Guild Wars 2

Seems a lot of people don’t understand Glicko or the concept behind it.

Even though your server may be winning or losing by a large margin currenty, that does not mean you are gaining or losing points necessarily.

A lot of the servers in this thread complaining about being dominated are actually gaining a large number of points because they are doing a lot better than Glicko expects them to do.

That is the beauty of mixing the matchups up a little. To see how your server performs against a different higher or lower rated server and if you perform better or worse than expected than you will rise or fall accordingly.

As of this post, SOS is losing their tier but gaining a hefty 63 points to their rating. Because they are performing a LOT better than Glicko expects them to against those two higher rated servers.

Snowreap.5174:

Your rating doesn’t change based on whether you win or lose, it changes based on whether you beat expectations or not. Based on the difference between your rating and your opponent’s rating, Glicko-2 calculates an ‘E’ value, which is the score you are expected to get against that opponent. Your actual score (designated ‘s’) is compared to that — if s is lower than E you will lose rating points (even if you’re still winning the match), and if s is higher than E you will gain rating points (even if you’re losing the match).

If you want to calculate exactly how many points you need in order to avoid losing rating points, you need to first calculate E. The procedure to do this is described here:
http://www.glicko.net/glicko/glicko2.pdf

The formula to calculate s has been posted by ArenaNet in the past, and it works out to:
s = ( sin((x/(x+y)-0.5)*pi) + 1 ) / 2, where x is your in-game score and y is your opponent’s in-game score.

Setting s = E, and solving for x, yields:

x = y * (0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)

In other words, your score must be [(0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)] times as high as your opponent’s score in order to avoid losing rating points.

You have to do this calculation separately (including the calculation of E) for each of your two opponents.

-ken

WIll we see a reduction in the overall range of rating (i.e. between SoR and ET in NA)? I would guess that when the “T1” people face off against the T2 (or lower) servers, the T1 servers ratings will reduce quite a bit as there is a 150 point gap between JQ and DB at the moment (and without doing the maths I would guess that does not represent the current relative strength of those servers). ET and FC have both risen in rating this week.

Yeah, ANet should really add he current rating evolution to the match-display at B!
Probably not everyone has millenium.org open at 2nd screen

It would produce a lot of
- Ok, score isn’t good, but hey we do great!
- Oh damm, this score is still not sufficient to keep our rating!

We’re having fun and should’ve been in T3 anyway. With the new matching system, that means that most weeks we’ll be facing off against other T3ish worlds like Mag, YB and KN. Some weeks we’ll be facing off against T2 worlds. Some weeks we’ll be facing off against T4 worlds. Rarely we’ll encounter T1 or T5 worlds. The system is a good one and is working as intended. We’re currently projected to be top of the group we’d normally be paired up with that includes worlds that are only on par with CD – a world that we regularly beat by 90-100k points. This is just sensible.

Honestly, I just wish folks would recognise that just because we’re not winning and we’re not going to win (although, hell, even second place isn’t that far off), it’s not the end of the world. Hell, if all of the people being negative about the match actually spent that time playing and having fun, we’d probably be running dead evens on score. A number of folks seemed to write off the match on the basis of the first few hours, despite everyone on the world knowing how poorly we do on reset nights

I think for the most part things are going to be a love-hate relationship with the new system.

There are positives: As someone earlier mentioned, we see many low ranked servers doing much better than expected against their higher tier adversaries. While they aren’t winning, Glicko is properly adjusting their ratings (servers like FC) to reflect more accurately server WvW strength, which will make ratings a lot more accurate a tool to use to predict how match-up will go, assuming things work out consistently. Servers like Kaineng logically SHOULD lose rating when lower ranked servers do much better than expected against them, especially since ratings like those were gained from numbers that they may not necessarily possess currently. (With many having transferred of Kain to my understanding.) So that’s working as intended.

Furthermore you also have the benefit of meeting new servers in battle and making new acquaintances, and getting out of old stale match-ups where the flow of battle and outcome is generally for most deadlocked tiers, predictable until the strongest server finally moves up. The excitement of this alone is a god send to many.

Now as for the negatives, they are pretty severe. You have match-ups like the current tier 8 match-up with SF being paired with HoD and ET. We of FC and ET got to see SF for months, and that they were a tier 6 server was a no-brainer. Seeing them thrown back somewhere where they proved for months that they don’t belong is.. unfortunate, to say the least.

The system seems to be assuming that based on scores and rating, matchup history alone (which is not really a completely accurate measure to go by) that HoD and ET together might stand a chance against SF of at least not being uber-dominated. Unfortunately while HoD is formidable, ET at this time has a terrible numbers issue that isn’t necessarily reflected in ratings and match-up history of their previous opponents and simply won’t be able to keep SF from simultaneously flooring both servers. Issues like that may be corrected overtime as the system acquires more updated data though, we will have to see.

Situations like tier 8 will be happening in other tiers also, and that’s the unforunate weak point of this system.

Another is that hard-earned ratings that some servers were fighting tooth and nail to reach a proper matchup for weeks-months are no longer having the gravity they used to in determining placement, which will upset some. I think EB is in that boat in particular.

Probably this whole package will be received with mixed feelings across the board. For me, from what I’ve seen so far its been enjoyable, but I certainly wouldn’t of been saying that if it was us down with SF right now, so meh.

Detharos.3157:

The system seems to be assuming …

The system assumes only one thing:

Every server should be able to prove the correctness of his rating against any server within a Rating-Area of +/-400 from time to time.

The more servers are ranked in between 2 servers and the more their rating differs the less likely (and therefore less often) they meet each other.

That’s it! not more and not less.

Mad Rasputin.7809:

Seems a lot of people don’t understand Glicko or the concept behind it.

Even though your server may be winning or losing by a large margin currenty, that does not mean you are gaining or losing points necessarily.

A lot of the servers in this thread complaining about being dominated are actually gaining a large number of points because they are doing a lot better than Glicko expects them to do.

That is the beauty of mixing the matchups up a little. To see how your server performs against a different higher or lower rated server and if you perform better or worse than expected than you will rise or fall accordingly.

As of this post, SOS is losing their tier but gaining a hefty 63 points to their rating. Because they are performing a LOT better than Glicko expects them to against those two higher rated servers.

The problem is servers ranked similarly to TC and FA are gaining rating from much easier match ups and this system in that is pretty much punishing TC and FA because of RNG and that is complete kitten.

I think this is hitting a smaller percentage of SOS harder. Our NA players like myself are getting torched. Some of us enjoy the “see what we can do before getting run over” part of the game,but I do understand people just walking away also. Hopefully we don’t loose too many people in the long run and the system shows balance for us down the road.
The idea of showing how our server is actually gaining or loosing in this system in game would help imo

Oh, I agree.

It was so much better when the old system locked you in to a monotonous matchup for two months.

One week is far too little time to truly understand the hopelessness of a one-sided match.

Asudementio.8526:

The problem is servers ranked similarly to TC and FA are gaining rating from much easier match ups and this system in that is pretty much punishing TC and FA because of RNG and that is complete kitten.

9 month of wrong match-ups do not recover within 1 week

Dayra.7405:

Asudementio.8526:

The problem is servers ranked similarly to TC and FA are gaining rating from much easier match ups and this system in that is pretty much punishing TC and FA because of RNG and that is complete kitten.

9 month of wrong match-ups do not recover within 1 week

You’re right best to just keep adding more and more kitten problems with the way match-ups are determined. Great idea.

http://mos.millenium.org/eu/matchups

Just look at this and see the matchup system was a mistake. 4 matchup’s are going fine,the rest looks terrible.Also Rof must be having a lot of fun playing vs all those deers and evil rabbits they meet on the battlefield..

Sovereign.1093:

Winning makes or break people. This guy isn’t used to losing; so the system isn’t making him but breaking him. He just needs to learn that it is a part of wvw to win and lose some of the time.

Or all of the time depending on the server.

Slashco.3927:

Title says it all. I know I enjoyed my old server matchup in T4. Yes, we won every week in points because of a strong oceanic presence, but in North American prime time, we’d have even numbers and sometimes be outmanned. Was fun. Going up 2 tiers on a “random” multiplier is the worst idea in the history of gaming.

No fun = no more GEM $ for Anet from me.

Sounds like the can’t win = no fun bus showed up.

Dayra.7405:

@Roo Stercogburn: Looks like you did not understood the new system.
So far Gunnars hold is the match winner with +56 rating points, Riverside is second with +23 points and Piken the looser of the match with -72pts.

Rating is what you are fighting for! Rating is the thing that remains with your server when the match is over. GH currently fights over expectation, and that’s why you win most.

And this is nice in the new system! It has different challenges for you. It can be either
- Defend your rating by dominating a weaker server (Piken fails to do so currently)
- win through in an even match (Riverside does this vs Piken)
- Resist the domination of a stronger server (GH’s does this well at moment)

In fact you can have every mixture of the above as you have 3 sides in the match.

Both scores matter. It hard to motivate anyone except the hard core on increases in Glicko.

hex.3218:

Slashco.3927:

Title says it all. I know I enjoyed my old server matchup in T4. Yes, we won every week in points because of a strong oceanic presence, but in North American prime time, we’d have even numbers and sometimes be outmanned. Was fun. Going up 2 tiers on a “random” multiplier is the worst idea in the history of gaming.

No fun = no more GEM $ for Anet from me.

Sounds like the can’t win = no fun bus showed up.

It is hard for a server to stay motivated when they are beaing badly beat. Losing is not fun anyway you look at it.

At the end of the day a player shouldn’t need to go research Glicko system statistical data or spend any time calculating a servers progress. They should be able to just log into a game and have a good time. The rest is on the devs shoulders. For some people this new system worked. For others, such as myself, it didn’t. I’m still in the same kittenty match up I’ve been in for two months. And given the history of what I’ve seen in WvW in the past several months, I dont expect much difference next week. Even if we do have a new matchup. This rut of a matchup for the past two months and this weeks unluck roll of the new system has certainly taken then wind out of a lot of Anvil Rock players sails. I fear it’s just an avalanche effect now. I’ll be surprised if AR bounces back anytime soon.

Tenebrous.2451:

It is hard for a server to stay motivated when they are beaing badly beat. Losing is not fun anyway you look at it.

SoS is losing badly in T2, but they’re bringing the fight like crazy. They have an indomitable spirit. A lot of servers out there fight harder when they’re losing, and have fun doing it.

Aye, I couldn’t care less about Glicko. I want fights which I may lose but have a chance to win. I want overarching strategy in which we can make choices, take risks and have a chance of winning. Might lose territory but in a match up to at least have a chance of seeing a difference.

I don’t give a kitten about some meta rating that spreadsheet-lovers can get excited about if my in-game experience sucks like a nuclear hoover.

Snowreap.5174:

Dodforer.2470:

It’s not that it’s the worst idea ever, just parts of it don’t make sense at all.

Looking at http://mos.millenium.org/na/ right now, Yaks is ahead by 10-12k, yet we’re losing 3 points in evolution, CD is almost 12K down and they’re UP 60 points in evolution….how the hell are we winning yet still losing, and how is CD losing and yet still gaining so much?

FA and TC have dropped 19-23 points, while SoS has gone up almost 48, and they’re getting hammered. DB is up almost 30, Mag is down almost 69, and SBI is up almost 47, and again SBI is getting hammered.

[…]

Maybe math has changed in the last 3 decades since I was in school, but the numbers don’t make sense at all. It looks like teams that are getting beaten are actually moving up, teams that are winning are moving down and you’re going to end up with T1 servers fighting T5 servers because they’re being brought into parity range through this system. I’m not saying these servers are bad or other servers are better, sometimes it’s a matter of population and coverage, but if the intent is to make matches like this happen more frequently I don’t see how it will benefit WvW overall.

Your rating doesn’t change based on whether you win or lose, it changes based on whether you beat expectations or not. Based on the difference between your rating and your opponent’s rating, Glicko-2 calculates an ‘E’ value, which is the score you are expected to get against that opponent. Your actual score (designated ‘s’) is compared to that — if s is lower than E you will lose rating points (even if you’re still winning the match), and if s is higher than E you will gain rating points (even if you’re losing the match).

If you want to calculate exactly how many points you need in order to avoid losing rating points, you need to first calculate E. The procedure to do this is described here:
http://www.glicko.net/glicko/glicko2.pdf

The formula to calculate s has been posted by ArenaNet in the past, and it works out to:
s = ( sin((x/(x+y)-0.5)*pi) + 1 ) / 2, where x is your in-game score and y is your opponent’s in-game score.

Setting s = E, and solving for x, yields:

x = y * (0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)

In other words, your score must be [(0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)] times as high as your opponent’s score in order to avoid losing rating points.

You have to do this calculation separately (including the calculation of E) for each of your two opponents.

-ken

This math is completely sound, but what it doesn’t take into account is that the Glicko system depends on a few “givens” that are not happening in this situation.

1. New opponents with the base score entering the system.

2. As with ratings in sports like baseball, there is a fixed team, or at most, very little inter-match player movement. (Free server transfers anybody?)

3. An adequate fall off for past scores influence on the current match ups, accounting for the server movement discussed above.

4. A larger population of opponents, the Glicko system gets more accurate with a larger population which is necessary for the randomized matches.

5. The stagnation from the above limited opponent pool leads to increases in players on the winning teams and decreases on the losing teams, which even if they stayed static as they were during the initial matching, would further depress the losers cumulative score.

6. An actual fixed team size or reasonable opponent pool. When you are outmanned to the extent some servers are, there is no amount of talent, capability or coordination to will make up for superior numbers, especially with the “snowball” effects of systems like WxP.

You play chess against one person, start with the same number of pieces and the winners don’t get extra pawns.

I am all in favor of scaling the outmanned buff, because right now, it just looks like a dinner bell for Golem rushes.

Hickeroar.9734:

Tenebrous.2451:

It is hard for a server to stay motivated when they are beaing badly beat. Losing is not fun anyway you look at it.

SoS is losing badly in T2, but they’re bringing the fight like crazy. They have an indomitable spirit. A lot of servers out there fight harder when they’re losing, and have fun doing it.

You can make that case for almost all servers at some level because they have some level of dedicated WvW players. The lower tier servers tend to have more marginal players where they show up if they think they can help us win, often because the PvE scene on those servers is pretty robust.

I think some people misunderstood what I was trying to say. I don’t care at all about rankings and ratings. I don’t care that SOS is “performing better than we are expected to have performed”. I care about having fairly evenly balanced fights (at least numerically) during North American prime time. The “better than expected performance” on ratings is really due to our Oceanic presence.

I know that before this update, in NA primetime, SOS would have the outmanned buffed at times on all 4 maps (so we weren’t dominating based upon numbers); hence, we weren’t just facerolling CD and SBI as some have suggested. For anyone that was on SOS/SBI/CD, they knew there was often roaming groups of 5-15 in all 3 BLs. Zergs did happen from time to time, but they were often confined to EB. Now, our T2 opponents field superior numbers to SOS in all BLs and we get steamrolled.

For those on SBI and CD that hated the fact that SOS would win every week due to oceanic presence, I understand that and don’t begrudge them wanting something to change. But, at least they didn’t get constantly steamrolled during the time they actual play.

I play for fun fights – not for score. That’s why I won’t be supporting Anet any longer. It just isn’t fun anymore.

Slashco.3927:

I think some people misunderstood what I was trying to say. I don’t care at all about rankings and ratings. I don’t care that SOS is “performing better than we are expected to have performed”. I care about having fairly evenly balanced fights (at least numerically) during North American prime time. The “better than expected performance” on ratings is really due to our Oceanic presence.

I know that before this update, in NA primetime, SOS would have the outmanned buffed at times on all 4 maps (so we weren’t dominating based upon numbers); hence, we weren’t just facerolling CD and SBI as some have suggested. For anyone that was on SOS/SBI/CD, they knew there was often roaming groups of 5-15 in all 3 BLs. Zergs did happen from time to time, but they were often confined to EB. Now, our T2 opponents field superior numbers to SOS in all BLs and we get steamrolled.

For those on SBI and CD that hated the fact that SOS would win every week due to oceanic presence, I understand that and don’t begrudge them wanting something to change. But, at least they didn’t get constantly steamrolled during the time they actual play.

I play for fun fights – not for score. That’s why I won’t be supporting Anet any longer. It just isn’t fun anymore.

You’re earning 400+ points during prime time on your server. I still don’t understand how you think you are getting steamrolled. Do you know that to get steamrolled, you have to be incapable of holding 2/3rds of the map by yourself?

If you really can’t think of anythin to tell your guild, why don’t you try this? http://www.youtube.com/watch?v=UXLVFnl3WcE

The only real match that should not have happened is DB/Mag/SBI. I am not saying this due to the fact I am on SBI I am saying this based upon the evolution. DB is going to be Top Tier 2 and the rating differance is too large I honestly think that DB has a shot at T1. Luckely Their increased rating will make this less likely to happen in the future. It will take some time for the rating to become what they should be. I think I will give it a few months. Really I should give it 30 weeks.

As for getting steamrolled. Prime time when I play we have put up some vary good fights against Dragonband. Our Evolution score agrees with that. The fact that the matchup thread has pepole saying that also agrees. I am actually having fun with it. Fighting SOS was just annoying and getting old fast. Mostly because that fight was lost due to population as diffrent times. If SOS was mosly NA preseance who knows what the outcome could have been. The map got wiped at night and then both SBI and CD came over and took back our lands durring the day with CD and SBI fighting after that was done. With this we get wiped at night and have to fight to keep it durring the day.

Slashco.3927:

For those on SBI and CD that hated the fact that SOS would win every week due to oceanic presence, I understand that and don’t begrudge them wanting something to change. But, at least they didn’t get constantly steamrolled during the time they actual play.

I prefer a quick painless loss with no hope rather then slow loss due to never meeting the majority of you. I am sure many would agree with me on this.

I’ll agree. the new system made things worse. Unf teh way PlayNc and Anet work is they will let it go on rather than admit mistake and shove false praise down our throats telling us its what we wanted…

I honestly think all they need to do is tweak it. I would be in favor of a hard cap put in place so you could not face a server X amount of rating points ahead or behind you.

And I’m on SBI. Sure we are getting crushed in score. And I feel for our non-NA primetime players. But the NA primetime has been awesome. There’s actual strategy going on. I can’t wait to get home tonight to log in.

I feel worse for EB – all set to move up to a competitive tier and bam, stuck in the same old matchup. But this time its likely only for a week. That is why I favor this matchmaking method (with a tweak), you only have to deal with a bad matchup for a week. And then you get to move on and have face new and different opponents. Much more fun that way.

pulsecodesgnl.3470:

I disagree. I enjoy the new matchups even though we were temporarily thrown back down in t8 via the random element. Glicko2 is finally going to be able to do what it was intended to do in the first place, and build REAL scores that should’ve been built since the system had started. It’s clearly evident by the scores that the tiers are not nearly as rigid as they appeared to be which is great.

Yes, there is a HUMONGOUS gap for tiers 1 and 2, and unfortunately it will linger for a couple weeks and will wholly allow for unbalanced matches; but good hell, why do you deserve preferential treatment over the other 6 tiers?

Deal with it, quit crying, if there is an actual outbalanced problem after the scores start to settle from the neccessary random variable matches for Glicko2, THEN Anet will adjust the random variable at that time. There are more tiers than your own Tier 1 and 2, quit acting like spoiled brats.

Says the T6 roflstomping 2 T8 servers.

Say what you like, but the new server matchups mean I will now finally get a chance to pound DB

I think that it will work out in the long run. I can see how there would be some really good data after 2-3 weeks with this new system. For example, Sea of Sorrow’s isnt nearly ready for T2, but judging by their performance in T2, they’ll do fine in T3 instead of T4 where they used to be.

lunyboy.8672:

pulsecodesgnl.3470:

I disagree. I enjoy the new matchups even though we were temporarily thrown back down in t8 via the random element. Glicko2 is finally going to be able to do what it was intended to do in the first place, and build REAL scores that should’ve been built since the system had started. It’s clearly evident by the scores that the tiers are not nearly as rigid as they appeared to be which is great.

Yes, there is a HUMONGOUS gap for tiers 1 and 2, and unfortunately it will linger for a couple weeks and will wholly allow for unbalanced matches; but good hell, why do you deserve preferential treatment over the other 6 tiers?

Deal with it, quit crying, if there is an actual outbalanced problem after the scores start to settle from the neccessary random variable matches for Glicko2, THEN Anet will adjust the random variable at that time. There are more tiers than your own Tier 1 and 2, quit acting like spoiled brats.

Says the T6 roflstomping 2 T8 servers.

I deleted my comment after I realized that there was quite a few t1/t2 posts in here agreeing with me. That being said, so what? Winning != fun. You realize that Sorrows Furnace was stuck in the Glicko Ghetto of being trapped in t8 for 3 months right?

And yeah, we were winning, but it was boring as all hell. This week has been nice, it has been a refresher from the coverage wars of t6, and go check out our weekly thread and how much everyone is enjoying it.

If you really think you need to win order to enjoy WvW you are going to be sorely dissapointed time after time. There is a large majority of our core players on SF that prefer when we are actually losing so it gives us something to do, something different than the mindnumbing winning with a zombie horde in t8/t7. But at the same time it’s nice to have that kind of a matchup where things are less intense as well.

Johje Holan.4607:

I honestly think all they need to do is tweak it. I would be in favor of a hard cap put in place so you could not face a server X amount of rating points ahead or behind you.

And I’m on SBI. Sure we are getting crushed in score. And I feel for our non-NA primetime players. But the NA primetime has been awesome. There’s actual strategy going on. I can’t wait to get home tonight to log in.

SBI (rank 13) is doing spectacularly well. They’re only 6.5K points behind #7 Maguuma, despite Mag being rated almost 300 points higher. at this rate, SBI will easily be back in the #12 spot in the rankings next week.

SBI may not be in the same class as DB, but so far they’re showing that they can be competitive vs. Mag. If Maguuma wants to stop bleeding so many rating points, they are going to have to devote more attention to keeping SBI down (although I’m sure DB is going to make that hard for them to do).

-ken

Snowreap.5174:

Johje Holan.4607:

I honestly think all they need to do is tweak it. I would be in favor of a hard cap put in place so you could not face a server X amount of rating points ahead or behind you.

And I’m on SBI. Sure we are getting crushed in score. And I feel for our non-NA primetime players. But the NA primetime has been awesome. There’s actual strategy going on. I can’t wait to get home tonight to log in.

SBI (rank 13) is doing spectacularly well. They’re only 6.5K points behind #7 Maguuma, despite Mag being rated almost 300 points higher. at this rate, SBI will easily be back in the #12 spot in the rankings next week.

SBI may not be in the same class as DB, but so far they’re showing that they can be competitive vs. Mag. If Maguuma wants to stop bleeding so many rating points, they are going to have to devote more attention to keeping SBI down (although I’m sure DB is going to make that hard for them to do).

-ken

At the moment I think SBI and Mag are more worried about bleeding points from DB. They have plenty to share with us. If Mag fights SBI and ignores DB then they will not get the same amount of points then if they take all the points from DB.

I think however that the fact that SBI vs Mag vs DB is the only match that is completely out of whack just shows you that the new system is working,.

pulsecodesgnl.3470:

lunyboy.8672:

pulsecodesgnl.3470:

I disagree. I enjoy the new matchups even though we were temporarily thrown back down in t8 via the random element. Glicko2 is finally going to be able to do what it was intended to do in the first place, and build REAL scores that should’ve been built since the system had started. It’s clearly evident by the scores that the tiers are not nearly as rigid as they appeared to be which is great.

Yes, there is a HUMONGOUS gap for tiers 1 and 2, and unfortunately it will linger for a couple weeks and will wholly allow for unbalanced matches; but good hell, why do you deserve preferential treatment over the other 6 tiers?

Deal with it, quit crying, if there is an actual outbalanced problem after the scores start to settle from the neccessary random variable matches for Glicko2, THEN Anet will adjust the random variable at that time. There are more tiers than your own Tier 1 and 2, quit acting like spoiled brats.

Says the T6 roflstomping 2 T8 servers.

I deleted my comment after I realized that there was quite a few t1/t2 posts in here agreeing with me. That being said, so what? Winning != fun. You realize that Sorrows Furnace was stuck in the Glicko Ghetto of being trapped in t8 for 3 months right?

And yeah, we were winning, but it was boring as all hell. This week has been nice, it has been a refresher from the coverage wars of t6, and go check out our weekly thread and how much everyone is enjoying it.

If you really think you need to win order to enjoy WvW you are going to be sorely dissapointed time after time. There is a large majority of our core players on SF that prefer when we are actually losing so it gives us something to do, something different than the mindnumbing winning with a zombie horde in t8/t7. But at the same time it’s nice to have that kind of a matchup where things are less intense as well.

My apologies, I misread your “quit crying” comments as being pointed at everyone in the thread, not just the upper tiers.

I have been enjoying the match up despite the spawn camping, corpse jumping, and general classiness I have come to expect.

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

lunyboy.8672:

Snowreap.5174:

Dodforer.2470:

It’s not that it’s the worst idea ever, just parts of it don’t make sense at all.

Looking at http://mos.millenium.org/na/ right now, Yaks is ahead by 10-12k, yet we’re losing 3 points in evolution, CD is almost 12K down and they’re UP 60 points in evolution….how the hell are we winning yet still losing, and how is CD losing and yet still gaining so much?

FA and TC have dropped 19-23 points, while SoS has gone up almost 48, and they’re getting hammered. DB is up almost 30, Mag is down almost 69, and SBI is up almost 47, and again SBI is getting hammered.

[…]

Maybe math has changed in the last 3 decades since I was in school, but the numbers don’t make sense at all. It looks like teams that are getting beaten are actually moving up, teams that are winning are moving down and you’re going to end up with T1 servers fighting T5 servers because they’re being brought into parity range through this system. I’m not saying these servers are bad or other servers are better, sometimes it’s a matter of population and coverage, but if the intent is to make matches like this happen more frequently I don’t see how it will benefit WvW overall.

Your rating doesn’t change based on whether you win or lose, it changes based on whether you beat expectations or not. Based on the difference between your rating and your opponent’s rating, Glicko-2 calculates an ‘E’ value, which is the score you are expected to get against that opponent. Your actual score (designated ‘s’) is compared to that — if s is lower than E you will lose rating points (even if you’re still winning the match), and if s is higher than E you will gain rating points (even if you’re losing the match).

If you want to calculate exactly how many points you need in order to avoid losing rating points, you need to first calculate E. The procedure to do this is described here:
http://www.glicko.net/glicko/glicko2.pdf

The formula to calculate s has been posted by ArenaNet in the past, and it works out to:
s = ( sin((x/(x+y)-0.5)*pi) + 1 ) / 2, where x is your in-game score and y is your opponent’s in-game score.

Setting s = E, and solving for x, yields:

x = y * (0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)

In other words, your score must be [(0.5 + arcsin(2*E-1)/pi) / (0.5 – arcsin(2*E-1)/pi)] times as high as your opponent’s score in order to avoid losing rating points.

You have to do this calculation separately (including the calculation of E) for each of your two opponents.

-ken

This math is completely sound, but what it doesn’t take into account is that the Glicko system depends on a few “givens” that are not happening in this situation.

1. New opponents with the base score entering the system.

2. As with ratings in sports like baseball, there is a fixed team, or at most, very little inter-match player movement. (Free server transfers anybody?)

3. An adequate fall off for past scores influence on the current match ups, accounting for the server movement discussed above.

4. A larger population of opponents, the Glicko system gets more accurate with a larger population which is necessary for the randomized matches.

5. The stagnation from the above limited opponent pool leads to increases in players on the winning teams and decreases on the losing teams, which even if they stayed static as they were during the initial matching, would further depress the losers cumulative score.

6. An actual fixed team size or reasonable opponent pool. When you are outmanned to the extent some servers are, there is no amount of talent, capability or coordination to will make up for superior numbers, especially with the “snowball” effects of systems like WxP.

You play chess against one person, start with the same number of pieces and the winners don’t get extra pawns.

I am all in favor of scaling the outmanned buff, because right now, it just looks like a dinner bell for Golem rushes.

You forgot one more,

Glicko also expects all three servers to fight each other equally. But there are times when two servers will join in a loose alliance against the third, producing distorted results in ratings.

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

Congrats at missing the point, and adding obliviousness to an already obfuscated thread. Performance > Winning. If you need to win to enjoy WvW you should probably quit while you are ahead.

ykyk.2740:

I don’t know what the hell matchup you are playing but SoS are giving us some of the best fights I’ve seen in awhile. The ppt differential is because SoS can’t field all maps during NA prime, but you guys more than hold your own on the maps you have a presence on, and it’s showing in the rating gain.

I concur, SoS is kickin butt out there. The only thing holding them back is coverage, which to me, should not be the deciding factor in how well one server does. I like the PPT system, but I don’t think it should be the primary way to get points.. I only think it should be a small factor.

How might rankings look if there was no PPT system? How might they look if points were only awarded through kills (not from siege kills), and the capturing/defending of towers/keeps/camps?

I’ve seen many instances of small groups not only taking out larger groups, but also taking and defending towers and keeps.

pulsecodesgnl.3470:

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

Congrats at missing the point, and adding obliviousness to an already obfuscated thread. Performance > Winning. If you need to win to enjoy WvW you should probably quit while you are ahead.

Getting Crushed in a match and gaining points isn’t “winning”

I’m on SBI, this matchup isn’t anymore “fun” then the previous matchup of getting zerged all day long.

People like yourself who believe that this “point” scale is good for the game completely miss the point of actual fun fights.

Winning in itself isn’t really goal for most servers, Because we’ve realized long ago that those who generally win matchup’s have more population…

I could give two craps if I gain points in a matchup, just like I give two craps if SBI wins a matchup..

What I do care about though is actual fun fights.

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

But only 2 matches this week have the server with the highest rating gaining more rating points. In 6 out of 8 NA matches, the highest-rated server is actually losing points.

Working as designed. The inability for points to flow down from higher tiers was a key deficiency in the old matchup system.

-ken

Xsorus.2507:

pulsecodesgnl.3470:

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

Congrats at missing the point, and adding obliviousness to an already obfuscated thread. Performance > Winning. If you need to win to enjoy WvW you should probably quit while you are ahead.

Getting Crushed in a match and gaining points isn’t “winning”

I’m on SBI, this matchup isn’t anymore “fun” then the previous matchup of getting zerged all day long.

People like yourself who believe that this “point” scale is good for the game completely miss the point of actual fun fights.

Winning in itself isn’t really goal for most servers, Because we’ve realized long ago that those who generally win matchup’s have more population…

I could give two craps if I gain points in a matchup, just like I give two craps if SBI wins a matchup..

What I do care about those is actual fun fights.

If you are completely unable to find fun fights you are playing wrong or looking for the game to change to your preferred meta. There are plenty others that seem to enjoy the match, and popping into your weekly thread, there are plenty on your own server that agree.

Why should the system cater to you? What makes you and your preferences more important than everyone elses?

The point of WvW is not PPT or score, it is rating.

• If you fight a stronger opponent and hold your own (while losing PPT), you will gain rating.
• If you fight a weaker opponent and win by only a little (while winning the PPT), you will lose rating.

If you randomly get assigned an opponent that is too strong for you, you are not doomed. Instead, this is a great chance to steal some of their rating points. If you randomly get assigned to a weaker tier, be careful not to let your guard down, or you might lose rating.

Every battle has meaning now.

There is always the opportunity for success, even against impossible odds. This will become the new meta, it is just a matter of time before everyone realizes it.

Why would you call this a bad thing?

Snowreap.5174:

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

But only 2 matches this week have the server with the highest rating gaining more rating points. In 6 out of 8 NA matches, the highest-rated server is actually losing points.

Working as designed. The inability for points to flow down from higher tiers was a key deficiency in the old matchup system.

-ken

If they’re losing points, then they’re going to be paired against more and more servers where they completely outnumber their opponent..

How is this not hard to understand…No one cares if a Tier 1 server is losing points and a Tier 4 server is gaining points if the Tier 1 server is zerging the ever living crap out of the Tier 4 server 24/7.

pulsecodesgnl.3470:

Xsorus.2507:

pulsecodesgnl.3470:

Xsorus.2507:

Rofl..

Only 1 match this week doesn’t have the Server with the Higher Rating being the one winning.

Congrats at missing the point, and adding obliviousness to an already obfuscated thread. Performance > Winning. If you need to win to enjoy WvW you should probably quit while you are ahead.

Getting Crushed in a match and gaining points isn’t “winning”

I’m on SBI, this matchup isn’t anymore “fun” then the previous matchup of getting zerged all day long.

People like yourself who believe that this “point” scale is good for the game completely miss the point of actual fun fights.

Winning in itself isn’t really goal for most servers, Because we’ve realized long ago that those who generally win matchup’s have more population…

I could give two craps if I gain points in a matchup, just like I give two craps if SBI wins a matchup..

What I do care about those is actual fun fights.

If you are completely unable to find fun fights you are playing wrong or looking for the game to change to your preferred meta. There are plenty others that seem to enjoy the match, and popping into your weekly thread, there are plenty on your own server that agree.

Why should the system cater to you? What makes you and your preferences more important than everyone elses?

Judging by the fact that our WvW population has decreased dramatically this week..I’m going to guess most of them share my opinion on the matter.

Also, since you’re leading on your server by close to 60k points…I think its safe to say you might be enjoying the system more being on the side zerging the crap out of everyone else.

Lets pair you against some T1 servers next week and see how much enjoyment you have.

Xsorus.2507:

If they’re losing points, then they’re going to be paired against more and more servers where they completely outnumber their opponent..

How is this not hard to understand…No one cares if a Tier 1 server is losing points and a Tier 4 server is gaining points if the Tier 1 server is zerging the ever living crap out of the Tier 4 server 24/7.

The only reason this EVER happened was because ANet left a stagnant system in place for far too long, and not using Glicko2 like it was intended to be utilized. Yes, it is going to be a kitten for a bit while it fixes itself, and the bandwagon servers of tier 1 and tier 2 may end up winning a match, but so what?

If you’re really going to emoquit because of one week, then good riddance. WvW will be better off without you.

Xsorus.2507:

Judging by the fact that our WvW population has decreased dramatically this week..I’m going to guess most of them share my opinion on the matter.

Also, since you’re leading on your server by close to 60k points…I think its safe to say you might be enjoying the system more being on the side zerging the crap out of everyone else.

Lets pair you against some T1 servers next week and see how much enjoyment you have.

LOL @ fairweathers leaving as an argument for changing the system. Yeah, we are winning, if you are really using that as an argument you are not at all familiar with the Glicko Ghetto or Sorrow’s Furnace, at all.

Every server that isn’t winning will lose their fairweathers, and that will show your true WvW population, if you really think that’s something we are unfamiliar with please review the historical matchup data on mos.millenium.

If you’re losing vast numbers because you aren’t winning then you don’t really enjoy WvW as much as winning, are you? Deal with a kitten match for a week and get over it. Crying in excess because you feel entitled to something different just creates additional problems and bad fixes.

pulsecodesgnl.3470:

Xsorus.2507:

If they’re losing points, then they’re going to be paired against more and more servers where they completely outnumber their opponent..

How is this not hard to understand…No one cares if a Tier 1 server is losing points and a Tier 4 server is gaining points if the Tier 1 server is zerging the ever living crap out of the Tier 4 server 24/7.

The only reason this EVER happened was because ANet left a stagnant system in place for far too long, and not using Glicko2 like it was intended to be utilized. Yes, it is going to be a kitten for a bit while it fixes itself, and the bandwagon servers of tier 1 and tier 2 may end up winning a match, but so what?

If you’re really going to emoquit because of one week, then good riddance. WvW will be better off without you.

Xsorus.2507:

Judging by the fact that our WvW population has decreased dramatically this week..I’m going to guess most of them share my opinion on the matter.

Also, since you’re leading on your server by close to 60k points…I think its safe to say you might be enjoying the system more being on the side zerging the crap out of everyone else.

Lets pair you against some T1 servers next week and see how much enjoyment you have.

LOL @ fairweathers leaving as an argument for changing the system. Yeah, we are winning, if you are really using that as an argument you are not at all familiar with the Glicko Ghetto or Sorrow’s Furnace, at all.

Every server that isn’t winning will lose their fairweathers, and that will show your true WvW population, if you really think that’s something we are unfamiliar with please review the historical matchup data on mos.millenium.

If you’re losing vast numbers because you aren’t winning then you don’t really enjoy WvW as much as winning, are you? Deal with a kitten match for a week and get over it. Crying in excess because you feel entitled to something different just creates additional problems and bad fixes.

https://dviw3bl0enbyw.cloudfront.net/uploads/forum_attachment/file/75598/gw115.jpg

Yea I can see why you must be having fun.

grin

By the way, we’re not losing vast numbers because we aren’t winning, We were Red on our matchup for months…. We stood no chance of beating SoS…

We’re losing numbers because no one likes to fight 10 to 1 odds..

titanlectro.5029:

The point of WvW is not PPT or score, it is rating.

• If you fight a stronger opponent and hold your own (while losing PPT), you will gain rating.
• If you fight a weaker opponent and win by only a little (while winning the PPT), you will lose rating.

If you randomly get assigned an opponent that is too strong for you, you are not doomed. Instead, this is a great chance to steal some of their rating points. If you randomly get assigned to a weaker tier, be careful not to let your guard down, or you might lose rating.

Every battle has meaning now.

There is always the opportunity for success, even against impossible odds. This will become the new meta, it is just a matter of time before everyone realizes it.

Why would you call this a bad thing?

Because it is not tracked in game in a visable way. Again lower tier (if they are like mine) servers have more marginal players which turn out for WvW when there it seems their participation will make a difference. If it does not there is always PvE.

Rating is not an explicit in game stat. like the score is. Explaining it is kinda pointless because it is often counter intuitive, and it is mostly the WvW involved players that will check mos.millenium.org and understand the rating system.

With matchups where you are getting beat into the ground, yet gaining rating, there is all these marginal players see is the score, so they are less motivated to turn out, which means we will be consistantly outnumbered (often by quite a bit) and have a problem drawing more people into liking WvW in general. You can not sustain ratings on that.

Xsorus.2507:

https://dviw3bl0enbyw.cloudfront.net/uploads/forum_attachment/file/75598/gw115.jpg

Yea I can see why you must be having fun.

grin

Retort with out of context ad hominem’s. At least you went out on a losing argument in a classy manner, oh wait…

If your server can’t deal with a week of being outmatched that should say enough in and of itself. Even HoD and ET are still fighting hard, and their numbers are prevalent. HoD just capped a majority of EB and still have a presence in our borderlands despite our overwhelming numbers and the fact that we will recap most of it shortly.

What’s your excuse? Improve your tactics or leave WvW like the rest of the fairweathers until the rest of your server has done the work for you.

pulsecodesgnl.3470:

Xsorus.2507:

https://dviw3bl0enbyw.cloudfront.net/uploads/forum_attachment/file/75598/gw115.jpg

Yea I can see why you must be having fun.

grin

Retort with out of context ad hominem’s. At least you went out on a losing argument in a classy manner, oh wait…

If your server can’t deal with a week of being outmatched that should say enough in and of itself. Even HoD and ET are still fighting hard, and their numbers are prevalent. HoD just capped a majority of EB and still have a presence in our borderlands despite our overwhelming numbers and the fact that we will recap most of it shortly.

What’s your excuse? Improve your tactics or leave WvW like the rest of the fairweathers until the rest of your server has done the work for you.

Rofl, HoD just capped a majority of EB… really?

OMG..I’m sure that will let them overcome the 60k Point advantage you already have.

Not to mention, You have a 200 point Score advantage over HoD…..Big difference no doubt…But DB has a 400 Point Advantage over SBI…

We’ll see how good you’re doing next week when you’re not the one paired against 2 servers you vastly outnumber.

We’ll see how much fun you’re going to be having.

Xsorus.2507:

Rofl, HoD just capped a majority of EB… really?

OMG..I’m sure that will let them overcome the 60k Point advantage you already have.

Not to mention, You have a 200 point Score advantage over HoD…..Big difference no doubt…But DB has a 400 Point Advantage over SBI…

Get. Over. Winning. If that was the case, 2/3rds of the entire game population would emoquit like you claim your server is. If you aren’t having fun with WvW, get out, it’s not meant for you.

pulsecodesgnl.3470:

Xsorus.2507:

Rofl, HoD just capped a majority of EB… really?

OMG..I’m sure that will let them overcome the 60k Point advantage you already have.

Not to mention, You have a 200 point Score advantage over HoD…..Big difference no doubt…But DB has a 400 Point Advantage over SBI…

Get. Over. Winning. If that was the case, 2/3rds of the entire game population would emoquit like you claim your server is. If you aren’t having fun with WvW, get out, it’s not meant for you.

People aren’t emo quitting over not winning, Its like you just don’t understand the concept since you’re on the zerging server.

The only good thing about this system is you will probably eventually get paired against a server that vastly outnumbers by a large margin, and we can watch your terrible server emo quit.

Looking at your server history, its been months where you’ve not dealt with being vastly outnumbered.

When you get paired against a server who’s got 400 points on you, you’re going to be in for a rude awakening.

Xsorus.2507:

People aren’t emo quitting over not winning, Its like you just don’t understand the concept since you’re on the zerging server.

The only good thing about this system is you will probably eventually get paired against a server that vastly outnumbers by a large margin, and we can watch your terrible server emo quit.

You mean the fairweathers? Yes, good riddance. They left when we we’re matched up against IoJ and DH and it was much more fun, people actually paid attention in map chat, it was easier to organize co-ordinated tactics, and the challenge gave us something to improve on.

The core of SF always looks forward to a fight where we are outmatched, we always learn and improve our tactics based on the scenario. When we win we get inflated again, and some of the new/fairweathers get picked up into the core. It’s how things work, relying on your fairweathers for fun is where you lose out on enjoyment.