k + hot join games?

You can clearly see that they have 4k ranked, 2 k almost in unranked as well as almost 1k in the old solo que.

So maybe about 7k games in a queing form.

Also i would like to point out that HOT JOIN games used to be super fun 8 vs 8 when the game had underwater combat and other things.

Hot join only became bad once they released sky hammer and killed 5 vs 5.

Anet did a great job in destroying that huge part of the game. It used to be you didnt que up in ranked for the old school tickets, you went into GW2 to find a server for 8 vs 8 and they were always full.

Now no one plays in hot join, ( the spam would be crazy if they had 8 vs 8 now) very few players are playing ranked and to the OP.

Nothing will change for win streaks or losing streaks because the player gap in the game is huge simply because no one is playing it.

8v8 hotjoin on Capricorn was good fun in the old days yeh. Not really strategic or skillfull and stuff, but fun it was.

play in the morning during the week, just skip work or whatever. you can basically carry every game. but really, don’t skip work or school or whatever.

AlphaWolvesGamer.5790:

As much as I hate to be that person who says this, if you’re not the one who isn’t carrying their weight on your team, then someone on your team is causing drag vs someone on the enemy team carrying. As long as your dealing DPS and taking objectives, you should be okay for yourself. However, the downfall of Solo queue is the 4 random people you’re going to deal with every single game, and spoilers: These people are picked from the same pool of people you were playing with during this 2-3 hour period, so its highly likely you had someone on your team who just wasn’t doing enough and you were unfortunate to have that person on your team several times in a row. You’ll probably get a lot of “get good scrub!” hate cuz you’re complaining about losing, but sometimes, no matter how good you are, you can’t carry a sinking ship.

Simple Answer: Solo Queue has “those moments” – Especially when you’re pulling from the same queue of people over the 2-3 hour time frame you posted.

It’s been happening to me too: The worst parts are when my gains only give me +13 point score, but my losses give me like -20 point score, meaning despite me having an approximate 50% win ratio this season, I’m still losing points.

This happens every season.
It may be true there are optimal times of play and there are things to improve individually but there’s a constant that usually causes this issue (especially as a Solo player).

This Usually starts happening after a certain point because the matchmaking algorithm is built for Solo Queue but given all the Duo-Queue players in the same leaderboard and rating, the system ends up folding in over itself to compensate. The system likely can’t balance the matchmaking from the pool of players available and more often than not Solo players suffer.

If Duo-ing was not part of the Solo-Queue ladder, this would happen less often and players wouldn’t be losing so many points for a loss.

Source: https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm

Lol this discussion again. You all do know these discussions fall on deaf ears outside the those adversely effective.

If you are a competitive pvp that’s not abusing out of match METAs with a duo partner. And you are not just Yolo queuing ranked without a care in the world of winning or losing. You all my friends are simply collateral damage. You all are bound to be not longed for this game’s sPvP mode.

My honest advice. Seek shelter somewhere that’s going to give it to you. Stop trying to force a place that not your home to become your home. It’s just not going to happen.

You can ignore all the writing on the wall. However it still will remain on the wall. Nothing you can do to change it. So you might as well change yourself and move on.

this all, specially those statements of plat-plat-plat-gold-silv vs silv-silv-gold-silv-silv and the read in the wiki (https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm) tell me the MM is looking first to find a proper 5-man-roster and then find a proper opponent.
this would declare much of the stated problems. maybe i misread the wiki, please clarify if so…

wouldn’t it be better overall to find a total of 10 players and if found team them up afterwards? like overall more balanced matchups? how could you just not do so?

well maybe im wrong…
please someone clarify me

Yeah, seeing it happen too – now on my 9th loss. Happy times in PvP….

All of the responses, even the OP’s original post seems to be misunderstanding the reason why a player will go on a 10+ game losing streak and that reason is match making that is literally “stacked against you”, where a plat 2 solo que is forced to get teams like: (you) as plat 2, gold, gold, gold, gold vs. teams like: leg, plat, plat, plat, gold. Start adding people to your friends list and start paying attention to these lopsided rating distributions during lose streaks. You’ll see that it is true.

The real question is:
Why are some players forced to play through these mandatory streaks of bad match making, resulting in frequent resetting of MMR/Rating levels and extremely volatile placement “Continuously ping ponging between 1600ish and 1350ish all season”, whilst others will tag a top 100 positioning directly out of placement and only ever fluctuate between 1600, to 1620 and maybe as low as 1580? Those are the two different types of patterns you see within the algorithm, if anyone cares to pay attention at all. You never hear of any stories that are in between these two extreme types of patterns. Honestly it would indicate a Control Group A and Control Group B type situation where some players are omitted from the mandatory bad match streaks whilst others are being used as stepping stones for others, within the algorithm.

I have a thread written in detail about this algorithm activity:
https://forum-en.gw2archive.eu/forum/game/pvp/Season-8-My-Match-Making-Response-Please/first

To me the issue isn’t players are stacked. The issue is the rating gap on a team. Legends and golds shouldn’t be on the same team. Right now in order to speed up queue times, the system will expand the rating of players that can be added to a team. If an all legend team was facing a plat 1/2 team, they would get very little for the win and the plat team would lose maybe 6 pts but could get 20 pts for a win.

I was at 1650 yesterday, Right now I am at 900, I think I won approximately 3 matches within the last approx 24 hours.

I have players on my team trebbing walls, I have players on my team talk nonsense nonstop and then turn a perfectly good match to hell, I have players who zerg around off the cap point, not capping anything, I have players zerg around who after kill left the point uncap, I have players who dont know what they are doing but still giving bad plan at the beginning of the match, and yet every brainless players think if someone say a plan that it must be follow without thinking and without reading the mini map, I have player ran away from the best only to be kill, any possible scenario that humiliate the hell out of me is already played out and i am reliving them in every match now.

the matching system is wrong somewhere. I dont know what it is but its not right. Or please, please just dont let me think that i’ve work my ass off, solo queuing from nothing to high gold rank to be humiliated in every match be be kick off to low silver now, I dont think I am going to stay in silver either, because the players are getting worse and the familiar “bot” player are starting to surface again.

Humiliation is what is left and if it wasn’t for the chest. I dont think there are anything left in gw2 thats worth playing.

I learned to stop after the 3 loss in a row. Just go do something else in-game or out of game, continue the next day. Helps to keep your sanity, a lot.

Ragmon.6350:

I learned to stop after the 3 loss in a row. Just go do something else in-game or out of game, continue the next day. Helps to keep your sanity, a lot.

I stop after the first loss per day.

But meanwhile on my F2P acct, i am pretty much on an 8 win streak :/

I had a lot of loses yesterday. Some wins, but not enough. I am just doing this for the legendary gear now

Why do people do this? Why do you go on a massive loss streak and act like matchmaking is working against you. It isn’t. If you lost 10 in a row you need to re-evaluate yourself and see what YOU are doing wrong and fix it to stop that from happening again. This is just like that other guy with his massive rant about losing 15 games in a row, but hes a skill clicker.

8 seasons in you people need to understand that the population is low and MM isn’t going to spend 30mins giving you a perfectly balanced game. If you want to win more then either start improving or play something cheese(condi thief/ventari rev)

You might be right. I currently have 2k games played. I play every profession and l have played them all almost evenly. My most played class is thief and it has around 360 games played. All the others are around 190+ so I definitely need to practice one class before I jump around to others.

What class to choose…

NotoriousNaru.1705:

Why do people do this? Why do you go on a massive loss streak and act like matchmaking is working against you. It isn’t. If you lost 10 in a row you need to re-evaluate yourself and see what YOU are doing wrong and fix it to stop that from happening again. This is just like that other guy with his massive rant about losing 15 games in a row, but hes a skill clicker.

8 seasons in you people need to understand that the population is low and MM isn’t going to spend 30mins giving you a perfectly balanced game. If you want to win more then either start improving or play something cheese(condi thief/ventari rev)

Big talk from smurfing culprit #1
Come back and post after you’ve actually experienced legitimate matches.

Trevor Boyer.6524:

NotoriousNaru.1705:

Why do people do this? Why do you go on a massive loss streak and act like matchmaking is working against you. It isn’t. If you lost 10 in a row you need to re-evaluate yourself and see what YOU are doing wrong and fix it to stop that from happening again. This is just like that other guy with his massive rant about losing 15 games in a row, but hes a skill clicker.

8 seasons in you people need to understand that the population is low and MM isn’t going to spend 30mins giving you a perfectly balanced game. If you want to win more then either start improving or play something cheese(condi thief/ventari rev)

Big talk from smurfing culprit #1
Come back and post after you’ve actually experienced legitimate matches.

You didn’t lose 15 games in a row because mm was working against you. You lost 15 in a row because you are actually very bad at the game(i watched ur vid). You have 0 game knowledge, 0 matchup knowledge, 0 map awareness, and 0 mechanics seeing as you skill click everything. Yet you complain about your loss streak.

If you can name one decent player that had mm work against him and they ended up losing 15 in a row maybe i’ll take a meme like yourself more seriously.

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

Lmao your memes never seem to disappoint.

You can’t lose 15 games in a row, cry on the forums, and expect anyone good at this game to take you seriously. I don’t think you truly grasp how bad ANYONE has to be to lose FIFTEEN ENTIRE GAMES in a row. If matchmaking was so flawed like you claim it to be, then why is there not a single semi decent or good player complaining about them having FIFTEEN games stacked against them. Here’s the reality of your situation Mr. I like to click my skills, of those 15 games you lost you probably had 2 maybe 3 stacked against you, and the other 12-13 you were just not good enough to win. I’m sorry to burst your buble but its true.

Do you notice a pattern here? Good players keep climbing up, and bad players keep losing(15 loss streak xd what a meme). MM isn’t out to get anyone most people coming on this forum crying about where they are and where they should be, are usually just delusional and are exactly where they deserve to be.

Faux Play.6104:

To me the issue isn’t players are stacked. The issue is the rating gap on a team. Legends and golds shouldn’t be on the same team. Right now in order to speed up queue times, the system will expand the rating of players that can be added to a team. If an all legend team was facing a plat 1/2 team, they would get very little for the win and the plat team would lose maybe 6 pts but could get 20 pts for a win.

Nah, a lot of the holders of legend titles aren’t all that good. I’ve fought primordial legends who drop DH traps while my pet is biting his ankle, really dubious plays like that.

Trevor boyer…Anyone that has clicked on ur forum signature instantly knows why you lost that many games and that its definitely not because of matchmaking.
You have a video of yourself playing symbol guard and clicking all your skills, not to mention the actual gameplay is super bad.
It’s kinda funny also that a skillclicker thinks by going on an alt they can hit legend.
Not that legend is supposed to be for good players, many people got there playing cheese builds and queuedodging, but even if we’re only talking about those guys – none of those cheese people skill click and are at least average/slightly above average skillwise.

Plat matches are frequently decided based on a couple top tier players in those matches. That means that a lot of your matches might be out of your control simply because you’re not as good as they are. However, if you’re losing 10 or 15 in a row, it’s almost definitely because you are upset and making more mistakes than usual. You push into bad situations because you think you need to carry. You complain because your team died while you were 1v3, but you fail to notice that you pushed into it while your teammates were on respawn.

I thought the whole ‘smurf’ trick doesn’t work anymore anyways, i.e. if a skilled player is on an alt paired with another skilled player on their main, the matchmaking will take the highest skill level anyways?

Trevor Boyer.6524:

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

The math behind the MMR system actually would predict around /-150. If you run the numbers in glicko 2 for individual players, you get a minimum possible deviation of 60. However, since you are matching 5 people on a team you have to sum the square of the errors. That results in a deviation of 134. What that means is 7 out of 20 people are more than +/- 134 frim where they are currently rated, and 1 person in 20 is more than -/ 268, and 3 out of 1000 are more than -/+402.

NotoriousNaru.1705:

Why do people do this? Why do you go on a massive loss streak and act like matchmaking is working against you. It isn’t. If you lost 10 in a row you need to re-evaluate yourself and see what YOU are doing wrong and fix it to stop that from happening again. This is just like that other guy with his massive rant about losing 15 games in a row, but hes a skill clicker.

8 seasons in you people need to understand that the population is low and MM isn’t going to spend 30mins giving you a perfectly balanced game. If you want to win more then either start improving or play something cheese(condi thief/ventari rev)

I’m not acting as if the system is against me, I’m stating something that happened to me twice and statiscally speaking it shouldn’t have and that’s why it called my attention. I’m slowly recovering my rating now and guess what; I didn’t have to “re-evaluate” myself and fix anything which obviously helps but is not the case. Just kept playing when most pvpers do and not at 3am.

To sum up, I just wanted to see the numbers on those 2 teams over those 10 or 15 loses, and what happened, no rant, no complains, etc.

NotoriousNaru.1705:

Lmao your memes never seem to disappoint.

You can’t lose 15 games in a row, cry on the forums, and expect anyone good at this game to take you seriously. I don’t think you truly grasp how bad ANYONE has to be to lose FIFTEEN ENTIRE GAMES in a row. If matchmaking was so flawed like you claim it to be, then why is there not a single semi decent or good player complaining about them having FIFTEEN games stacked against them. Here’s the reality of your situation Mr. I like to click my skills, of those 15 games you lost you probably had 2 maybe 3 stacked against you, and the other 12-13 you were just not good enough to win. I’m sorry to burst your buble but its true.

Do you notice a pattern here? Good players keep climbing up, and bad players keep losing(15 loss streak xd what a meme). MM isn’t out to get anyone most people coming on this forum crying about where they are and where they should be, are usually just delusional and are exactly where they deserve to be.

Psychological projection is a defense mechanism people subconsciously employ in order to cope with difficult feelings or emotions. Psychological projection involves projecting undesirable feelings or emotions onto someone else, rather than admitting to or dealing with the unwanted feelings.

You may be playing on a different account but you are the same person.

Levijeh.1467:

NotoriousNaru.1705:

Why do people do this? Why do you go on a massive loss streak and act like matchmaking is working against you. It isn’t. If you lost 10 in a row you need to re-evaluate yourself and see what YOU are doing wrong and fix it to stop that from happening again. This is just like that other guy with his massive rant about losing 15 games in a row, but hes a skill clicker.

8 seasons in you people need to understand that the population is low and MM isn’t going to spend 30mins giving you a perfectly balanced game. If you want to win more then either start improving or play something cheese(condi thief/ventari rev)

I’m not acting as if the system is against me, I’m stating something that happened to me twice and statiscally speaking it shouldn’t have and that’s why it called my attention. I’m slowly recovering my rating now and guess what; I didn’t have to “re-evaluate” myself and fix anything which obviously helps but is not the case. Just kept playing when most pvpers do and not at 3am.

To sum up, I just wanted to see the numbers on those 2 teams over those 10 or 15 loses, and what happened, no rant, no complains, etc.

Ironically this was posted right below a statistical explanation of why your rating will vary….

Faux Play.6104:

Trevor Boyer.6524:

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

The math behind the MMR system actually would predict around /-150. If you run the numbers in glicko 2 for individual players, you get a minimum possible deviation of 60. However, since you are matching 5 people on a team you have to sum the square of the errors. That results in a deviation of 134. What that means is 7 out of 20 people are more than +/- 134 frim where they are currently rated, and 1 person in 20 is more than -/ 268, and 3 out of 1000 are more than -/+402.

That logic is largely invalid because match sorting is not based on glicko nor does it match people of equal scores against each other.

Good thing that trevor boyer baddie isn’t denying he lost 15 in a row because hes bad and skill clicks and it looks like several people have also watched his signature video and have come to the same conclusion. My work here is done.

Bye lads

SneakyTouchy.6043:

Faux Play.6104:

Trevor Boyer.6524:

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

The math behind the MMR system actually would predict around /-150. If you run the numbers in glicko 2 for individual players, you get a minimum possible deviation of 60. However, since you are matching 5 people on a team you have to sum the square of the errors. That results in a deviation of 134. What that means is 7 out of 20 people are more than +/- 134 frim where they are currently rated, and 1 person in 20 is more than -/ 268, and 3 out of 1000 are more than -/+402.

That logic is largely invalid because match sorting is not based on glicko nor does it match people of equal scores against each other.

Getting players of similar ratings most certainly factors into the scoring for matchmaking per their wiki. It isn’t the only thing however.

The deviation associated with your rating is very relevant. The deviation of your rating is how it bounds your rating. It is saying it knows with 64% certainty that you are within one deviation of your rating, 95 confident you are within two deviations, and 99.7% confident you are within 3 deviations of your rating.

Faux Play.6104:

SneakyTouchy.6043:

Faux Play.6104:

Trevor Boyer.6524:

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

The math behind the MMR system actually would predict around /-150. If you run the numbers in glicko 2 for individual players, you get a minimum possible deviation of 60. However, since you are matching 5 people on a team you have to sum the square of the errors. That results in a deviation of 134. What that means is 7 out of 20 people are more than +/- 134 frim where they are currently rated, and 1 person in 20 is more than -/ 268, and 3 out of 1000 are more than -/+402.

That logic is largely invalid because match sorting is not based on glicko nor does it match people of equal scores against each other.

Getting players of similar ratings most certainly factors into the scoring for matchmaking per their wiki. It isn’t the only thing however.

The deviation associated with your rating is very relevant. The deviation of your rating is how it bounds your rating. It is saying it knows with 64% certainty that you are within one deviation of your rating, 95 confident you are within two deviations, and 99.7% confident you are within 3 deviations of your rating.

Except this logic isn’t applicable due to the matching process. It has too wide of a spectrum and doesn’t balance the teams while forming them.

This is why season 3 locked us into either permanent win streaks or permanent loss streaks, because the matching process was designed to be biased. Glicko requires the most balanced matching possible in order for it to function, or the assumption that you as a player will always have a 50% chance of being on the better team. Once that’s achieved, it’s accuracy is then strongly dependent on small variations of MMR sums. It would probably be most wise to assume there’s a curvature between MMR value and a player’s actual capability, but we don’t even have a linear assumption in place right now.

Since the relationship between matching algorithms and glicko are complex, I’d probably have to build a MATLAB simulation, but there’s clearly a threshold somewhere that fully disallows glicko from being relevant at all. Season 3 and 4 proved to break that threshold and I suspect we are currently lingering close to it.

I’ve learn win lose it doesn’t matter because it’s a joke… get the rewards and that’s it.

Most balanced match making possible? I don’t believe that at all. The glicko system would almost certainly need fluctuation of purposeful lopsided matches being thrown to different control groups at certain times for there to be point fluctuation between them at all. I’m not going to sit here and write you a 10 page thesis on this one but if you were to sit down and actually try to find ways to improve the glicko, you would self discover what I’m talking about. I had discovered that, without purposeful match stacking from time to time, a glicko ran system designed for individual ratings that is played through 5 man vs. 5 man teams, will actually implode over time and not expand, in terms of the margin of difference between high ranked players and low ranked players. This is mostly due to the mixing and matching of how the system is supposed to work and if the match making is actually making balanced matches, Each individual rating of a given player only represents 20% of his contribution and the other 80% directly represents other players that he is playing with. Hrmm now stop and think about this logically for a minute. You don’t need to run a simulation to understand that if matches were actually balanced in creation, the beginning of a season may start with highs looking like 1800 and lows looking like 900 but over the course of time, due to balanced match making, highs will be forced to carry more 900s in their team and 900s will be getting carried by more 1800s in their team. This should be resulting in 1800s being dragged down by 900s whilst 900s are dragged up by 1800s. All in the end, imploding that margin of difference in the whole number integer ratings. 1800 – 900 should end up looking more like 1500 – 1300 in the end. It certainly shouldn’t be expanding like we see, into 2300+ through 600. Even given the volatility in difference between rating gained and rating lost. There is also other supportive aspects of this effect:

• A 1800 wins a game, gains like +7 rating but when he loses a game, he gains like -40.
• A 900 wins a game being carried by a 1800 in his team, against another team with a 1800. The 900 gains like +25 points but if he loses only loses like -8
• The higher an individual’s rating gets, according to match making, he should be pulling in more lows to go against similar players of his rating also being forced to carry lows. He should not be getting stacked high teams against lower rated parties.
• The lower an individual’s rating gets, according to match making, he should be being grouped with higher ranked players to go against other lows who are teamed with high ranked players. He should not be getting stacked into bad teams that are repeatedly fed to high ranked players.
• Upon all of the above, we all know that guild wars 2 conquest is a game with hard limitations on how much one can carry a match. You have so much life value, only so many stunbreaks, only so much disengage, you only deal so much damage and above all, 1 good player can only defend 1 node at a single given time. If we were to compare some match up like Abjured vs. TCG, the match would be balanced. But if we were to run some control group situation and make the Abjured substitute Magic Toker out for some subpar player or let’s even say “me”, do you have any idea how much that would matter in the outcome of the match? Point being to take from this example: Great players are not immortal carry gods. If great players are supposed to go against other great players within this que system, they should have varied luck with who they draw in for their matches, resulting in varied wins and varied loses. They SHOULD NOT be rolling stacked teams of high ratings vs. lower ratings all the time. And if they are doing that frequently, something is wrong with the match making. Because they should be getting balanced matches for competition, even if it looks like: leg leg plat gold gold vs. plat plat plat plat plat. So why are they frequently getting matches like: leg leg plat plat plat vs. plat plat gold gold gold? That is high level kitten right there and no one can convince that it isn’t.

So if anyone were to run the numbers realistically here, they would see that in the presence of fair and equal match making, the margin of difference between high ranked players and low ranked players should be imploding over time, not expanding. At the very least, peak ratings for individual players should be settling at a realistic margin and not going past a certain point. It certainly should NOT be expanding. Ever expanding numeric margins suggest the presence of:

• High levels of meta match manipulation remaining within the community
• High levels of stacked match making that is designed to happen within the algorithm to make sure that the glicko works as intended for individual ratings within a 5v5 game.
• Remaining segregation within the system, something similar to division segregation, probably ran by listed MMR “which is indeed separate from Rating”. And MMR is not always gained or lost fairly as arenanet has us all agree to in the TOS before being allowed to download and play Guild Wars 2. This indicates the very real possibility of account favoritism.

Other things to point out:

• During this past season I have noticed more evidence of the control group A and control group B theory. I’ve noticed about 6 to 8 of my close friends will always begin lose streaks together on the same day and the same time “these people are not playing together, they are grouped into their own duos with people I don’t play with”. We all have bad match making streaks for this strangely exact same time frame, usually lasts for a couple days, then on the exact same day at the exact same time, it goes away. Then everyone gradually climbs back up the ladder at the same time while suddenly a different group of guys is on a magic lose streak to feed us? This wave of sudden shift in match quality has happened two or three times this season, at the same time, to the same groups of people, all in synchronicity. This is a statistical anomaly and shouldn’t/wouldn’t be happening over and over like automated clockwork, least it indicates control groups within the community and switches being flipped to create lose streaks as stepping stones. Why would they do this? Pretty sure it’s to make the glicko expand, rather than implode.
• The only way to make sure that glicko margins expand rather than implode with such a small community, would be to have some group of players varied all throughout the ratings, not just the top 50, who are pushing wins much more often within the balanced match making than anyone else, regardless of their actual skill levels. Then you would need a group of players varied through the ratings list, not just bronze players, who are somehow pushing 50% or lower win rates, regardless of their actual skill level. If the match making were actually balanced, it wouldn’t matter who was good and who was bad because they would be being mixed together in a balanced fashion, creating balanced matches. Simply put boys, the more balanced the match making is with average party mmr vs. average mrr, the more balanced the margin of difference begins to look from high rating to low rating, meaning it will begin to implode. To prevent this, lopsided matches must occur, especially within an ever shrinking community. It’s the only way to make sure there are enough differences in integers to reach from bronze 1, up to legend. Otherwise we have everyone hovering 1200 – 1600. The problem begins to get worse the smaller the community gets.

But why would they do this? It isn’t to screw with anyone, it’s simply to make sure the system works. But who is on the good list and who is on the bad list for win streaks and lose streaks and why are they allocated into these control groups? Well that’s a topic for a different thread but I can say this:

• Identified power players within the community have a loud voice and many of them stream. Power players who complain about bad match making is bad commercialism.
• Streamers in general, even the subpar skilled ones – No one wants to watch a loser on twitch, it’s boring. And it’s bad commercialism.
• People who are not identified power players, who do not stream, have no voice. No one cares if they complain about 20+ game lose streaks in a forum.
  Go figure how to allocate control group A and control group B

At any rate, I am truly convinced something like this is happening after all these years of viewing the patterns within the algorithm between the different demographics of players, concerning their status within the community and ability to represent Guild Wars 2, commercially.

Go ahead and toss all the pretentious remarks you want at me, explain to me how bad of a player I am and try to convince me how wrong the above statements are. But I think you’ll find it difficult to do that with everyone amongst the large list of players who are noticing the same things that I have, 5 years into this trail of evidence.

SneakyTouchy.6043:

Faux Play.6104:

SneakyTouchy.6043:

Faux Play.6104:

Trevor Boyer.6524:

You’re missing the point entirely, Naru. Even if I was a terrible player, my MMR/Rating should be settling at some point. It should not be ping ponging between 1600 and 1300 all season each and every season, due to bad match making streaks that function on an automated schedule that is as predictable as clockwork. Leg Plat Plat Plat Gold vs. (Me) Plat Plat Gold Gold Gold, for 15+ games in a row, is not balanced match making.

A few things that need to be said:

• You’re “God Of PvP” title was not real. We all know what you did to obtain it.
• If I were to begin smurfing, my 1500ish – 1600ish rating would sky rocket to 1800+ quite easily. Very very easily. But I’m sure you already know all about how that works.
• I don’t think you are as good of a player as you believe you are. If you want to put some push and shove behind your name again and if you are that hard up to defame mine, we should set up a time to do a non smurfing environment 1v1. It would be good for your twitch ratings. Will you take me up on this after all of the big talk or will you dodge?

~ Let me know, Naru

The math behind the MMR system actually would predict around /-150. If you run the numbers in glicko 2 for individual players, you get a minimum possible deviation of 60. However, since you are matching 5 people on a team you have to sum the square of the errors. That results in a deviation of 134. What that means is 7 out of 20 people are more than +/- 134 frim where they are currently rated, and 1 person in 20 is more than -/ 268, and 3 out of 1000 are more than -/+402.

That logic is largely invalid because match sorting is not based on glicko nor does it match people of equal scores against each other.

Getting players of similar ratings most certainly factors into the scoring for matchmaking per their wiki. It isn’t the only thing however.

The deviation associated with your rating is very relevant. The deviation of your rating is how it bounds your rating. It is saying it knows with 64% certainty that you are within one deviation of your rating, 95 confident you are within two deviations, and 99.7% confident you are within 3 deviations of your rating.

Except this logic isn’t applicable due to the matching process. It has too wide of a spectrum and doesn’t balance the teams while forming them.

This is why season 3 locked us into either permanent win streaks or permanent loss streaks, because the matching process was designed to be biased. Glicko requires the most balanced matching possible in order for it to function, or the assumption that you as a player will always have a 50% chance of being on the better team. Once that’s achieved, it’s accuracy is then strongly dependent on small variations of MMR sums. It would probably be most wise to assume there’s a curvature between MMR value and a player’s actual capability, but we don’t even have a linear assumption in place right now.

Since the relationship between matching algorithms and glicko are complex, I’d probably have to build a MATLAB simulation, but there’s clearly a threshold somewhere that fully disallows glicko from being relevant at all. Season 3 and 4 proved to break that threshold and I suspect we are currently lingering close to it.

Glicko does not need balanced matches to properly rate players. People on a team need to be the same or as close as possible to the same rating for it to work best. If you try to evenly distribute the good and bad players between two teams you will end up driving everyone to the same rating.

Trevor Boyer.6524:

Simply put boys, the more balanced the match making is with average party mmr vs. average mrr, the more balanced the margin of difference begins to look from high rating to low rating, meaning it will begin to implode. To prevent this, lopsided matches must occur, especially within an ever shrinking community. It’s the only way to make sure there are enough differences in integers to reach from bronze 1, up to legend.

this is true regardless of the population size. If you don’t try to put player of similar ratings on the same team, you will drive everyone to the average rating. The math will keep the tails of the distribution from expanding roo much by increasing a players deviation if they play too many matches outside their skill level.

What you are observing is more likely due to people logging into alts, or other fluctuations in player population.

Pvp is all about luck nowadays,
Your pary must have members who
1. Its not their first round of pvp ever
2. They know what classes are
3. They know what a skyhammer, stilness, Lord, etc is.
4. They dont run with 3 people to the close node at the start of the game.
5. At least one players isn’t afk, or goes afk when you’re 50 points behind.

Of the above 5, you may select 1 to exclude, The others wont be in your party

I am a decent player and started at 3 gold after placement – the strangething was that in 2 even 3 gold matches there were lots bad players and got like a 20+ straight of defeats (!). Some even with ridiculous scores. Something aint right with the match making engine for sure.

I don’t care about my personal rating rather than some fun and good (balanced) matches. If it is about giving ‘less queue time matches’ I personally prefer to wait more in order to get a balanced game.

It sounds like you got placed higher than you deserved, Dreddo. That can happen in placements because the deviation from each match is so high.

All those ’’arguments’’ from DuoQ pro’s here aren’t touching the true problem. The MMR system is just skewed. Trust me, the loss streaks I have seen in previous seasons (not this season for some reason) weren’t caused by anger/rage after some losses, or by being too bad, or by being placed at a wrong division after the first 10 matches. The MMR’s of 5 team members combined should by nearly equal to the opposing team. I’ve seen people posting proof of a lot of situations in which this was not the case, and that is the very most important reason for a lot of losses in those loss streaks.

Think of another reason, and then think about the probability that that reason is causing for example 10 losses in a row. For 1 person in a 5-man team game, come on… its highly unlikely.