Glicko. You have to dominate for weeks to move up or send someone down a tier permanently. It took YB a LONG time to crack T1 and they had to utterly dominate T2 to make that happen. As in YB effectively scored more points than the other two T2 servers combined for something like six weeks. Even then it took TC stumbling a bit in T1 for them to move up.
Given the way the matches are setup, Glicko sort of works but unfortunately the entire system is built around stale matches even if most of the time they are more fair (but not really).
Sarcasm For Hire [SFH]
“Youre lips are movin and youre complaining about something thats wingeing.”
Actually, TC and FA were letting DB win so that DB could gain glicko so that DB could stay in T2 as tc and FA want SoS out cuz SoS do nothing but siege hump, and bunker towers. And run from fights the only time they try to fight is when its 20v10 and they are super condi so yeah.
Forever Against Stacked Servers
Virual [VRUS] Alien Lunatics [StFu] Nocturnal Sxaddx [Nuts] Ft. Aspenwood
That which is dead may eternally lie, but with great aeons even death may die.
We knocked SOS out of a t2 spot for last week for the first time in forever. SOS typically loses their t2 matches by a large margin.
DB WON in T2.
SOS moves back up and DB moves back down into ZZZ t3 matchup.
wtf?!
You didn’t knock out SOS at all. You random rolled into the t2 match (less than 5% chance I believe). At the end of the match SOS still has the higher Glicko and better chance of rolling the t2 match……
You all are like 1840 glicko score… They are 1880 something glicko score. TC and FA are still in the 1900s…
So even though FA and TC let you all win, just saying… that one week. You’re own server’s glicko score didn’t improve enough to keep you in T2. You had a big chance of losing T2 at the next match up… And as of right now, its about a 2.8% chance or so of being back in T2 with a matchup of FA and TC.
You should focus on Mag instead of whining.. the only reason you won that matchup is because FA and TC tanked it.
What a joke. Last time we matched with maguuma it was DB 364874 SBI 242625 and Mag with an awful 141222.
We more than doubled their score, and you think we should worry about them because they tryharded the first 3 hours of the week?
SOS hadn’t won their tier in forever [ever?] and we did. Seems like glicko isn’t the right ranking system for such a small pool with stale and tired matchups and winners not showing progress for winning.
Why would anyone bother when you dominate for months and months and when you finally move a tier and actually win it treats you like you lost and puts you back down.
Right now 2.806% chance of getting rolled in T2. Last week it was 3.0 <something> chance and you got lucky to get to T2 early. More than likely your next matchup next week will be Stormbluff Isle and Maguuma.
NOTE: this changes throughout the week so keep checking it, but you won’t get to like 100% like blackgate and T1 servers are in staying in T1 or staying in T2 for some time.
He’s not the only one who has criticized the glicko rating for WvW. It doesn’t reflect changes in population very well. It is very slow to adjust. It requires servers to overstack and essentially destroy two other servers in order to move up, or to completely stop playing and destroy militia morale in order to move down. Not fun for anyone. Anet has the power to change the situation by increasing the variance on the randomized match-up rolls so that they are maybe +/-150 or +/-200 glicko rating instead of +/-100, but they don’t.
Chaba Tangnu
Founding member of [NERF] Fort Engineer and driver for [TLC] The Legion of Charrs
RIP [SIC] Strident Iconoclast
I’m not saying there is an error in the implementation of glicko here…
I’m saying that using glicko is more suited to chess which it was designed for which has a player pool of hundreds of thousands of unknowns matching against each other with no prior histories and where skill level per ‘player’ usually increases at a steady rate.
That’s just not how wvw goes though. WvW ‘player’ skill level varies with lots of factors including population pve events real life events and the like. glicko seeks to reduce variance that can result in too many points being awarded in match ups where there is no prior game history to draw conclusions from and it does a good job at that. It places a lot of weight on old wins and minimizes recent fluctuations.
WvW does not need that correction and should embrace servers pushing for wins and being rewarded for them.
Anet’s glicko system has been fundamentally flawed since release. The entire thing needs to be scrapped, and the new borderland only makes things worse.
I’m sure that’s why DB was double teamed the whole first part of the week. Thank you for the nice welcome to T2 guys.
Nice one!
LOL – DB was not double teamed. ROFL. Delusional. The case is that FA and TC DID want DB to win. The problem is DB doesn’t understand glicko. TC had glicko to lose, so if they would have actually focused continually on TC and not FA, their glicko would have increased. As it is they played stupid and this is what happened. They would have had to have double the score of TC and they didn’t.
As far as SoS, they usually focus FA only as does TC (this can be seen by TC opening an FA keep and then stepping back and allowing SoS to cap a keep while TC keeps FA out from defending,so usually it is a double team on FA, which is boring as all kitten. It was nice to see a real free-for-all for a change.
We knocked SOS out of a t2 spot for last week for the first time in forever. SOS typically loses their t2 matches by a large margin.
DB WON in T2.
SOS moves back up and DB moves back down into ZZZ t3 matchup.
wtf?!
Because there was a very small chance that SoS could drop and DB come up. DB is still rank 7 and SoS is still rank 6. That’s all. It was a T3 server in a T2 matchup. Eventually DB will stay move into T2 permanently. https://leaderboards.guildwars2.com/en/na/wvw
Maybe I’m a thorn in your perfection
A heretic’s voice in your head
A stargazer, releaser
Welcome to Glicko Hell! It’s kind of like Hotel California – just look at Sea of Sorrows. They’ve maintained their T2 standing and before the matching shakeup last week hadn’t won a match in a year! A YEAR!
Welcome to Glicko Hell! It’s kind of like Hotel California – just look at Sea of Sorrows. They’ve maintained their T2 standing and before the matching shakeup last week hadn’t won a match in a year! A YEAR!
Well, glicko is based on how ‘evenly’ a server plays. If one time zone has huge coverage and the other times zones are minimal, the glicko will show this. OCX seems to be big on DB and not so much in other times, so this is part of the reason.
I’m sure that’s why DB was double teamed the whole first part of the week. Thank you for the nice welcome to T2 guys.
Nice one!
LOL – DB was not double teamed. ROFL. Delusional. The case is that FA and TC DID want DB to win. The problem is DB doesn’t understand glicko. TC had glicko to lose, so if they would have actually focused continually on TC and not FA, their glicko would have increased. As it is they played stupid and this is what happened. They would have had to have double the score of TC and they didn’t.
As far as SoS, they usually focus FA only as does TC (this can be seen by TC opening an FA keep and then stepping back and allowing SoS to cap a keep while TC keeps FA out from defending,so usually it is a double team on FA, which is boring as all kitten. It was nice to see a real free-for-all for a change.
No sir, you are delusional to think DB could of moved up so quickly by doubling the score. lol. again, nice one. Then you deflect by bringing up SoS, FA, TC fights, which have no bearing in the discussion.
BTW, I was on TC before I moved to DB, so you’re not sending any cowpatties this way.
Given the way the matches are setup, Glicko sort of works but unfortunately the entire system is built around stale matches even if most of the time they are more fare (but not really).
Not really, it’s not Glicko’s fault that NA tiers are so separated by Glicko, it reflects your stacking behavior quite adequately.
The match-making based on Glicko can generate very varying matches, look at this weeks EU T1-4: 3-1-8, 9-2-4, 10-11-5, 8-7-12
Anet won’t do anything, because one of us is going to implode in a month or two anyways.
Won’t say which one.
/me runs out the door.
This PvF brought to you by doors™
You’ll just have to rebuild errrrr or who ever implodes first that is.. DB went through the same thing but they are moving back up it looks. I remember golem week fighting them down in Bronze or bottom of silver. They probably get more new players because they have Dragon in their name lol
So you take it easy the week you get new blood in your match up and play hard when you are back to the same servers you’ve been facing? LOL ok
Yes that is exactly it. Sos has been the weak link in t2 for a long time, Fa/Tc liked fighting db so they both didnt hit you guys much so you can gain glicko and stay in t2. Since sos is in t2 again this week, they have to hit them super hard to make them lose glicko and bring you back up.
I’m sure that’s why DB was double teamed the whole first part of the week. Thank you for the nice welcome to T2 guys.
Nice one!
LOL – DB was not double teamed. ROFL. Delusional. The case is that FA and TC DID want DB to win. The problem is DB doesn’t understand glicko. TC had glicko to lose, so if they would have actually focused continually on TC and not FA, their glicko would have increased. As it is they played stupid and this is what happened. They would have had to have double the score of TC and they didn’t.
As far as SoS, they usually focus FA only as does TC (this can be seen by TC opening an FA keep and then stepping back and allowing SoS to cap a keep while TC keeps FA out from defending,so usually it is a double team on FA, which is boring as all kitten. It was nice to see a real free-for-all for a change.
No sir, you are delusional to think DB could of moved up so quickly by doubling the score. lol. again, nice one. Then you deflect by bringing up SoS, FA, TC fights, which have no bearing in the discussion.
BTW, I was on TC before I moved to DB, so you’re not sending any cowpatties this way.
I was also on TC – moved to FA. Y’all were so much better than SoS. Understand how Glicko works…..
Since sos is in t2 again this week, they have to hit them super hard to make them lose glicko and bring you back up.
I hope they listen to Buffy. I was ready to go back to T2, dumped all my gold for it. T1 is good and all but the politics and blobbing is still strong with the old guard. Why can’t they just go for fights and fun like YB. ((((
Not really, it’s not Glicko’s fault that NA tiers are so separated by Glicko, it reflects your stacking behavior quite adequately.
The match-making based on Glicko can generate very varying matches, look at this weeks EU T1-4: 3-1-8, 9-2-4, 10-11-5, 8-7-12
Do you know the populations of NA servers, or you just making up bullkitten? The way anet’s glicko system works, its nearly impossible for glicko to transfer between tiers. Look up the history of NA servers: T1 servers have had 2100+ rating since mos launch, T2 is generally around 1900, 8th ranked server has a big glicko drop from 6th, with 7th somewhere inbetween. Do you think NA somehow perfectly maintains this ratio of stacked servers? Or is it more likely that the glicko system is fundamentally flawed for wvw?
Not really, it’s not Glicko’s fault that NA tiers are so separated by Glicko, it reflects your stacking behavior quite adequately.
The match-making based on Glicko can generate very varying matches, look at this weeks EU T1-4: 3-1-8, 9-2-4, 10-11-5, 8-7-12
Do you know the populations of NA servers, or you just making up bullkitten? The way anet’s glicko system works, its nearly impossible for glicko to transfer between tiers. Look up the history of NA servers: T1 servers have had 2100+ rating since mos launch, T2 is generally around 1900, 8th ranked server has a big glicko drop from 6th, with 7th somewhere inbetween. Do you think NA somehow perfectly maintains this ratio of stacked servers? Or is it more likely that the glicko system is fundamentally flawed for wvw?
I’m thinking you don’t know the history.
Did you know that Eredon Terrace was No.1 at launch?
If glicko were reflective of population half these servers would drop like a rock. Instead these cancers called “push weeks” are possible. Can be a no-show for weeks and finally when glicko starts to wake up you just PvD a little.
No it’s slow and bad. Takes all the meaning out of winning. Again the server that won should not drop a tier.
If glicko were reflective of population half these servers would drop like a rock. Instead these cancers called “push weeks” are possible. Can be a no-show for weeks and finally when glicko starts to wake up you just PvD a little.
No it’s slow and bad. Takes all the meaning out of winning. Again the server that won should not drop a tier.
I guess you really don’t understand Glicko. That is only ONE SMALL part of it (and that is a mod that A.Net did to the Glicko formula). The other part is how well the server population performs and how even that performance is. If Server population were the only criteria, the smaller servers would be higher tiers than they are.
I guess you really don’t understand Glicko. That is only ONE SMALL part of it (and that is a mod that A.Net did to the Glicko formula). The other part is how well the server population performs and how even that performance is. If Server population were the only criteria, the smaller servers would be higher tiers than they are.
I guess you don’t understand population and timezone coverage. I’ve tried the smaller, low tier servers and they would not be in higher tiers at all. They hardly even muster a fight during NA prime. You would not see a dead server winning. If it went by winners move upwards and 3rd place risks downwards we would get better matchups.
Not really, it’s not Glicko’s fault that NA tiers are so separated by Glicko, it reflects your stacking behavior quite adequately.
The match-making based on Glicko can generate very varying matches, look at this weeks EU T1-4: 3-1-8, 9-2-4, 10-11-5, 8-7-12
Do you know the populations of NA servers, or you just making up bullkitten? The way anet’s glicko system works, its nearly impossible for glicko to transfer between tiers. Look up the history of NA servers: T1 servers have had 2100+ rating since mos launch, T2 is generally around 1900, 8th ranked server has a big glicko drop from 6th, with 7th somewhere inbetween. Do you think NA somehow perfectly maintains this ratio of stacked servers? Or is it more likely that the glicko system is fundamentally flawed for wvw?
You have one good point: I don’t know the current NA-population distribution. It may indeed be the case that not one is left in T1, but none noticed as it isn’t tested since a while. The last test of T1 vs T2 was last Sept. (week 35/2015, http://mos.millenium.org/na/matchups/history/313) where Blackgate played in T2. At that time the 280pts distance between Blackgate and SoS and FA got confirmed.
For Glicko and my outsiders correct estimation of NAs server (manpower/coverage) strength, it must be tested regularly. Currently it is tested only seldom. However whenever (especially during the seasons) it was tested in the past, the Glicko rating turned out to be quite adequate, and the matches to be very single-sided. E.g. the last season in Sept./Oct. 2014: http://mos.millenium.org/na/matchups/history/213
So yes, to be really adequat Glicko has to be empirical tested by matches. However these tests were quite painful for the involved server in the past, as Glicko was mostly right and the matches very single-sided. The forum was full of cries during the seasons and when the random-modifier was higher as it is today (till season 2, if I remember right).
The last test of T1 vs T2 was last Sept. (week 35/2015, http://mos.millenium.org/na/matchups/history/313) where Blackgate played in T2. At that time the 280pts distance between Blackgate and SoS and FA got confirmed.
Dude the last test of T1 vs T2 was TC….
They were the weak link and now their stuck down in T2. Blackgate was actually winning their matchups when glicko went and put the winner down a tier. That’s glicko for you……..
The last test of T1 vs T2 was last Sept. (week 35/2015, http://mos.millenium.org/na/matchups/history/313) where Blackgate played in T2. At that time the 280pts distance between Blackgate and SoS and FA got confirmed.
Not the best example. I wasn’t active at this point in time, but I can guarantee you BG pushed hard in order to stay in t1, while SoS and FA tanked because they definitely didn’t want BG stuck in t2. Week 36 would be a better example, but it probably has problems too.
Blackgate and YB performed better in these tests then TC, that’s why TC stayed in T2.
Yep and it only took months of demoralization and people quitting in frustration to find out the T2 winner should move up and the T1 loser should move down. Thanks glicko!
That’s more a problem of tiers consisting of exactly 3 servers. If the server do not fit this fixed structure, e.g. there are 4 “T1” server and only 2 “T2” server then this is unavoidable.
Yep and it only took months of demoralization and people quitting in frustration to find out the T2 winner should move up and the T1 loser should move down. Thanks glicko!
What’s the alternative? 1-2 demoralizing weeks per month!
or
More stringent balance-enforcement, but cries in the forum were very loud as well, when the system hindered you 2 month ago to further overstack T1
Oh oh. Do you remember season 1? Glicko and Swiss Tournament are definitely better than people. I’ll take holiday from the forum when ANet make matches manually and the people know that.
