I don’t think they divide the leagues. We’ve seen plenty of anecdotes in the forums about Silver players matched with / against Platinum and higher players.
ArenaNet stated they’ve shifted more towards the quicker queues than… hmm evenly matched queues for this season. I guess we’ll see a popular push for more rigidly segregated leagues, even if it means longer queue times.
I’m sure some people remember S1-4 queue times becoming nearly unplayable after reaching the higher leagues.
I think you skipped over the point of our (Exodore and I) conversation… He said Placements is a whole lot more accurate than I make it out to be. I was proving him wrong with facts.
I said in my original post that achieving an accurate Skill Rating takes about 100 games. Now who’s skipping?
I suggest you refrain from replying if you can’t follow the conversation we were having.
You’re not gonna bully me out of my post. I’m sorry the soft-reset and (maybe) unfair placement matches didn’t park you and a lot of people where they felt they belonged. But the system only works after people play dozens of matches.
I highly doubt we’ll face another soft-reset in Season 6, so you can probably relax.
I think you skipped over the point of our (Exodore and I) conversation… He said Placements is a whole lot more accurate than I make it out to be. I was proving him wrong with facts.
I said in my original post that achieving an accurate Skill Rating takes about 100 games. Now who’s skipping?
I suggest you refrain from replying if you can’t follow the conversation we were having.
You’re not gonna bully me out of my post. I’m sorry the soft-reset and (maybe) unfair placement matches didn’t park you and a lot of people where they felt they belonged. But the system only works after people play dozens of matches.
I highly doubt we’ll face another soft-reset in Season 6, so you can probably relax.
I agree that games get a lot better after 100-150 games but I think that’s match making and not the skill rating….
I enquired in another thread if there was a hidden match making algorithm that puts similar skilled players together while disregarding actual skill rating but no one could really answer that.
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
I agree that games get a lot better after 100-150 games but I think that’s match making and not the skill rating….
aren’t they the same thing?
I’m not entirely sure with this season, appears people are all over the place. But Krewe is right, the games DO get better after 100-150 games, even if that means you’re silver playing with/against platinum. So technically to get into gold, you need to play at a level higher than platinum.
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
I’m curious where your faith in these test-run Placement matches came from because so many people got screwed over because the system “wasn’t perfect.”
A DC player meant an instant loss. Some unlucky blokes have gotten 3x DC in a row. My 60% winrate means, statistically, i’ll win 6 out of my 10 games. Well I got a DC and another 1 unlucky game where some one dc’d but came back before 2m.
Meanwhile you have those certain few who duo on alt account where one person DC’s to prevent a loss… cheating the system completely. You’re right it’s not perfect, it’s down right broken.
“so many people”
I doubt that. You’re using a few accounts from the vocal minority and assuming it’s the norm. When I look at my friends who PvP, all but one landed in the ballpark (within a couple hundred points) of where they should have been (based on personal assessment and where they are now). So clearly it’s not perfect, but nor is it too flawed to trust.
In the case of DCs and AFK trolls, they typically account for a small percentage of matches and affect teams equally. The only exception would be intentionally trying to snipe someone’s matches.
Microsoft Trueskill is another rating system. The FAQ has a comparison between Glicko and Trueskill, but Microsoft’s write-up is just a little biased. https://www.microsoft.com/en-us/research/project/trueskill-ranking-system/
Glicko2 can be adapted to games of random teams and a matchmaker applied. Many games besides GW2 have done it.
I don’t think they divide the leagues. We’ve seen plenty of anecdotes in the forums about Silver players matched with / against Platinum and higher players.
Overwatch does not have league restrictions on matchmaking. However, they do have a much larger playerbase, so finding a close match is easier and faster.
One important thing to note is that Overwatch does limit the skill differential on a group queue. I think it’s 1500 rating (in a system of 0-5000) between highest and lowest member of the group.
I’m curious where your faith in these test-run Placement matches came from because so many people got screwed over because the system “wasn’t perfect.”
A DC player meant an instant loss. Some unlucky blokes have gotten 3x DC in a row. My 60% winrate means, statistically, i’ll win 6 out of my 10 games. Well I got a DC and another 1 unlucky game where some one dc’d but came back before 2m.
Meanwhile you have those certain few who duo on alt account where one person DC’s to prevent a loss… cheating the system completely. You’re right it’s not perfect, it’s down right broken.
“so many people”
I doubt that. You’re using a few accounts from the vocal minority and assuming it’s the norm. When I look at my friends who PvP, all but one landed in the ballpark (within a couple hundred points) of where they should have been (based on personal assessment and where they are now). So clearly it’s not perfect, but nor is it too flawed to trust.
In the case of DCs and AFK trolls, they typically account for a small percentage of matches and affect teams equally. The only exception would be intentionally trying to snipe someone’s matches.
Microsoft Trueskill is another rating system. The FAQ has a comparison between Glicko and Trueskill, but Microsoft’s write-up is just a little biased. https://www.microsoft.com/en-us/research/project/trueskill-ranking-system/
Glicko2 can be adapted to games of random teams and a matchmaker applied. Many games besides GW2 have done it.
I don’t think they divide the leagues. We’ve seen plenty of anecdotes in the forums about Silver players matched with / against Platinum and higher players.
Overwatch does not have league restrictions on matchmaking. However, they do have a much larger playerbase, so finding a close match is easier and faster.
One important thing to note is that Overwatch does limit the skill differential on a group queue. I think it’s 1500 rating (in a system of 0-5000) between highest and lowest member of the group.
Do you think it may be beneficial for GW2 to implement the same cap on matches and skill rating?
Say maybe 200? I know our playerbase is not as big and this could result in slightly longer queues but I think it would be better for measuring ones skill rating more accurately.
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
One important thing to note is that Overwatch does limit the skill differential on a group queue. I think it’s 1500 rating (in a system of 0-5000) between highest and lowest member of the group.
Do you think it may be beneficial for GW2 to implement the same cap on matches and skill rating?
Say maybe 200? I know our playerbase is not as big and this could result in slightly longer queues but I think it would be better for measuring ones skill rating more accurately.
The purpose of the differential cap in OW was to stop top players from trolling average and bottom players (average group rating is used for matchmaking there) – basically stop outlier matches.
Assuming 1200 is the true mid-point in GW2 ratings, the equivalent differential in GW2 would be 720. I would say dropping that differential to 500 would be reasonable, but anything more and you’re hurting the ability for friends to play together.
ArenaNet definitely dialed matchmaking toward the faster queue times / less closely rated players for this season; I usually don’t wait more than 2 minutes.
I’d be willing to bet we’ll see a “shave” towards more closely rated players / longer queue times for Season 6. I’m behind a change in that direction.
just to put things into perspective regarding total pvp population, according to gw2efficiency, a skill rating of 1850 is about where the top 1% of the population is.
Looking at the current NA leaderboards, 1850 MMR is at about rank 70.
Thus, one can extrapolate that there are maybe up to 7,000 players in TOTAL on NA that have played 10 or more PvP matches in the last one and a half months.
Miyoshinono, Yama no Akikaze Sayofukete, Furusato Samuku, Koromo Utsu Nari.
500 rating is too far, even 400 is. Anything above 250 should be scrapped, there’s a massive skill difference between someone at 2000 and 1750.
I’d rather see something less drastic as the player-base’s Skill Ratings come closer in line with their actual skill level.
I guess you have to form an opinion on how close the majority of players are. If most people are where they should be, then a drastic shift in league segregation could work. If they aren’t, then it might unfairly keep people in a league where they don’t belong (that works up and down).
And I’m still a little weary of starting new accounts (or players new to sPvP) at 1200. A brand new player most-likely doesn’t have the average set of skills that an active PvPer does. And I personally think alt-flipping is dumb, because gaming the system doesn’t help the developers.
just to put things into perspective regarding total pvp population, according to gw2efficiency, a skill rating of 1850 is about where the top 1% of the population is.
Looking at the current NA leaderboards, 1850 MMR is at about rank 70.
Thus, one can extrapolate that there are maybe up to 7,000 players in TOTAL on NA that have played 10 or more PvP matches in the last one and a half months.
That’s actually not 100% accurate, as gw2efficiency is going to include EU and NA so 1% being at 1854 rating might not be applicable to NA. It may or may not be close. Assuming you are correct though, the other number available of gw2efficiency is that there are currently 25,748 accounts with rating on them that are registered on gw2efficiency. This is out of a population of 109,006 accounts that have registered. I don’t think people in China use gw2efficiency… do they?
When ground-targetted bone minion explosions become a thing, I will change this signature.- 2013 http://twitch.tv/alemfi/
Dota2 uses what is similar to Glicko based on their descriptions.
Microsoft’s Trueskill has a Bayesian approach just like Glicko. The biggest selling point of Trueskill is that it has built-in support for random teams whereas Glicko2 has no official method for that.
500 rating is too far, even 400 is. Anything above 250 should be scrapped, there’s a massive skill difference between someone at 2000 and 1750.
Yes, a 2000 player vs. a 1800 player is night-and-day. But what about a difference between 1000 and 1400? Your example of the very top is an outlier. You can’t hinder the majority of players to address a very specific case. And if a top player did queue with someone much lower, do you think their carry potential at the average rating is strong enough to outweigh the risk of losing big?
just to put things into perspective regarding total pvp population, according to gw2efficiency, a skill rating of 1850 is about where the top 1% of the population is.
Looking at the current NA leaderboards, 1850 MMR is at about rank 70.
Thus, one can extrapolate that there are maybe up to 7,000 players in TOTAL on NA that have played 10 or more PvP matches in the last one and a half months.
That’s not a very good estimate. It’s unclear if GW2Efficiency is counting real rating or decayed rating. The leaderboard is decayed, so you’re likely underestimating by a lot.
And I’m still a little weary of starting new accounts (or players new to sPvP) at 1200. A brand new player most-likely doesn’t have the average set of skills that an active PvPer does.
“Average” is based on the population; it’s not based on any set of skills or knowledge. If 90% of players smash their face into their keyboard to control their character, then someone who smashes their face to PvP is pretty average.
As long as the statistic isn’t accounting to the number of games played it’s worth nothing. You can’t take rating out of the context. I’m at 1300 rating with ~20 games. How is this accurate? Right, it isn’t at all
[orz] below mediocre – we sponsor Arenanet
Piken Square EU, maybe soon on your server.
1000 and 1400 is also a massive difference in skill.
Looking through my friends list, the guys at 1.4k are pretty solid with rotations/builds but they lack PvP experience (raid players) and the 1k guys are my friends who aren’t even raiders and just do open world with minimal understanding of builds.
We can pull numbers all day but the fact is, there can never be any meaningful rating/matchmaking if you put some people 400 rating apart on the same/other team.
just to put things into perspective regarding total pvp population, according to gw2efficiency, a skill rating of 1850 is about where the top 1% of the population is.
Looking at the current NA leaderboards, 1850 MMR is at about rank 70.
Thus, one can extrapolate that there are maybe up to 7,000 players in TOTAL on NA that have played 10 or more PvP matches in the last one and a half months.
Math fail. Rating distributions in Glicko2 (and most rating systems) are Gaussian, not linear.
logic fail.
Top 1% of the population is still 1% of the population regardless of if the distribution is gussian, or linear or normal or circular or inverted.
Miyoshinono, Yama no Akikaze Sayofukete, Furusato Samuku, Koromo Utsu Nari.
Top 1% of the population is still 1% of the population regardless of if the distribution is gussian, or linear or normal or circular or inverted.
Ya, you got me there.
But your analysis is still flawed because NA and EU have different rating distributions. EU’s ratings spread out a bit more, so it’s picking up a disproportionate amount. If you go to 2.5%, that’s currently rank 194 in NA, which ends up being over 7700.
Further, there’s a good chunk of players who have decayed off the leaderboard. GW2Efficiency is likely using real rating and not decayed rating. If it used decayed, the 50% mark would be well below 1200.
Top 1% of the population is still 1% of the population regardless of if the distribution is gussian, or linear or normal or circular or inverted.
Ya, you got me there.
However, the leaderboard is decayed rating, so you’re probably underestimating by a lot.
Possibly, possibly not.
If people with decay all grinded their MMRs back up, the top score that indicates the top 1% will probably just go up a bit, and the number of people in that top 1% will stay similar.
Miyoshinono, Yama no Akikaze Sayofukete, Furusato Samuku, Koromo Utsu Nari.
Dota2 uses what is similar to Glicko based on their descriptions.
Microsoft’s Trueskill has a Bayesian approach just like Glicko. The biggest selling point of Trueskill is that it has built-in support for random teams whereas Glicko2 has no official method for that.
500 rating is too far, even 400 is. Anything above 250 should be scrapped, there’s a massive skill difference between someone at 2000 and 1750.
Yes, a 2000 player vs. a 1800 player is night-and-day. But what about a difference between 1000 and 1400? Your example of the very top is an outlier. You can’t hinder the majority of players to address a very specific case. And if a top player did queue with someone much lower, do you think their carry potential at the average rating is strong enough to outweigh the risk of losing big?
just to put things into perspective regarding total pvp population, according to gw2efficiency, a skill rating of 1850 is about where the top 1% of the population is.
Looking at the current NA leaderboards, 1850 MMR is at about rank 70.
Thus, one can extrapolate that there are maybe up to 7,000 players in TOTAL on NA that have played 10 or more PvP matches in the last one and a half months.
That’s not a very good estimate. It’s unclear if GW2Efficiency is counting real rating or decayed rating. The leaderboard is decayed, so you’re likely underestimating by a lot.
And I’m still a little weary of starting new accounts (or players new to sPvP) at 1200. A brand new player most-likely doesn’t have the average set of skills that an active PvPer does.
“Average” is based on the population; it’s not based on any set of skills or knowledge. If 90% of players smash their face into their keyboard to control their character, then someone who smashes their face to PvP is pretty average.
See I’m on the side of the fence that says anything above 250 should be scrapped and remade, I wouldn’t mind the longer queue times now that rating is involved. I know what I provided may be one specific case to you, but it’s not. these games are frequent and sometimes a few in a row and that’s just for me, I’ve read plenty more cases across these forums and these are just the ones people notice, plenty more probably go unnoticed..
Simply put: If people are in the same game with a 600 rating difference, it automatically debunks any meaning to rating at all. Yes it’s extreme but it’s also applicable to the lower gaps from say Gold T1 and Platinum…
Good news is they are changing season 6 but for now all we can do is speculate.
The new matchmaker is not worth risking game stability in a hotfix either. The earliest build to ship these on with adequate testing that doesn’t throw off the rest of the studio will be the off-season build.
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
Bleh. Skill rating and matchmaking are horribly designed. Most everyone is stuffed into gold in crap matches.
This season was the most boring since the game launched.
Total fail.
…wait, wait, you are saying that most players would up in the middle bracket, and that is a problem with the MMR system?
Because, uh, sounds exactly like what you would expect in a system that reflects actual skill in MMR, and MMR in the bracket. Most players are … wait for it … average! Just like you. Yay!
everyone should go back and re-evaluate there own play. I do not consider myself to be a “GOOD” player. I was able to make it to PLAT while drunk and blazed on a GS D/W all wells chrono rune necro. Yeah i said GS Necro.
Attention Moderators I am not
S P E E D Starr #0 Necro NA or
I Am NeXeD awful d/D ele NA
everyone should go back and re-evaluate there own play. I do not consider myself to be a “GOOD” player. I was able to make it to PLAT while drunk and blazed on a GS D/W all wells chrono rune necro. Yeah i said GS Necro.
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Attachments:
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Trying to figure out a way for that matchup to happen during NA primetime, but I’m at a loss. That’s one for Evan to investigate. You may want to make a new thread about this so that it gets more visibility.
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Trying to figure out a way for that matchup to happen during NA primetime, but I’m at a loss. That’s one for Evan to investigate. You may want to make a new thread about this so that it gets more visibility.
I’m pretty sure they’ve seen it and the other games I posted… I Don’t mind these games, gives me something to learn from but I am getting over the design of the system and how I’m supposed to climb, so it’s not the match I dislike.
Games are really fun when it’s full of Platinum players, I’ve actually got a chance to hold my own but seeing Naru hit top 5 last night was impressive, nj if you read this
I mostly play for the new Free-For-All arena in PvP lobby.
….. And Elementalist.
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Trying to figure out a way for that matchup to happen during NA primetime, but I’m at a loss. That’s one for Evan to investigate. You may want to make a new thread about this so that it gets more visibility.
I’m pretty sure they’ve seen it and the other games I posted… I Don’t mind these games, gives me something to learn from but I am getting over the design of the system and how I’m supposed to climb, so it’s not the match I dislike.
Games are really fun when it’s full of Platinum players, I’ve actually got a chance to hold my own but seeing Naru hit top 5 last night was impressive, nj if you read this
Wow. He must have been farming lower ranked players all night long.
I was in your group yesterday and added all of us at the end to check rankings.
They had 1751 (no242), 1931 (now no5), 1924, 1379, not sure
We had 1377, 1327, 1621 (me), 1625, not sure
I blocked those 2 to avoid, but they were representing as ‘offline’. So, I had to stop playing as no way I can avoid them. In addition, we duo had to queue for about 4 mins (not the instant pop-up) and lost 10 points for the match.
It should have been more balanced matchmaking first of all or at least less loss in this case given it was an almost guaranteed loss from the start.
everyone should go back and re-evaluate there own play. I do not consider myself to be a “GOOD” player. I was able to make it to PLAT while drunk and blazed on a GS D/W all wells chrono rune necro. Yeah i said GS Necro.
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
This definitely gets a slow clap from me into a standing ovation.
(cheers cheers)
MY GOD THE EPIC BATTLE BETWEEN SILVER PLAYERS AND LEGENDARY WAS A SIGHT TO BEHOLD!!!!!!!!!!!!!!!!!!!!!!!!
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Trying to figure out a way for that matchup to happen during NA primetime, but I’m at a loss. That’s one for Evan to investigate. You may want to make a new thread about this so that it gets more visibility.
I’m pretty sure they’ve seen it and the other games I posted… I Don’t mind these games, gives me something to learn from but I am getting over the design of the system and how I’m supposed to climb, so it’s not the match I dislike.
Games are really fun when it’s full of Platinum players, I’ve actually got a chance to hold my own but seeing Naru hit top 5 last night was impressive, nj if you read this
Wow. He must have been farming lower ranked players all night long.
I was in your group yesterday and added all of us at the end to check rankings.
They had 1751 (no242), 1931 (now no5), 1924, 1379, not sure
We had 1377, 1327, 1621 (me), 1625, not sure
I blocked those 2 to avoid, but they were representing as ‘offline’. So, I had to stop playing as no way I can avoid them. In addition, we duo had to queue for about 4 mins (not the instant pop-up) and lost 10 points for the match.
It should have been more balanced matchmaking first of all or at least less loss in this case given it was an almost guaranteed loss from the start.
Its funny, ive taken a few screen shots myself and i have a few in another thread as well as my own.
Back when we had solo que i use to see the top players all the time but the difference was they were on both teams. Literally 1 game Josre, Nos, Etc Etc players on 1 side and then tarcis, toker on the other and then me.
Its why you saw 65% win rates from the best players in the game.
Now its like 90% win rate gets you into legendary………. makes a whole lot of sense, hip hip hooray!!!!!!!!
I’m trying, I’m just not as good as these players :’( (duo q ranked 19/20)
So hard to climb out of Silver I swear….
Trying to figure out a way for that matchup to happen during NA primetime, but I’m at a loss. That’s one for Evan to investigate. You may want to make a new thread about this so that it gets more visibility.
I’m pretty sure they’ve seen it and the other games I posted… I Don’t mind these games, gives me something to learn from but I am getting over the design of the system and how I’m supposed to climb, so it’s not the match I dislike.
Games are really fun when it’s full of Platinum players, I’ve actually got a chance to hold my own but seeing Naru hit top 5 last night was impressive, nj if you read this
Wow. He must have been farming lower ranked players all night long.
I was in your group yesterday and added all of us at the end to check rankings.
They had 1751 (no242), 1931 (now no5), 1924, 1379, not sure
We had 1377, 1327, 1621 (me), 1625, not sure
I blocked those 2 to avoid, but they were representing as ‘offline’. So, I had to stop playing as no way I can avoid them. In addition, we duo had to queue for about 4 mins (not the instant pop-up) and lost 10 points for the match.
It should have been more balanced matchmaking first of all or at least less loss in this case given it was an almost guaranteed loss from the start.
Its funny, ive taken a few screen shots myself and i have a few in another thread as well as my own.
Back when we had solo que i use to see the top players all the time but the difference was they were on both teams. Literally 1 game Josre, Nos, Etc Etc players on 1 side and then tarcis, toker on the other and then me.
Its why you saw 65% win rates from the best players in the game.
Now its like 90% win rate gets you into legendary………. makes a whole lot of sense, hip hip hooray!!!!!!!!
A lot of people weren’t around when they had the old solo queue but to me its proof that this system can certainly be improve. Once they effectively got rid of synch queueing even the best of the best were around 2/3 win rate. Under this system you have a non negligible amount of people winning like 80% or more of there games. Some of it could be the effect of duoing but there really does seem to be a not insignificant amount of games where the majority of the highest rated player in the match are on one side or the lowest player by a significnat margin doesn’t have a counterpart on the other team. For whatever reason it seems like this mostly benefits the higher ratesd players on average. They need to tighten things up if it means longer queues so be it. I would be perfectly fine waiting 5 min or so for more competitve matches.
There’s actually a pro league NA duo (hint: same guy who throws Helseth’s games on EU) right now who are cheating by having alts “disconnect” or throw on the other team. I don’t doubt that some of the top players are very good and capable of high win rates but there’s a lot of cheating going on in NA which explains some of these win rates.
The other thing is, legendary players can match up vs silver which makes their win rate also very high since there’s no way a silver player is going to beat a legendary duo without having a legendary duo on his own team which doesn’t happen often.
